# Research Papers — 2026-07 · bundle page 152 of 222 > Full-text markdown of papers #7551–7600 of 11075 published in 2026-07, newest first. AI-readable mirrors of the full 11075-paper 2026-07 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-07/page/152 [← Previous bundle](https://aipod.dev/papers/archive/2026-07/page/151.md) [Next bundle →](https://aipod.dev/papers/archive/2026-07/page/153.md) ### Ceci n'est pas une pipe: AI systems as semantic abstractions Authors: Jade Alglave, Patrick Cousot Published: 2026-07-10 Categories: cs.AI, cs.PL arXiv: https://arxiv.org/abs/2607.09489v1 An AI system's output is not the fact or world state it appears to describe, but rather an engineered representation. We propose a semantic framework to describe AI systems, to be able to examine the correctness of such representations. To do so, we distinguish what is justified by accepted domain knowledge, what reference sources say, and what the system can currently use. This allows us to give precise definitions to common failures: extrapolation, refuted or unsupported assertion, sources versus knowledge mismatch, stale or refuted source, added hypotheses, unsupported use... We hope our framework gives a useful vocabulary for specifying and checking AI systems whose outputs, citations, tool calls, and world-changing actions must be justified by reliable claims and explicit authority rather than apparent fluency. --- ### SigLIP-HD by Fine-to-Coarse Supervision Authors: Lihe Yang, Zhen Zhao, Hengshuang Zhao Published: 2026-07-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.09488v1 High-quality visual representation is a long-standing pursuit in computer vision. In the context of multimodal LLMs (MLLMs), feeding higher-resolution images can produce more fine-grained visual tokens. However, it introduces additional computational and design complexity, due to multiple forward passes and post-processing of increased tokens. Before simply adopting a higher resolution, have we truly unlocked the model's full perception capability at a standard resolution? Therefore, we study an interesting problem: how to achieve fine visual perception under lower cost without larger images. We present SigLIP-HD in this work. The core is a highly simple fine-to-coarse supervision design. We enforce the coarse feature of a mid-resolution image to mimic the fine-grained feature of its high-resolution version. We build this framework on the advanced SigLIP 2 model. Our final model produces better visual tokens at exactly the same inference budget. It is validated on extensive MLLM benchmarks and consistently delivers stronger results than our baseline model, especially on OCR-related tasks. --- ### Neural Collapse Is Forbidden: Information Floors in Language Models Authors: Bruno Abrahao Published: 2026-07-10 Categories: cs.LG, cs.CL, stat.ML arXiv: https://arxiv.org/abs/2607.09487v1 Within-class variance in language-model representations is commonly read as incomplete neural collapse. We argue it is allocated information storage, and that the allocation obeys a law. A one-line centering identity voids a family of simplex equiangular-tight-frame claims, including our own earlier ones; in dimensionless variance shares across 14 models, macro-category structure carries only 4-12% of representational variance and within-token context carries 79-91%, stable across a 100x parameter range. On the theory side, token-level weight decay penalizes a category in proportion to its type count, not its occurrence mass, reducing next-token prediction to an imbalanced K-class problem whose optimum orders category norms by type count. A converse floor, proved for binary categories, forces within-category dispersion to be at least proportional to the conditional mutual information I(token; context | category). The law holds: identity dispersion, not total variance, tracks this information across every tested model and partition, under a model-free estimate and even across models, where one model's information predicts another's dispersion; and over pretraining the category share overshoots, decays, and partially recovers, because the information it must carry never left. --- ### Decoupling Language Guidance from Backbones for Text-Guided Medical Segmentation Authors: Yungeng Liu, Xuanzi Fang, Haijin Zeng, Qi Dai, Yongyong Chen Published: 2026-07-10 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.09481v1 Text-guided medical image segmentation leverages clinical semantics to improve lesion delineation, yet many existing models bind cross-modal fusion, supervision, and decoder design into a task-specific architecture. Such tight coupling makes it difficult to reuse language guidance modules across heterogeneous vision and text backbones, and often requires redesigning the network when the encoder pair changes. This paper presents BTHA, a backbone-transferable hierarchical adapter framework for text-guided medical image segmentation. BTHA is built around a stable feature-level interface: given multi-scale visual features and a text representation, it injects semantic guidance through shape-preserving adapters while maintaining the decoder-side tensor contract. To make this interface effective, we introduce a Hierarchical Coarse-to-Fine Supervision Strategy that decomposes learning into global image-text alignment, multi-scale auxiliary localization, and boundary-aware final mask refinement. We further design a Scale-Adaptive Gated Semantic Guidance (SAGSG) adapter, where resolution-specific gates adaptively control textual injection and channel recalibration suppresses redundant cross-modal responses. Evaluations across diverse vision and text backbones show that the same adapter and supervision design remains effective across convolutional and transformer-based visual encoders as well as different language encoders. Experiments on four public datasets further demonstrate that BTHA improves strong text-guided baselines with modest computational overhead. --- ### Foveation-Guided Dynamic Token Selection for Robust and Efficient Vision Transformers Authors: Ibrahim Batuhan Akkaya, Kishaan Jeeveswaran, Bahram Zonooz, Elahe Arani Published: 2026-07-10 Categories: cs.CV, cs.LG, cs.NE arXiv: https://arxiv.org/abs/2607.09480v1 The human visual system (HVS) employs foveated sampling and eye movements to achieve efficient perception, conserving both metabolic energy and computational resources. Drawing inspiration from this robustness and adaptability, we introduce the Foveated Dynamic Transformer (FDT), a foveation-guided dynamic token-selection architecture that integrates these mechanisms into a vision transformer framework. The FDT exhibits strong resilience to various types of noise and adversarial attacