# Research Papers — 2026-09 · bundle page 44 of 266 > Full-text markdown of papers #2151–2200 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/44 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/43.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/45.md) ### Page-Aware Retrieval-Augmented Generation for EvalLLM 2026: A Five-Variant Study on French PDFs Authors: Abdelhak kelious Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34776v1 We study retrieval-augmented generation (RAG) for questions about French PDF documents when both the answer and its supporting document pages are evaluated. Five system variants add dense retrieval, rank fusion, reranking, and query decomposition to a BM25 baseline. On 595 challenge questions, the complete system scores 0.4450 MRR@10 and 0.4013 Recall@10, compared with 0.3430 and 0.2994 for BM25. Dense retrieval alone and a simple lexical--dense fusion both underperform BM25. Reranking improves the hybrid system, whereas adding query decomposition produces the largest further gain, with higher latency and more detected output artifacts. The complete system slightly exceeds the reported anonymous overall mean on two answer metrics but falls below it on most page-retrieval metrics. These results identify accurate page selection, rather than semantic retrieval in isolation, as the main opportunity for improvement in this setting. --- ### Before the Token Commits: Trajectory-Level Benchmarking of Visual Hallucinations in Diffusion VLMs Authors: Yadong Wang, Siping Yue, Yu Tian, Chuanxing Geng, Xiang Chen Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34772v1 Multimodal diffusion language models generate responses by iteratively unmasking tokens, making each answer the endpoint of a multi-step trajectory rather than an immediate commitment. Hallucination benchmarks built for autoregressive models evaluate only the final output, and therefore cannot determine whether an unsupported claim in diffusion VLMs appears late or has already stabilized before any answer token is revealed. We introduce DynaHall, a trajectory-level benchmark of annotation-backed binary visual propositions covering object existence, counting, attributes, and relations, with controlled hard negatives graded by visual prior. DynaHall is paired with a commitment-aware protocol that records the intermediate answer tendency at every unmasking step alongside the committed output. Across five diffusion VLMs from three architecture families, visual hallucination is settled before commitment: an unsupported answer is already the preferred state while the answer position is still masked, and later unmasking steps rarely reverse it, so the failure is not introduced at the write step. This holds across decoding schedules, answer formats, and open-ended generation. DynaHall also exposes failures hidden by final-output metrics, including counting and relation collapse, prior-driven false positives, and attribute errors whose direction changes by type. Guided by this diagnosis, PGS (Pre-commitment Gradient Steering) edits still-masked answer states to reduce false positives, bringing the affirmation rate close to balance, and transfers to another architecture without degrading general ability. DynaHall and PGS suggest that hallucination should be measured and mitigated along the generation trajectory of diffusion VLMs, not only at the final answer. --- ### When Do Model Internals Help? Exploring the Role of Representation Engineering in LLM Safety Authors: Tianyi Guan, Jianhui Chen, Liangming Pan Published: 2026-09-28 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.34771v1 Reliable AI safeguards require both control mechanisms that reduce unsafe behavior and monitoring mechanisms that detect safety risks during model interactions. Established behavioral safeguards include alignment methods that optimize model outputs and text monitors that assess interaction text. Representation engineering instead reads or modifies internal model states, but the relative strengths of these approaches remain unclear because they are often evaluated under different settings. We present a matched evaluation across two tracks. For safety control, we compare DPO, a behavioral alignment method, with three representation steering methods across robustness, practicality, and granularity. DPO provides the strongest overall control and generally improves with increasing training data, although its safety can degrade after subsequent benign fine-tuning. Representation steering remains competitive primarily in low-data settings, particularly with high-quality contrastive data. For safety monitoring, we compare representation probes with fine-tuned and open-weight text monitors across full-response detection, early detection, and computational cost. Specialized text monitors achieve the strongest overall detection accuracy, while representation probes remain competitive at substantially lower marginal cost. Finally, monitor-guided interventions recover much of the safety lost by DPO after benign fine-tuning, with little additional over-refusal. Overall, representation engineering does not generally replace behavioral safeguards, but offers practical advantages under specific conditions and can provide complementary safety benefits. --- ### Reference-Grounded Data Curation for Instruction-Following Thai-English Machine Translation Authors: Thodsaporn Chay-intr, Krittapad Harnchang, Mahannop, Thabua, Kobkrit Viriyayudhakorn, Thanaruk Theeramunkong Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.34770v1 Instruction-following machine translation (IF-MT) requires respecting prompt-level rules on terminology, formatting, and register. Rule compliance typically trades off against translation quality, a tension that general-purpose IF data augmentation methods do not address. We propose Reference-Grounded Data Curation, a two-phase pipeline that extracts every supervised constraint from a reference translation that already satisfies it, ensuring feasibility by construction. Phase 1 applies Instruction-Following Difficulty (IFD) scoring to retain the hardest-but-learnable instances from an English-Thai parallel pool. Phase 2 extracts constraints from each reference target and keeps only generations satisfying every constraint, yielding the 1.97M-record Grounded dataset. We fine-tune open-weight bases on Grounded to produce ChindaMT, a Thai-English translation family at 4B, 2B, and 0.8B parameters. Under length-controlled pairwise judging, ChindaMT outperforms or matches every same-size baseline at every tier on both plain translation and under explicit rules, reaching up to a 68.4% win rate against the strongest baseline. The recipe transfers cleanly across Qwen generations. We release model weights, the Grounded dataset, and evaluation suites. --- ### LongPuzzleBench: Evaluating GUI Agents on Long-Horizon Visual Puzzles Authors: Bingo Zhang, Haochuan Lu, Zongjie Li, Genjian Li, Ari Yu Zhang, Chaozheng Wang Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.34769v1 GUI agents need long-horizon visual reasoning: they must interpret a changing interface while keeping a multi-step plan viable as earlier actions constrain later ones. Existing benchmarks evaluate grounding, computer use, and game play, but rarely test whether agents stay coherent across long chains of coupled decisions. Long-horizon visual puzzles expose this capability directly: a legal move that looks like progress can make the puzzle unsolvable, and the loss shows only several moves later. We introduce LongPuzzleBench, 114 levels in six puzzle games played through native GUI actions, where one objective can take a human over a thousand actions on persistent boards and dead ends go unannounced. With Native GUI Actions alone, the strongest agents solve most objectives, but success falls sharply on harder, longer boards: seven of ten general-purpose agents solve nothing harder than Medium, and none completes Bolt Unscrew Hard, which a human solves along with every other objective. Code Execution CUA does not close this gap, and its scores mix visual solving with algorithmic search. Controlled diagnostics trace these failures to one limitation that neither rules, state hints, nor failure memory removes: agents judge each move by the visible progress it makes, not by the future options it leaves. --- ### Privacy-Preserving Full-Body Meshing from mmWave Radar via Mesh Foundation Model Supervision Authors: Shuxing Zhang, Yongquan Ni, Zhenyu Ding, Yawen Lin Published: 2026-09-28 Categories: cs.AI, cs.CV arXiv: https://arxiv.org/abs/2609.34768v2 Millimeter-wave (mmWave) radar enables privacy-preserving human perception, but the extreme sparsity of point clouds from commercial single-chip sensors (mean ~6.5 points/frame; ~28% empty frames) has confined prior art to body-part keypoints or discrete action classification. We present a cross-modal teacher-student framework that lifts commercial radar to full-body, per-frame, metric 3D mesh reconstruction with per-joint uncertainty. Three innovations: (1) a mesh-foundation-model teacher - SAM 3D Body produces whole-body MHR ground truth (70 joints, 18,439 mesh vertices) from a single RGB frame with zero training, slashing annotation cost by orders of magnitude; (2) StudentPoseFormer - set encoding with masked attention pooling, a temporal Transformer, and a CVAE multi-hypothesis head that outputs both the pose mean and per-joint variance, honestly reporting where the radar cannot see; and (3) a multi-stage ground-truth quality pipeline (confidence gating, depth validation, temporal smoothing, bone-length consistency, bad-frame rejection) plus systematic information-lever ablations. On the public MM-Fi benchmark (same TI IWR6843 sensor, cross-subject), our full configuration reaches 7.45 cm 12-joint MPJPE, with ablations proving the causal value of point accumulation (k = 3, -0.34 cm), Doppler (-0.85 cm; -2 cm at the wrist on fast actions), and velocity loss (-0.27 cm). On our own synchronized radar + RGB-D corpus with block-level held-out splits, the pipeline achieves 21.47 cm end-to-end (per-joint hierarchy from 4.8 cm at the hip to 34.7 cm at the wrist - matching physical information limits), could be improved to 15 cm with ~30k diverse samples, and a scaling law shows sample diversity, not volume, is the binding constraint. Deployment inference is radar-only - no camera, no image. --- ### Beyond Reconstruction Loss in Post-Training Quantization: Balanced Fitting for Large Vision-Language Models Authors: Minchan Kang, Kyeonghye Park, Seungyeon Sa, Seoyoung Cho, Daeshik Kim, Yucheol Cho Published: 2026-09-28 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.34765v1 Post-training quantization (PTQ) enables efficient deployment of large vision-language models (LVLMs), but is typically calibrated on a small set while expected to generalize across diverse downstream tasks. Although recent PTQ methods for LVLMs incorporate sensitivity signals, they still minimize reconstruction loss with respect to the full-precision model, potentially over-preserving FP behavior and calibration-specific bias. Rather than treating quantization solely as an error to be minimized, we observe that it can also provide beneficial regularization for certain layers and modalities. Motivated by this observation, we propose Balanced Fitting, a quantization effect-based framework that balances precision and regularization beyond reconstruction-based optimization. By measuring layer- and component-wise quantization effects for weights, vision activations, and text activations, Balanced Fitting combines fine-grained fitting for sensitive components with coarser fitting to exploit potential regularization benefits. Experiments on multiple LVLMs show that our method consistently outperforms prior PTQ approaches under both weight-only and weight-activation quantization, while lower reconstruction loss does not reliably translate into better downstream performance. The source code is publicly available at https://github.com/kmc3661/BFQ --- ### WeaveData: A Multimodal Data Analysis System with Self-Critiquing and Self-Evolving LLM Plans Authors: Min Jia, Shihao Zhou, Jun-Peng Zhu, Peng Cai, Kai Xu, Chao Zhang, Li Li, Aoying Zhou, Heng Long, Qiu Cui, Liu Tang, Qi Liu Published: 2026-09-28 Categories: cs.DB, cs.AI arXiv: https://arxiv.org/abs/2609.34764v1 Multimodal data analysis, which answers questions over relational tables, text, and images, has attracted growing attention in the data management community. Large language models (LLMs) enable such analysis in natural language by generating analysis plans over relational and semantic operators. However, LLM-generated plans are error-prone: a plan may silently compute something other than what was asked, fail during execution, or return a result that misses the question. This paper presents WeaveData, a multimodal data analysis system with self-critiquing and self-evolving LLM plans. First, WeaveData generates a typed logical plan for each question and critiques it step by step before execution, and it checks the executed result against the question afterwards. Second, WeaveData evolves a plan that fails or misses the question: it diagnoses the failure with the actual data, reuses the results that remain valid, and accumulates planning experience for later questions. Third, WeaveData grounds planning in a metadata knowledge graph of all modalities, clarifies ambiguous questions with the user, and backs every model judgment with evidence in an interactive notebook. We demonstrate WeaveData on two public multimodal datasets. --- ### PanoVLN: Towards Effective Panoramic Vision-and-Language Navigation Authors: Zhen Wang, Changpeng Wang, Zhe Liu, Zhangyang Qi, Yuxiang Lu, Zimo Zeng, Donglian Qi, Xi Chen Published: 2026-09-28 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2609.34759v1 Recent vision-language models (VLMs) have advanced vision-and-language navigation (VLN), enabling models to predict navigation actions from visual observations and language instructions. In this work, we explore VLN with panoramic observations and introduce PanoVLN. The motivation is straightforward: more complete visual context should enable better-informed navigation decisions. For example, a panorama can reveal a passage outside a perspective camera's field of view, allowing the model to identify the intended route without additional exploration. However, we find that simply replacing perspective images with panoramas yields only limited gains. Our diagnosis suggests that fully exploiting wider visibility requires modifications to action prediction, training supervision, and visual representation. First, wider visibility supports longer-horizon action planning. We make the model predict longer action sequences, enabling larger turns and subsequent movement from a single panorama. Specifically, we introduce a confidence-guided execution (CGE) strategy that dynamically determines how many predicted actions to execute before replanning. Second, wider visibility also brings more complex route choices. We therefore construct training routes with frequent branching points and clear instructions to provide targeted supervision for route selection. Third, panoramic navigation requires understanding spatial relationships across viewing directions, beyond recognizing individual landmarks. We combine semantic and geometric features from RGB panoramas to capture both scene content and spatial layout without adding visual tokens. With a 4B backbone and RGB-only input, PanoVLN surpasses the previous SOTA by 11.9% and 8.7% in success rate on R2R-CE and RxR-CE Val-Unseen. Real-world experiments on a quadruped further demonstrate faster navigation with fewer pauses than prior VLN methods. --- ### Statistical Benefits of Fine-Tuning from Pretrained Initialization in Diagonal Linear Networks Authors: Alexandre Declèves, Etienne Boursier, Nicolas Flammarion Published: 2026-09-28 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.34756v1 Adapting pretrained models to downstream tasks with limited data has become a central paradigm in modern deep learning. Yet, despite its widespread practical success, how fine-tuning leverages information from pretraining remains poorly understood theoretically. We study fine-tuning from pretrained weights through the lens of sparse linear regression and two-layer diagonal linear networks. In our setting, pretraining provides information through the support (and signs) of the initialization predictor, which may contain coordinates relevant to the downstream task. We show how pretrained information reshapes the implicit bias and training dynamics, and can thereby reduce the sample complexity of recovering the target parameters and support. In particular, for a clean initialization with correctly inherited signs, we show that the required sample size is comparable to that of a weighted Lasso estimator that explicitly exploits the pretrained support through a suitably chosen regularizer. Our results thus show how information encoded in pretrained weights can be implicitly exploited by gradient-based fine-tuning, reducing the amount of data needed to recover a downstream task. --- ### Draft-KV: Learning Useful Latent Communication Between Language Models Authors: Linquan Wu, Shichang Meng, Tianxiang Jiang, Haoyu Yang, Peng Zhong, Fengming Zhu, Xi Peng, Linqi Song, Jacky Keung, Jingyu Zhang Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.34754v1 Latent communication passes internal states between language models instead of decoded text, but higher receiver accuracy does not show that the receiver used the message content. Across five method-dataset pairs, replacing each message with one from an unrelated question changes accuracy by at most 0.60 points, even when communication adds 15.44 points over the receiver alone. Thus the interface can supply the gain while making the sharer dispensable. Draft-KV instead sends the key-value states formed while the sharer drafts an answer to the current question. Linear projections place these states in a side memory read through a gated attention branch, and progressive training moves from message reconstruction to answer supervision under a guard on harm from mismatched messages. Both models remain frozen and the interface trains 1.05M parameters, 348x fewer than C2C. With a Qwen3-8B sharer, a frozen Qwen2.5-0.5B-Instruct receiver reaches 78.04% on MMLU-Redux, versus 37.45% alone and 36.40% with reassigned messages. At fixed interface size, scaling the sharer from 0.6B to 8B raises accuracy from 46.11% to 78.04%; communication also transfers to held-out tasks and can exceed both models when each holds different evidence. --- ### A Unifying Framework of Concept-based Explainable AI with Completeness Guarantees Authors: Vojtěch Kůr, Adam Kukučka, Tomáš Brázdil, Vít Musil Published: 2026-09-28 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2609.34750v1 Concept-based explanations describe neural network predictions through human-understandable properties of inputs called concepts. The field encompasses approaches that differ in how they define and represent concepts and connect them to model predictions. We introduce a theoretical framework that describes these approaches in a common mathematical language and supports a shared analysis of their properties. For concept discovery, which identifies concepts automatically within a latent space of a trained model, we employ a concept autoencoder view. An encoder extracts concept representations from the model's latent space, and a decoder uses them to reconstruct the original latent representation. The autoencoder's reconstruction error measures how accurately its decoder recovers the original latent representation. We revisit model completeness: how well the concepts can reproduce the model's outputs. We show that model incompleteness of the concepts can be bounded by the autoencoder's reconstruction error. The autoencoder view also provides a common way to define individual concept attributions, which measure each concept's contribution to a prediction. We establish when these attributions sum to the model's prediction, and bound the discrepancy otherwise, thus providing attribution completeness guarantees. --- ### CoDrive: Cross-Vehicle World-Consistent Video Generation with Precise Trajectory Control for Cooperative Driving Authors: Yu Meng, Baining Zhao, Junta Wu, Tengfei Wang, Rongze Tang, Haiyu Zhang, Wenqiang Sun, Chen Gao, Zhibo Chen, Xinlei Chen, Yong Li, Xiao-Ping Zhang et al. Published: 2026-09-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.34749v1 Real-world driving is inherently multi-agent, yet most existing driving world models generate observations from a single ego vehicle. Independently extending them to multiple vehicles does not ensure that different agents observe a consistent shared world. We present CoDrive, a cross-vehicle, multi-view driving video generation framework that jointly generates observations of vehicles sharing the same dynamic scene with precise camera-trajectory control. CoDrive interleaves local self-attention, which models spatiotemporal dependencies among the views of each vehicle, with global self-attention, which enables information exchange and consistency modeling across vehicles. To explicitly encode their spatial relationships, all camera trajectories are represented in a shared world coordinate system and injected into the attention layers through projective relative positional encoding. We further adopt a progressive mixed-task training strategy that combines large-scale real-world single-agent data with synthetic cross-agent interaction data, allowing the model to benefit from real-world appearance distributions while learning cross-agent consistency from simulation. For systematic evaluation, we introduce CoDrive-Bench, a benchmark covering real and synthetic multi-vehicle scenarios and evaluating trajectory controllability, scene geometry consistency, and instance-level consistency. Experiments show that CoDrive improves trajectory controllability and cross-agent geometric and instance consistency while maintaining competitive visual quality. --- ### No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation Authors: Seonghyeon Kim, Chaeyun Jang, Noah Lee, Boseop Kim, Juho Lee Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.34745v1 Multi-teacher on-policy distillation (MOPD) combines independently developed domain teachers into a single student by distilling their predictions on student-generated samples. We study a setting where teachers share a reference model but undergo different post-training procedures, and find that MOPD can struggle to recover some teacher capabilities. Because distillation occurs on student-generated prefixes, the student initialization can strongly affect subsequent recovery. However, initial benchmark performance is not a reliable predictor of a good MOPD initialization. For example, merge initialization can start below SFT warm-up yet finish higher after MOPD. We further find that effective merging depends on both the relative teacher contributions and the overall merge scale, with some strong configurations lying outside the simplex of convex parameter averaging. Thus, selecting a good merge initialization requires evaluating not only its immediate performance but also the learning it enables under MOPD, making one-shot coefficient search difficult. We propose Iterative Merging for MOPD (IM-MOPD), which starts from a uniform merge and progressively adds task-vector increments for under-recovered domains during distillation. In a 5-domain setting, IM-MOPD achieves higher average normalized recovery than MOPD with either uniform merge initialization or SFT warm-up, showing that effective teacher contributions can be determined progressively during training. --- ### Optimizing and Securing the Modern Watermarking Channel for Images Authors: Enoal Gesny, Eva Giboulot Published: 2026-09-28 Categories: cs.CR, cs.CV arXiv: https://arxiv.org/abs/2609.34744v1 To comply with recent regulations requiring traceable generated content, modern watermarking has adopted multi-bit post-hoc watermarking schemes. These modern designs rest on an encoder-decoder pair implemented as deep neural networks. These models are usually treated as pure black-boxes trained end-to-end, with the noise of the watermarking channel modeled through a fixed set of geometric and valuemetric transforms applied to watermarked images. We argue that this purely empirical approach leads to unquestioned design flaws and a lack of theoretical performance guarantees. This work proposes a general theoretical model of modern post-hoc watermarking schemes grounded in a statistical analysis of the outputs of the encoder/decoder pair. We show that these deep neural networks implicitly define a watermarking channel modeled as parallel AWGN channels, with messages transmitted using BPSK modulation. This imposes a binary alphabet, greatly limiting the capacity of these watermarking systems. Another fatal flaw is their lack of a secret key, making them intrinsically insecure. We make this notion of watermarking security precise for post-hoc schemes by linking it to the possibility of estimating the secret key under a given statistical model of the decoder's output. By putting together the results from this theoretical analysis, we introduce SNW: a novel post-hoc watermarking system that significantly outperforms existing state-of-the-art baselines in terms of capacity while also providing strong security guarantees. Notably, it does not depend on a fixed codebook or binary alphabet, allowing it to reach a rate close to Shannon capacity through the use of capacity-achieving error-correcting codes. --- ### Do Emotion Concepts Generalize Across Sources, Modalities, and Architectures in Vision-Language Models? Authors: Bohao Xing, Xin Liu, Kaishen Yuan, Deng Li, Rong Gao, Guoying Zhao, Xiaolan Fu, Heikki Kälviäinen Published: 2026-09-28 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.34742v1 Recent studies suggest that large language models encode emotion concepts as structured internal representations, but most existing work focuses on text and a single architecture. Therefore, we ask, do emotion concepts generalize across sources, modalities, and architectures in vision--language models (VLMs)? To address this, we construct CMES (Cross-Modal Emotion Stimuli), a multi-source collection of emotion-conditioned stories, real facial expressions, synthetic portraits, and synthetic emotion-evoking scenes. For each stimulus source, we extract a separate set of six Ekman emotion vectors from each of three VLMs. We report four main findings as follows: 1) Image-derived emotion vectors form a low-dimensional geometry similar to that of text-derived vectors. Valence is relatively stable across sources, while arousal varies more. 2) Text- and image-derived emotion vectors have modest cosine similarity but still show held-out cross-modal correspondence. Text-derived vectors can also steer image interpretation. 3) Cross-architecture correspondence remains even when native cosine is near zero. Transformations estimated from generic ImageNet activations recover both correspondence and causal transfer without using the six emotion vectors or their labels. 4) After aligning representations across architectures, we construct a shared emotion subspace that preserves affective geometry and selective steering effects. The corresponding consensus emotion vectors also generalize to a held-out fourth architecture at two model sizes. These results suggest that emotion representations can share relational structure and causal effects across sources, modalities, and architectures, even when individual vector directions differ. --- ### Predictive Dual Smoothing for Column Generation Authors: Senne Berden, Noah Schutte, Andrea Lodi, Tias Guns Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.34740v1 Solving large-scale linear programs efficiently is an important challenge in many optimization settings. A key technique is column generation, which alternates between solving the master problem over a restricted subset of the variables, and using a pricing subproblem to identify new variables to add. The pricing subproblem is guided by the dual solution of the current restricted master problem, but oscillations in these dual solutions can substantially slow convergence. Dual stabilization methods address this issue. Dual smoothing is a common stabilization method, which guides the pricing subproblem using a combination of the current dual solution and duals from previous iterations. However, while past dual solutions can stabilize the dual trajectory, they do not necessarily guide pricing towards useful new variables. We therefore introduce predictive dual smoothing, which instead combines the current dual solution with a learned prediction of future duals to steer pricing towards variables that are more useful in subsequent iterations. The predictor is trained offline using supervision extracted from standard column generation trajectories and is used only to modify the pricing subproblem's objective function, while exact reduced-cost checks and fallback pricing with the unsmoothed duals preserve correctness. Experiments on cutting stock and generalized assignment problems show that predictive dual smoothing substantially reduces generated columns and wall-clock time relative to standard column generation and existing classical and learned stabilization methods. These gains extend to out-of-distribution instance sizes, and predictive smoothing provides further improvements when combined with strong classical stabilization. --- ### Beyond Token Alignment: Event Completion for Cross-Tokenizer On-Policy Distillation Authors: Jiacheng Liu, Jingwei Song, Qituan Zhang, Siheng Chen, Linfeng Zhang Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.34738v1 On-policy distillation (OPD) transfers knowledge between language models through teacher supervision on student-generated trajectories. With different tokenizers, a single teacher token may require multiple student tokens to generate, creating intermediate states where the event is entered but not yet completed. Existing cross-tokenizer methods align tokens or text spans to construct comparable prediction targets. We study a complementary problem after partial generation: once the student produces a prefix of a teacher token, multiple next tokens may complete the same remaining bytes, but the teacher only specifies the required completion rather than how probability should be divided among these valid continuations. We introduce Event-Set Completion Distillation (ESCD), which complements cross-tokenizer probability alignment with completion-set supervision. ESCD aggregates prefix-related teacher events and supervises the total probability of byte-compatible one-step student completions, avoiding tokenizer-dependent probability splits among individual tokens. The method reuses student trajectories and predictions, requiring neither additional rollouts nor changes to the student vocabulary. Experiments demonstrate consistent gains in mathematics, code, and scientific reasoning across model families and tokenizers, extending to large-scale MoE distillation from a 1T teacher to a 35B student. Local analyses show that retaining completion sets better matches the reference supervision, while one-step completion covers over 99% of observed compatible teacher mass after partial event entry in the studied tokenizer pairs. These findings support event entry and event completion as complementary supervision targets for cross-tokenizer knowledge transfer. Code will be released on GitHub. --- ### SeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing Authors: Vasilis Perifanis, Nikolaos Pavlidis, Symeon Symeonidis Published: 2026-09-28 Categories: cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.34736v1 Large language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query embeddings, model representations, preference data, or clusters of similar examples. Such approaches can be effective, yet the representation used for routing rarely states what a query actually requires. We introduce SeLMRoute, a routing framework that separates the extraction of candidate-independent semantic evidence from the learning of candidate performance and the application of deployment objectives. A decision model first evaluates a set of interpretable questions about the query, such as its reasoning requirements and use of external knowledge, with each judgment retained as a probability distribution. The resulting probabilistic semantic state is used by a lightweight supervised router to estimate candidate model performance. Routing objectives are applied after performance estimation, which allows the same semantic state to support performance-oriented and cost-aware decisions. On the LLMRouterBench (15 datasets, 20 candidate models, 11,481 queries), SeLMRoute achieves an average accuracy of $72.08\% \pm 0.45$, while grouped five-fold out-of-fold evaluation reaches $72.64\%$, compared with $69.23\%$ for the strongest fixed candidate. The representation achieves the highest mean performance among the evaluated semantic, dense, lexical, and domain-level representations. In a separate 13-model performance-cost setting, SeLMRoute improves performance in all five grouped splits, with a mean PerfGain of $2.66\%$. Our code is available at https://github.com/Indigma-Innovations/SeLMRoute. --- ### From Human Narrative to Harmonic Structure: A Human-Centered Investigation of Algorithmic Music Generation through the Chord Wheel Diagram Authors: Josef Pavlíček, Petra Pavlíčková, Irena Štrausová Published: 2026-09-28 Categories: cs.AI, cs.SD arXiv: https://arxiv.org/abs/2609.34735v1 Contemporary AI-based music generation can produce compositions that satisfy formal requirements of tonality and musical coherence. However, whether musical expression can be described by mathematical properties alone remains a fundamental question. Human composers operate within personal and cultural contexts that influence harmonic decisions and deliberate departures from established patterns. This study investigates six narrative-driven popular songs by Bob Dylan, Johnny Cash, and Ritchie Valens. Original human harmonies are compared with outputs of an explainable computational harmonizer operating on the same melodies without access to the original chord progressions. We examine harmonic vocabulary, functional persistence, repetition, non-diatonic events, and tension-resolution patterns using Chord Wheel Diagrams and BPMN-based representations. Results show that high melody-chord compatibility does not necessarily imply preservation of the original human harmonic decision pattern. Some generated harmonizations retain the economical structure of the reference, while others alter harmonic diversity or suppress distinctive events while remaining compatible with the melody. Rather than quantifying artistic quality, the study introduces narrative-conditioned harmonic structure as a complementary perspective for computational music analysis. The findings suggest that generative systems may benefit from modeling not only harmonic correctness, but also structural identity, context, and human compositional intention. --- ### SurgGMF: Fully Causal Gaussian Motion Forecasting for Anticipatory Surgical Scene Rendering Authors: Jingqian Sun, Yichao Tang Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34733v1 Dynamic surgical scene modeling is essential for robotic perception, simulation, and decision support. Although existing neural rendering methods enable efficient reconstruction and rendering of deformable surgical scenes, they remain primarily focused on observed-frame reconstruction rather than forecasting future scene states. To this end, we present SurgGMF, a fully causal Gaussian motion forecasting framework for anticipatory surgical scene rendering. Rather than predicting future RGB images directly, SurgGMF forecasts future Gaussian motion states represented by position, scale, and rotation residuals (X/S/R) from historical Gaussian motion fields. To prevent target leakage, we introduce a full-causal-last rendering protocol, where future Gaussian states are rendered without accessing target-frame Gaussian attributes while preserving causal appearance propagation. We evaluate SurgGMF on 12 EndoNeRF and StereoMIS video slices using neural temporal learners and classical dynamics baselines under a unified forecasting protocol. Learned Gaussian motion forecasting consistently outperforms classical dynamics baselines in render space, demonstrating gains beyond hand-crafted state extrapolation. Latency analysis further reveals an accuracy--efficiency trade-off: under the current implementations, TKAN achieves the highest accuracy, whereas GRU and LSTM provide more favorable module-level latency profiles. These results establish SurgGMF as a reproducible framework for causal Gaussian motion forecasting and advance surgical Gaussian representations from retrospective reconstruction toward predictive scene modeling. --- ### What Visual Generators Need from Teachers: Rethinking Representation Alignment Authors: Yongcong Wang, Hingchin Chen, Mingyu Fan, Shuo Jiang, Teer Zhang, Yucong Sun, Zijia Wang, Yiming Lu, Chengchao Shen Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34732v1 Representation alignment speeds up diffusion transformer training by pulling an intermediate block of the model (student) toward features of a frozen pretrained encoder (teacher). Which teacher layer to align, and for how long, is still set by convention, and each alternative costs a training run. We find that alignment helps where the student cannot linearly recover the teacher's features, not where it already resembles them. Since a deep teacher layer is largely predictable from the one below, we isolate what each layer adds, its increment, and measure how much of it an unaligned student recovers. The student fills the teacher's hierarchy from the bottom up and stalls near the top, which we call hierarchy filling: even after 400K steps it recovers almost none of the deepest. The recoverability gap is the unrecovered share of an increment, read from one unaligned checkpoint. In short runs that each align one teacher layer at one block, the gap nearly reproduces their ranking by FID improvement, and CKA, a measure of feature similarity, largely reverses it. Representation Alignment and Recoverability Estimation (RARE) picks the teacher layer with the largest gap before training. During training, it tracks each token's remaining distance to that layer, the online counterpart of the gap, weights tokens by it, and phases out the loss once the average distance stops falling. With SiT-B/2 on ImageNet $256\times256$, RARE reaches an FID of 18.02 without guidance and 4.46 with it, ahead of seven alignment baselines including REPA, iREPA and HASTE. It also trains in 14% fewer GPU-hours than iREPA. Its FID stays below iREPA's across model scales, teachers, datasets and backbones. --- ### Information-Theoretic Analysis of Next-Token Prediction under Markovian Data Authors: Masoud Kavian, Abdellatif Zaidi, Milad Sefidgaran Published: 2026-09-28 Categories: stat.ML, cs.IT, cs.LG arXiv: https://arxiv.org/abs/2609.34731v1 We develop an information-theoretic framework for generalization in next-token prediction under temporally dependent data. We consider independent trajectories generated by finite-memory Markov processes and distinguish algorithmic dependence, quantified by mutual information, from temporal dependence, characterized by mixing. For cross-entropy loss, we derive an expected generalization bound using the Donsker--Varadhan variational representation and a McDiarmid-type concentration inequality for Markov chains. A refinement captures the joint effect of context length and temporal mixing through the mixing properties of the history-state process. We then extend the bound through a rate--distortion formulation, replacing mutual information with the minimum information rate required to represent the learned model within a prescribed distortion in the generalization gap, yielding informative guarantees for deterministic algorithms over continuous hypothesis spaces. For margin-based prediction, we derive explicit bounds for linear and self-attention next-token predictors via noisy low-dimensional compression, revealing the roles of context length, model complexity, sample size, margin, and temporal mixing. Experiments on TinyStories and ETTh2 show that longer contexts can reduce both training and test losses, but typically reduce training loss more, enlarging the generalization gap. A complementary ETTh2 analysis identifies an effective predictive-memory scale near 24 hours, with no statistically supported improvement beyond this scale, offering a plausible explanation for test-performance saturation at larger contexts. --- ### Edge-Level Automorphism in GNNs: A Quantitative Framework and Effective Designs For Link Prediction Authors: Chen Shao, Donald Loveland, Tobias Käfer, Danai Koutra Published: 2026-09-28 Categories: cs.LG, cs.GT arXiv: https://arxiv.org/abs/2609.34729v1 Graph Neural Networks (GNNs) are effective for learning node and link embeddings through permutation-equivariant aggregation. However, standard GNNs collapse automorphic nodes, i.e., those with identical structural roles (or orbits) into indistinguishable representations, leading to the node automorphism problem. This collapse limits their expressive power and degrades link prediction performance. Existing approaches to characterize GNN expressiveness rely primarily on Weisfeiler-Lehman (WL) analyses, but these methods are typically qualitative and often misaligned with empirical results. To address this gap, we begin by introducing a novel quantitative framework to assess GNN expressiveness for link prediction. We first formalize edge-level automorphism through edge orbits, which capture the set of structural role pairs for nodes that share a link. Then, we introduce the edge automorphism ratio (EAR), a scalar metric that quantifies a GNN's ability to distinguish links in a given graph. We empirically demonstrate that EAR correlates strongly with performance, validating its practical benefit. Building on this insight, we design EDGE-ORBIT EQUIVARIANT GRAPH NEURAL NETWORK (EO-GNN), a GNN architecture that addresses automorphism collapse while preserving equivariance and incurring minimal computational overhead. EO-GNN accomplishes this through two core designs combined with WL-based node hashes: (i) automorphism-aware dropouts and (ii) subgraph orbit-biased aggregation. Empirical evaluations on synthetic and real graphs show improvements of up to 42.36% and 28.44%, respectively, in predicting links in scenarios with high automorphism. --- ### Dynamic Flow, Static Graph: KV Cache Reuse for Efficient LLM Serving on Mobile NPUs Authors: Zhengxiang Huang, Shengheng Chen, Chaoyue Niu, Yujie Sun, Zhaode Wang, Zeyu Zhao, Chengfei Lv, Fan Wu, Guihai Chen Published: 2026-09-28 Categories: cs.OS, cs.AI, cs.DB arXiv: https://arxiv.org/abs/2609.34727v1 On-device large language model (LLM) serving is a cornerstone of local-first personal intelligence, offering users data sovereignty, strong privacy guarantees, and freedom from cloud API latency and cost. Although KV caching is widely used to reduce latency in long-context inference, existing designs were primarily optimized for cloud GPUs with dynamic execution environments and abundant memory bandwidth. These architectural assumptions do not hold on mobile NPUs, where computation graphs must be statically compiled and both memory capacity and I/O bandwidth are severely constrained. In this work, we present a compute-storage co-design for mobile-centric prefix and non-prefix KV reuse. We first propose an intra-graph mechanism that maps selective KV recomputation onto static NPU graphs, reconciling algorithmic dynamicity with NPU staticity. We further develop an inter-graph scheduler to optimize chunk merging and minimize padding with dynamic programming. To address mobile bandwidth limitations, we introduce a hierarchical KV manager featuring a tree-hash-semantic hybrid structure, along with cost-aware prefetching and eviction policies. We also build a two-dimensional pipeline that overlaps KV loading, rerotation, and storage with NPU execution, hiding data-movement latency. Experiments across representative on-device workloads and LLMs show that our design reduces time-to-first-token (TTFT) by $40-60\%$ compared with no reuse and prefix-only caching. --- ### Geometry as Address: Routing Attention to Visual Memory for Long-Horizon Camera-Controlled Video Generation Authors: Zesong Yang, Weikai Chen, Liyuan Cui, Lutao Jiang, Runze Zhang, Yingda Yin, Xiaoyang Huang, Kai Yan, Keyang Luo, Wangguandong Zheng, Xin Wang, Hujun Bao et al. Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34722v1 Long-horizon camera-controlled video generation requires recovering previously observed content from an ever-growing visual history. Existing approaches either search historical context implicitly or reconstruct it into persistent 3D memory, facing inefficient memory access or accumulated geometric errors. Our key insight is that geometry need not explain the scene--it only needs to determine where visual memory should be read from, while attention decides what should be recovered. Based on this insight, we introduce GEAR, a Geometry-Enabled Attention Routing framework that uses geometry as an explicit token-level address for visual memory. Rather than fusing historical observations into a persistent global 3D representation, GEAR retains them as frame latents and uses per-frame geometry only to establish token-level correspondences with target views, thereby avoiding persistent error accumulation from global fusion. Guided by these correspondences, Geometric Correspondence Attention (GCA) selectively injects geometrically matched historical features into noisy target patches during denoising. We further introduce an Invisible Octree to accumulate visibility evidence and reject geometrically plausible but occluded correspondences. Extensive experiments demonstrate that GEAR achieves state-of-the-art visual quality, precise camera control, and revisit consistency, enabling minute-long video generation along challenging trajectories. --- ### DBCF: Dual-Branch Complementary Fusion of Foundation Models for Generalized Deepfake Detection Authors: Fengming Gu, Mingjie He, Zonghui Guo, Jie Zhangb, Shiguang Shan Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34720v1 As image generation and editing technologies have progressed substantially, facial forgeries pose significant challenges to privacy and public safety. Due to limited ability to capture forgery cues, existing small-scale forgery detection models often struggle to generalize across various domains and unseen manipulations. To address this limitation, researchers have turned to large-scale foundation models, which can provide richer representations and better generalization. Nevertheless, relying on a single foundation model alone remains insufficient for effective forgery detection. While models like CLIP offer robust global semantic cues, they lack the capacity to capture detailed local facial features. In contrast, DINO excels at capturing local structural features of faces, but provides weaker global semantic context. To fully utilize the synergies among multiple foundation models, we propose a hierarchical multi-granular framework that integrates complementary pretrained representations. Specifically, a Global Context Branch (GCB) based on CLIP captures holistic semantic cues, while a Fine-grained Cue Branch (FCB) built on DINOv3 captures localized structural irregularities. In addition, we design a feature fusion module that enables parameter-efficient adaptation of the frozen foundation backbones by adaptively extracting and integrating complementary features from the two models. By jointly leveraging global context and fine-grained cues, our method learns more comprehensive forgery representations and achieves strong cross-manipulation performance. Extensive experiments on multiple benchmarks demonstrate the benefit of the proposed design, particularly under cross-dataset and cross-manipulation settings. --- ### Quality Determines Direction, Length Shapes Magnitude: Length Control for Open-Ended Reinforcement Learning Authors: Zijun Weng, Zhongan Bi, Xuanang Gao, Xiaohui Hu, Shuangyong Song, Yongxiang Li, Kaidong Yu, Xuanjing Huang Published: 2026-09-28 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2609.34718v1 Reinforcement learning (RL) changes not only what language models say, but also how much they say, often increasing response length at the cost of token efficiency. Controlling this length growth is particularly challenging in open-ended RL because (i) response length is entangled with quality, (ii) open-ended tasks lack a natural success boundary for deciding when efficiency should be prioritized, and (iii) dense, graded rewards often yield small within-group quality margins, making quality-induced advantages especially sensitive to reward-level length shaping, which can perturb their magnitudes and even reverse their signs. We therefore adopt an asymmetric principle: quality should determine the direction of reinforcement, while length should only shape its magnitude. We instantiate this principle with Quality-Gated Length Advantage Shaping (QGLAS), which first computes advantages from quality rewards alone, then adds bounded bonuses only to shorter positive-advantage responses, leaving all other advantages unchanged. The bonus strength is further adapted to within-group quality separation, allowing conciseness to matter more when quality-favored responses are similar and less when their quality differences are clear. Across different model families, open-ended benchmarks, and reward sources, QGLAS consistently achieves a stronger quality--length trade-off than representative baselines. At approximately 30% compression, QGLAS retains 98.4--102.0% of the macro-average quality gains achieved by quality-only RL over the base model, compared with 68.3--75.5% for these baselines at comparable compression. --- ### ReMCTS: Reflection-Enhanced Monte Carlo Tree Search for Code Generation Authors: Huifei Wang, Xinying Huang, Yiheng Sun, Yifan Yuan Published: 2026-09-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.34717v1 Open-weight large language models (LLMs) can generate function-level programs from natural-language prompts, but plausible candidates still fail on hidden semantics and repeat mistakes across repair attempts. We present ReMCTS, an execution-grounded, memory-augmented, LLM-guided MCTS-style search framework. It organizes program candidates as tree states, retains branch-local debugging context, retrieves failure experience across branches, and distinguishes failed checks from unavailable evidence. On HumanEval and MBPP-Sanitized, visible-test ReMCTS improves over direct generation in 8 of 10 model-dataset pairs under held-out evaluation, whereas proxy-only search is less stable. Controlled tree-search, sampling, repair, and memory ablations characterize the source and limits of these gains. A 30-task HumanEval-X C++ pilot further demonstrates compatibility with compiler-backed execution, but does not constitute a broad multilingual evaluation. --- ### From Pixel Generation to Topological Inference: Structural Dual Super-Resolution for Trustworthy Cross-Physical-Domain Trabecular Morphology Learning Authors: Fan Zhang, Yi Zhang, Ling Wang Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34716v2 Clinical CT and UHRCT cannot resolve individual trabeculae, whereas synchrotron radiation microCT (SRuCT) provides high-resolution references but is not applicable for in vivo imaging. The two domains differ by a 32x resolution gap, are only coarsely paired, and exhibit severe physical differences including partial volume effects, noise, and artifacts. Existing super-resolution networks and pretrained-prior methods (GLEAN/StyleGAN2, Stable SR/LDM) underperform because they target pixel generation---diverse details and SSIM/PSNR---and do not model these physical differences. Pixel generation for a 32x resolution gap is intrinsically ill-posed. We propose a paradigm shift from pixel generation to topological inference: deterministically predicting invariant microstructures from macro-scale low-resolution inputs, evaluated by morphological parameters. The core of our 2D morphology learning lies in training on 2D slices while evaluating on 3D morphological parameters, ensuring that the learned representations capture true three-dimensional trabecular topology rather than 2D pixel statistics. We realize this paradigm via structural dual super-resolution, coupling forward physical degradation (micro-to-macro) with inverse structural inference (macro-to-micro) through structural duality constraints. The method is an end-to-end, few-shot, compact structural dual network (SDN), comprising a bidirectional modeling network, a multi-scale structural consistency discriminator, and four structural duality constraints. On the test set, SDN achieves morphological parameters largely consistent with SRuCT across six metrics, with SSIM reaching 0.8. Trained on 3.2 um SSRF data, the model generalizes well to 3.25 um isotropic BSRF data from an independent source, validating cross-source generalization and confirming trustworthy structural inference rather than pixel generation. --- ### PDE-JEPA: Predictive Representation Learning of Latent Dynamics Modeling for Parametric PDEs Authors: Zhentao Tan, Jianrong Zhang, Ruijie Quan, Yi Yang Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34715v1 Physical trajectories contain more than snapshots of a system: they also reveal how its states evolve under governing conditions. However, representation learning for parametric partial differential equations (PDEs) has largely relied on reconstruction-based objectives that emphasize recovering observed physical fields. In this paper, we investigate predictive representation pretraining as an alternative to reconstruction-based learning. We find that predictive representations preserve rich physical information, yet this advantage alone does not ensure accurate field evolution. Based on these observations, we introduce PDE-JEPA for parametric PDE dynamics. Specifically, we first train an encoder using a masked-latent prediction to capture the underlying regularities of PDE dynamics. To explicitly adapt the pretrained representation toward a more dynamics-aligned state space, we then introduce a geometry projector that aligns latent trajectory geometry with the evolution geometry of physical fields. Finally, building on this geometry-aligned latent space, we further develop a physics-structured latent predictor that decomposes the dynamics into parameter-independent evolution and parameter-dependent response components. Extensive experiments on nine widely used PDE benchmarks demonstrate that our framework outperforms existing state-of-the-art methods by an average of 33.4\% in-distribution, while achieving an average improvement of 51.4\% when extrapolating to unseen governing parameters. The project page is available \href{https://tanpig-x.github.io/PDE-JEPA/}{here}. --- ### RSI-Router: Evolving Subtask-Level LLM Routing and Skills for Cost-Efficient Agents Authors: Hao Li, Hangfan Zhang, Zhiyao Cui, Chunjiang Mu, Yiqun Zhang, Bo Zhang, Danyang Jia, Shuyue Hu Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34712v1 Practical deployment of large language model (LLM) agents requires strong task performance at affordable inference cost. For long-horizon agentic tasks, this performance-cost trade-off can be improved through within-task large-small model collaboration, as smaller models can handle some stages even when they cannot solve the full task. In this paper, we introduce RSI-router, a routing framework that constructs subtask-level model assignments and model-specific skills through recursive self-improvement over accumulated experience. Each iteration consists of four stages: Subtask Mining derives subtask definitions and identification rules from training trajectories; Routing Strategy Evolution proposes and evaluates diverse model assignments; Model-Specific Skill Evolution compares routed and large-model-only trajectories to diagnose failures and develop reusable execution skills; and Pareto-Optimal Router Selection updates the Pareto population using historical and newly generated routers while retaining dominated routers as experience for subsequent evolution. Routing between DeepSeek-V4.1-Flash and Qwen3.5-9B, RSI-router consistently surpasses the DeepSeek-only baseline at roughly half the inference cost (48.3%) across five agentic benchmarks. In particular, on ALFWorld, ScienceWorld, and WebShop, it cuts inference cost by 74.7-82.2% while simultaneously improving performance; on Terminal-Bench 2.0, it achieves a 16.7% relative performance gain at 18.0% lower cost. Moreover, RSI-router establishes a stronger performance--cost Pareto frontier than 9 routing methods. --- ### Learning Propagation Geometry from Message-Passing Feedback Authors: Yingxu Wang, Kunyu Zhang, Xinwang Liu, Mengzhu Wang, Siyang Gao, Chang Tang, Nan Yin Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.34711v1 Learning local geometry enables graph neural networks (GNNs) to adapt how they compare and integrate neighborhood information. However, estimating geometry from aggregated representations can overlook variation among individual messages and dependencies across feature dimensions. We propose GeoF, a recurrent framework that jointly evolves node features and propagation geometry through message-passing feedback. Each node maintains a local symmetric positive-definite geometry, initialized from a structure-aware prototype atlas and parameterized in block log-triangular coordinates. At each step, the geometry determines neighborhood weights, while triangular frame transport maps transformed source messages into the target node's local coordinates before aggregation. Weighted second-order statistics of residuals between aligned messages and the transformed target state capture directional variation and within-block dependencies, yielding a geometric update target. A shared controller learns complementary corrections through task supervision. A bounded log-triangular update combines these corrections, the target, and the previous geometric state while preserving positive definiteness. The geometry governs subsequent propagation, closing the feedback loop. With parameters shared across recurrent steps, task-specific readouts support node classification, link prediction, and graph classification. Experiments on benchmark datasets show that GeoF consistently outperforms state-of-the-art GNN baselines. --- ### FromPitch2Board: Benchmarking LLM Agents in Long-Horizon Football Management Authors: Peiyu Zang Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34710v1 Long-horizon agent benchmarks typically report how far an agent progresses, but do not identify whether its performance comes from the foundation model, scaffold, responsibility scope, match-control granularity, or horizon. We introduce FromPitch2Board, a deterministic football-management benchmark that studies five configurable factors through controlled comparisons on a single simulator, using paired seeds and a frozen calibration. We evaluate four foundation models and four agent scaffolds. In the Model Track, Coach points Z-scores span 0.19, while Manager points Z-scores span 0.68, with GPT-5.6 showing a sharp rise in passivity under responsibility expansion. Its responsibility ladder rises from 46.1 to 58.1 points with recruitment, then falls to 46.8 under full management, localizing the regression to the final responsibility boundary. Across that boundary, its skipped-decision rate rises from 1.1% to 57.9%. Within the Flash-Pro pair crossed across every scaffold, scaffold choice changes Manager points Z-scores by up to 0.48 relative to the fixed stateless scaffold. The 3Y cohort shows a directional reversal in mean ranking between years one and three, while a selected Claude Code+Pro configuration peaks in year three and remains below that peak, showing that responsibility scope and horizon expose behavior changes that a single headline score conceals. --- ### ResonAct: Streaming Metrics for Runtime Diagnosis and Self-Healing in Multi-Agent Systems Authors: Tarun Chintada, Neelamadhav Gantayat, Ishaan Romil, Renuka Sindhgatta, Soujanya Soni, Sameep Mehta Published: 2026-09-28 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.34701v1 Multi-agent systems (MAS) are increasingly used to automate enterprise workflows involving multiple specialized agents, external tools, and long-running task execution. Failures may arise from tool degradation, context propagation errors, coordination breakdowns, or repeated agent interactions that prevent task completion. While existing observability frameworks provide traces and logs, diagnosis and remediation are largely performed after execution completes, limiting opportunities for recovery during runtime. We present ResonAct, a runtime self-healing framework that enables continuous monitoring, diagnosis, and remediation of multi-agent systems through streaming operational metrics. ResonAct ingests execution traces, agent interactions, and tool invocations into a streaming analytics layer that continuously derives task progress, context health, and tool reliability metrics. These metrics serve as runtime control signals for detecting anomalous execution patterns and localizing root causes using a structured failure model. Based on the diagnosed failure, ResonAct dynamically selects remediation policies and performs actions. The framework operates as an external control plane, enabling intervention without modifying application agents or orchestration logic. We evaluate ResonAct across enterprise workflow scenarios and AppWorld benchmarks. The results show that the streaming metric-based analysis identifies execution degradations and localizes faults. Furthermore, policy-driven remediation improves task completion rates by up to 10.00 percentage points, with detection precision ranging from 70.59% to 82.91%, recall from 63.09% to 100%, recovery rates from 10.48% to 46.67%, and runtime overhead ranging from $-0.25%$ to 14.12% across the evaluated configurations. --- ### Triangular Resampling for Long-Horizon Motion Generation Authors: Kunhang Li, Yiyi Cai, Xiangyue Zhang, Fangyuan Tu, Yuhan Wu, Zhixiang Wang, Kaipeng Zhang, Haiyang Liu Published: 2026-09-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.34697v1 We introduce Triangular Resampling (TR), a post-training method for mitigating long-horizon error accumulation in motion diffusion models. Built on FloodDiffusion's triangular denoising schedule, TR addresses the mismatch between ground-truth-derived training windows and model-generated inference states. Replacing only completed motion history leaves this mismatch unresolved in partially denoised states within the active window. TR therefore extends rollout-based training to these states, using ground-truth clamping to limit excessive drift. For each replayed sample, TR draws one denoising threshold, shared across latent positions and replay updates, and replays multi-step triangular denoising without gradient tracking. After each update, states below the threshold are replaced with noise-matched ground truth, while those at or above it retain model predictions. The resulting latent window enters the standard training update. This rollout construction supports both supervised training (TR) and distribution matching (TR-DMD). On 120-second motion generation from HumanML3D test prompts, TR and TR-DMD achieve state-of-the-art FID AUC within their respective non-DMD and DMD comparison groups. Supervised TR reduces FID AUC by 40.9% and FID degradation slope by 55.3% relative to matched post-training without replay. --- ### Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control Authors: Yisheng Zhang, Tao Wang, Sicun Gao Published: 2026-09-28 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2609.34695v1 Reward shaping is fundamental to modern robotic control with deep reinforcement learning (RL), yet practitioners still rely heavily on heuristic principles borrowed from classical optimal control and trajectory optimization. Existing methods rarely distinguish reward terms that are intrinsic to the control objective from numerical regularizers, leading to brittle hyperparameter tuning. To determine which quantities a reward must contain, we study the stabilization control problem with a focus on zeroth-order (configuration) and first-order (velocity) information. We theoretically and empirically demonstrate that policy gradient methods can successfully solve stabilization tasks without first-order reward terms, adding such terms can instead introduce severe sensitivity as their scale grows. Conversely, our findings confirm that reward functions must be zeroth-order complete over goal-relevant coordinates, while the first-order state remains necessary in the policy observation under our low-dissipation assumptions. Overall, these results provide actionable and principled guidance for reward design in robotic RL. --- ### Predicting Delayed Train Trajectories on the Dutch Railway Network: Explainable AI Evaluation of Topological, Operational and Weather Features with Tree Based Ensemble Methods Authors: Jia Long Bao, Ali Mohammed Mansoor Alsahag, Seyed Sahand Mohammadi Ziabari Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.34692v1 The reliable prediction of passenger train delays is a critical component of railway management. While contemporary research frequently attempts to maximize absolute accuracy by deploying opaque deep learning architectures, the underlying data mechanics driving longitudinal predictive decay remain underexplored. Consequently, this study provides an explainable temporal robustness analysis of network-wide railway delay prediction. Focusing on the Dutch railway network, this research utilizes interpretable tree-based ensembles to integrate granular topological, environmental, and operational features. The overarching finding establishes that while feature-rich tree-based models improve simultaneous (within-month) prediction, predictive performance systematically degrades when evaluated across non-simultaneous (future) months. Furthermore, multi-horizon SHAP and dispersion analyses explicitly link this degradation to environmental feature volatility and instability within the statistical target definition. Ultimately, this thesis demonstrates that richer feature sets alone are insufficient to resolve long-term forecasting constraints, underscoring the necessity to transition toward dynamic, season-aware architectures anchored by absolute operational boundaries. --- ### Using LLMs to Detect LLM-Generated Texts: A Cross-Generation Analysis Authors: Haiyue Yuan, Jie Guo, Weidong Qiu, Zheng Huang, Ruizhe Li, Shujun Li Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.34691v1 Automated detection of LLM-generated texts (LGTs) is critical, yet dedicated detectors often struggle to generalize across domains and models. While general-purpose LLMs offer flexible zero-shot authorship classification with explanatory rationale, their detection behavior, especially regarding self-detection versus cross-detection across model generations, remains poorly understood. We systematically evaluate 15 LLMs spanning three model generations as both generators and detectors. Using a benchmark of 1,000 human-written texts and 15,000 LGTs (1,000 per model), we collected over 233,000 binary classifications alongside natural-language explanations. Our results reveal that detection efficacy is primarily driven by detector capability rather than generator provenance, although outputs from newer generators remain notably harder to detect. Crucially, statistical comparisons show no systematic advantage or disadvantage for self-detection across models. Error analysis further exposes generational bias shifts: first-generation detectors under-detect LGTs (high false-negative rates), second-generation detectors over-flag human texts (high false-positive rates), and the latest models achieve balanced trade-offs. Finally, we highlight significant inconsistencies in how different LLMs apply textual cues to justify their decisions. Code: https://github.com/hyyuan/detect-llm-generated-texts. --- ### Unified Trajectory Matching Policy Optimization: Diverse T2I Generation and VLA Generalization Authors: Zhiyuan Ma, Jiaming Li, Lingzhen Li, Yu Liu, Xuekai Zhu, Dingkang Liang, Kaiyan Zhang, Jianjun Li, Bowen Zhou, Xiang Bai Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34688v1 Reward-maximizing reinforcement learning (RL) is widely used to post-train stochastic diffusion and flow policies for text-to-image (T2I) generation. However, reward-maximizing RL causes policy mode collapse even under reference KL or entropy regularization, reducing the policy to a single high-reward mode. In T2I, this produces similar images and reward hacking. When extended to vision-language-action (VLA) models, the same collapse removes alternative successful strategies and weakens task and scene generalization. To address this limitation, we introduce Unified Trajectory Matching Policy Optimization (Uni-TMPO), a unified RL post-training framework for diffusion and flow policies. First, Uni-TMPO converts standardized rewards into a target distribution within each trajectory group and derives the policy distribution from trajectory log probabilities. Then, forward Kullback-Leibler optimization matches the two distributions instead of maximizing expected reward. A progress-conditioned coarse-to-fine scheduler efficiently constructs T2I trajectories. Within the unified framework, feedback-conditioned sampling uses updated observations to construct VLA trajectories. Extensive experiments show that Uni-TMPO achieves higher T2I rewards and VLA ID success rates than the strongest baselines. More importantly, it achieves the best T2I reward-diversity-efficiency trade-off and VLA generalization to held-out tasks and scenes, while real-robot evaluation demonstrates the value of multiple action strategies when the higher-reward target is blocked. --- ### VCN-Bench: A Video-Contextualized Navigation Benchmark for Spatial Reasoning over Prior Visual Experience Authors: Siqi Zhang, Meng Wei, Chenyang Wan, Shaohao Zhu, Shufan Shen, Xihui Liu, Zhihua Wei, Tai Wang, Jiangmiao Pang Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34687v1 Spatial reasoning is fundamental to embodied agents, yet it remains unclear whether spatial understanding can be carried forward to guide sequential interactions. Existing spatial-reasoning benchmarks typically terminate at offline predictions, while navigation benchmarks evaluate spatial reasoning as part of instruction following and exploration. We introduce VCN-Bench, a \textbf{V}ideo-\textbf{C}ontextualized \textbf{N}avigation benchmark for probing closed-loop spatial reasoning over prior visual experience in MLLMs. Given a prior video covering both the initial location and destination, the agent is tasked with reasoning out the instruction-specified target and navigating toward it with the inferred spatial context. Built on Matterport3D, VCN-Bench contains five instruction types, 100k training episodes, and 1,250 evaluation episodes. Navigation serves as the primary evaluation, while diagnostic goal identification helps distinguish destination-resolution errors from subsequent navigation failures. We further propose MV-DualVLN, a planning-oriented baseline that jointly leverages prior video and in-episode observations. Experiments reveal limited navigation performance, a substantial destination-resolution-to-navigation gap, and frequent navigation failures even after correct destination identification. --- ### Jailbreak Context Lingers: Divergent Safety Routing and Its Cross-Task Predictability in Tool Agents Authors: Xi Wang, Songlei Jian, Yiming Zhang, Bin Ji, Zhaoye Li, Ma Jun, Baosheng Wang, Jie Yu Published: 2026-09-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.34686v1 As large language models increasingly operate as tool-using agents, post-jailbreak safety feedback is often assumed to serve as a reliable safeguard; however, how lingering jailbreak context shapes subsequent agent behavior remains largely unexplored. To systematically examine this dynamic, we introduce a paired continuation framework across 192 parent tasks spanning 42 domains, evaluating 12,148 analyzed continuation pairs (curated from a 12,288-pair initially design) across eight diverse agents. We find that identical safety feedback induces sharply model-dependent behavioral routing rather than uniform protection: redirecting unsafe trajectories toward legitimate completion (\emph{rescue}), sustaining unauthorized execution (\emph{persistent unsafe}), or triggering over-refusal on benign tasks (\emph{collateral loss}). Through layer-wise activation patching, we discover a shared \emph{late-commit pattern} where causal intervention effects surge sharply near the final layers (relative depths of 0.958--0.984) despite an over 30-fold variation in peak magnitude across architectures. Crucially, critical-layer representations correlate with macroscopic routing outcomes, and intervening at these layers causally alters concrete next-step tool actions. Building on this causal foundation, we test whether localized intervention-derived features can serve as predictive proxies for full-trajectory routing outcomes on unseen parent tasks under leave-one-parent-task-out evaluation, finding that they provide viable predictive signals in responsive agents with peak ROC AUCs reaching 0.675 for \emph{rescue}, 0.777 for \emph{collateral loss}, and 0.702 for \emph{persistent unsafe}. These findings establish a mechanistic lens and a predictive baseline for anticipating the safety and utility trade-offs of post-jailbreak feedback in autonomous agents. --- ### Natural State-Prediction Accuracy can Hide Weak Controlled Responsiveness in VLA Readouts Authors: Hyungjoon Kim, Wonbin Son, Mi Young Lee, Jun Young Lee, Seungmin Rho Published: 2026-09-28 Categories: cs.RO, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.34684v1 Accurately decoding object states from the internal representations of vision-language-action (VLA) models does not establish that the predictions respond faithfully to changes in the target physical state. In natural observations, object state, robot configuration, occlusion, and task progress vary together, allowing contextual cues to contribute to prediction. In this paper, we introduce an evaluation framework that separates prediction accuracy, target-state responsiveness, and context stability using physically validated observations that cross target coordinates with robot contexts. We demonstrate that high natural-trajectory accuracy can coexist with weak controlled target-state responsiveness in fixed representation-readout pairs. Comparisons and interventions involving representations, readouts, and training data show that the three properties provide distinct diagnostic information. Furthermore, adding responsiveness and context sensitivity to a failure predictor based on initial state error and physical variables reduces policy-failure prediction error on new initializations relative to the specified baseline while same-observation controlled MAE is also informative. These findings motivate evaluating target-state responsiveness and context stability alongside natural prediction accuracy, and examining their relationship to actual policy behavior and task outcomes. --- ### AgentPerfBench: A Benchmarking and Evaluation Suite for Inference Performance of Agentic LLMs Authors: Cheuk Hang Lau, Zeyu Cao, Kevin Wong Cheuk Yin, Yao Lai, Haoran Wu, Nicholas D. Lane, Robert D. Mullins, Ilia Shumailov, Yiren Zhao Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.34683v1 The optimization of LLM serving engines, such as vLLM and SGLang, is largely benchmark-driven: optimizations, scheduling policies, hardware and system designs are all selected based on representative workloads. However, a significant mismatch has emerged in the agentic era. Existing benchmarks primarily focus on simple single-turn chatbot workloads. LLM applications are increasingly agentic: coding agents, terminal execution systems, and tool-use agents issue multi-turn requests with growing context lengths. We introduce AgentPerfBench, a benchmark suite for agentic inference. It uses real traces from agentic benchmarks, such as SWE-Bench and TerminalBench, alongside standard chat baselines. This enables benchmarking of models on multi-turn tasks involving tool calling, skill utilization, and increasing context lengths. AgentPerfBench also samples from empirical distributions of input length, output length, and turn count derived from the real traces, generating representative synthetic profiles for cheap and accurate measurements on new hardware. In addition, we further find that several existing benchmarks fail to accurately reflect real hardware performance for two key reasons: 1) they do not account for realistic context-length growth, and 2) they measure inference performance without operating at hardware saturation. We discuss these issues in detail and provide rich kernel-level Nsight Compute (NCU) traces to construct a new multi-dimensional roofline model that captures hardware-system limitations in both memory bandwidth and memory capacity footprint. The benchmarking suite then includes automated scripts to identify potential bottleneck conditions on emerging hardware when evaluated with diverse agentic traces. Together, these contributions quantify the chat-to-agentic gap in current inference benchmarks and characterise per-kernel GPU resource utilisation via roofline analysis. --- ### V-Gym: Enhancing Agentic Visual Reasoning via Skill-Data Co-Evolution Authors: Bei Yan, Yuecong Min, Jie Zhang, Junqi Yang, Shiguang Shan, Xilin Chen Published: 2026-09-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.34682v1 Advances in multimodal understanding, reasoning, and tool use enable agents to tackle increasingly complex visual reasoning tasks. By distilling past execution experience into reusable skills, agents can transfer lessons from both successes and failures into future reasoning, reducing repeated errors and improving capabilities. However, limited experience may produce unreliable, poorly generalizable skills, while static datasets may lack the targeted and diverse practice needed for refinement. To address this gap, we introduce V-Gym, an autonomous framework that iteratively co-evolves procedural skills and multimodal practice data from execution trajectories. During skill evolution, V-Gym analyzes trajectories to distill and refine hierarchical skills, updating procedural guidance and applicability conditions while retaining an update only if it improves validation performance. During data evolution, V-Gym selects generation seeds by balancing data utility and exploration, then translates trajectory-identified bottlenecks into diverse, targeted practice data that expand the data bank after quality checks. The resulting practice outcomes feed back into subsequent skill updates, closing the loop for continual skill refinement. Experiments across diverse multimodal reasoning benchmarks show substantial improvements over baselines with multiple backbone models. Its evolved skills generalize across domains and models, while evolved data support more effective skill refinement, enabling autonomous diagnosis, targeted practice, and continual self-improvement. --- ### SOLAR: A State-Driven Online Learning Rate Scheduler for LLM Pretraining Authors: Qiulin Shang, Binyu Wang, Yongqi Qiao, Songde Rao, Zhoutong Wu, Kun Yuan Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.34681v1 Learning-rate (LR) scheduling plays a central role in large language model (LLM) pretraining, yet current practice still relies heavily on hand-crafted heuristics such as Warmup-Cosine-Decay and Warmup-Stable-Decay. Because these schedules are fixed in advance, they cannot adapt to evolving optimization dynamics. Online learned scheduling within the Learning to Optimize (L2O) framework offers a dynamic alternative, but remains brittle at LLM scale due to noisy signals, delayed feedback, and the risk of catastrophic divergence. We propose SOLAR (State-driven Online Learning rAte scheduleR), a stabilized framework for reliable online LR adaptation. SOLAR uses a base schedule as a reference and learns bounded, state-dependent residual corrections for individual parameter groups. Each correction re-anchors to the base at every step, allowing the policy to adapt the LR without relearning the warmup-decay profile. A lightweight state representation and progress-aware reward guide online learning, while a Circuit-Breaker restores training after rare unsafe actions. Across autoregressive language-model pretraining, SOLAR improves final perplexity over tuned static schedules and automatic LR tuners for dense models from 60M to 1B, AdamW and Muon, and two MoE settings up to 3B. Matched 130M controls show that adding base anchoring and action bounds improves a global PPO controller from 27.09 to 23.74 final PPL, while group-wise control reaches 22.87 on the same two seeds. A residual policy trained on a 60M proxy can also be frozen and reused at larger dense scales without target PPO updates, remaining effective across a fourfold base-LR range. These results establish SOLAR as a practical learned LR controller for LLM pretraining. --- ### QuantForge: Discovering Residual Decompositions for MXFP4 Post-Training Quantization Authors: Qiulin Shang, Zhoutong Wu, Jie Hu, Kun Yuan Published: 2026-09-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.34680v1 Four-bit post-training quantization can reduce the memory demands of large language models, but preserving accuracy under strict MXFP4 W4A4 requires coordinating several design choices. Coordinate transforms change block-encoding errors, which in turn affect the residuals propagated through the network. The useful algorithmic decomposition is therefore not fully known before search. LLM-driven program evolution offers a way to explore these choices, but performance scores alone do not explain which design should change next. We introduce QuantForge, a PTQ discovery system that records competing explanations, selects controls that distinguish them, and checks that successor code implements the resulting conclusions. This residual compilation guides program revisions while retaining useful programs even when their original explanations are rejected. Remeasuring the revised program reveals the next error to address. This process discovers HiRes, a fixed MXFP4 quantizer that shapes coordinates, refines legal code assignments, and recovers errors along attention and MLP paths. Each stage acts on residuals measured after the preceding stage has executed. Across seven tasks, HiRes achieves the lowest seven-model Robust Fit (0.09300) and the lowest quantized Fit-7 at 32B. In matched-budget comparisons of LLM-driven program evolution, each with 240 evaluator calls, QuantForge reaches a held-out transfer target in six of eight runs, compared with three each for textual memory and reflection memory, and one for score-only evolution, despite evaluating fewer new programs. These results show that QuantForge improves the discovery of transferable PTQ algorithms by turning controlled evidence into subsequent program changes. --- ### From Preference to Reciprocity: Decentralized Matching with Empirically Grounded LLM-agent Based Modeling Authors: Wangxuan Fan, Xiaoyu Nie, Zhoutian Shi, Xiangcheng Meng, Shipei Zeng, Pin Gao, Yan Hu, Zhongxiang Dai Published: 2026-09-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.34679v1 Bipartite matching is a fundamental problem in game theory and market design. Classical approaches such as Gale--Shapley assume complete preferences and centralized computation, whereas many real-world matching processes are decentralized, asynchronous, and shaped by sequential interaction under limited information. We propose a dynamic bipartite matching framework that combines large language model (LLM) agents with contextual bandits. In a simulated Chinese marriage market, economically grounded LLM agents evaluate locally encountered candidates, while agent-specific Logistic-UCB models learn reciprocal acceptance from realized proposal outcomes. The mechanism therefore separates two decisions---\emph{whom do I like?} and \emph{who is likely to like me back?}---without requiring ex ante market-wide preference rankings. We first validate LLM-induced mate preferences against the empirical conditional-logit reference across multiple LLM backbones. In the $50\times50$ matching experiment, Bandit-UCB achieves the highest mean mutual welfare (56.01 versus 54.87 for Gale--Shapley), a smaller gender rank gap than the classical baselines, and the fewest blocking pairs among the LLM-ABM policies. Learned acceptance models show economically interpretable gender-differentiated associations, while counterfactual setups reveal no systematic unilateral advantage from prior search knowledge. Overall, these results support the advantages of decentralized matching with LLM-based behavioral modeling and online learning under incomplete information for economic simulation and computational social science research. --- ### Fair Fact-Checking: Closing the Cross-Lingual Gap in LLM Factual Judgement with RoSh Authors: Muhammad Ahmad, Fatemeh Seyedin, Adrian Weller, Dongwon Lee, Mahmoudreza Babaei Published: 2026-09-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2609.34678v1 Misinformation on social media remains a critical problem, and more and more people settle it by asking a language model instead of a fact checker. Whether models judge such claims reliably is debated; whether they judge them equally well in every language people ask in has gone almost unasked. We test eight models from five families, 3B to 70B, on 1,500 encyclopedic factual claims that exist in identical form in eight languages. English is judged better than every other language on every model, and the gap is widest on the smallest ones, where Llama-3B on Arabic is no better than guessing. Existing remedies retrain on more multilingual data or fit an unconstrained map between language representations, and neither asks whether the model already holds the answer and simply fails to say it. It largely does: a linear probe recovers the truth from the very activations the model fails to express. We propose RoSh, a per-language shift and rotation of the residual stream, computed in closed form at three layers, with no training and no weight modified. It improves every model and closes 75% of the gap on average, helping most where the model was worst: Arabic on Llama-3B goes from chance to nearly the English level, and a fifth fewer of the claims answered correctly in English are lost in translation. What remains is no longer a read-out failure: afterwards the head recovers as much of what is encoded outside English as it does in English. An unconstrained map fitted on the same pairs falls below the untouched baseline, so the orthogonality constraint is doing the work, and every model clears a scrambled-correspondence control and ten further controls. On the two benchmarks of the closest inference-time method, latent-space intervention, run with its own data and metric code, RoSh's gains are five to thirteen times larger. --- ### Learning What to Recall: Adaptive Multi-Cue Episodic Memory for World Models Authors: Beomsu Kim, Chieh-Hsin Lai, Bac Nguyen, Amir Bar, Jong Chul Ye, Yuki Mitsufuji Published: 2026-09-28 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2609.34677v1 World models predict future observations from current experience and actions, yet prediction can depend on observations seen far in the past. Episodic memory preserves past observations for later recall; however, as memory accumulates, it raises a fundamental question: which memories are useful for the current prediction, and which available retrieval cues should be trusted to find them? This is challenging because fixed criteria based on recency, pose overlap, or visual similarity can be unreliable across environments and queries. We propose Future-Aware Recall (FAR), a framework that learns episodic recall from future-aware predictive supervision and adaptive multi-cue scoring. During training, FAR measures predictive utility by the conditional log-likelihood of the realized future given recalled context, approximated by negative diffusion prediction loss, and uses it to train a retriever that remains future-blind at inference. The retriever learns cue-specific relevance and automatically determines which available retrieval cues, such as time, pose, vision, and audio, to trust for each query when selecting memories. Across three complementary settings, FAR outperforms hand-designed recall even with the same retrieval cues, automatically adapts which available cues to trust, and recalls the right history as the world changes. Together, these results establish FAR as a flexible, principled approach to episodic memory access in world models. ---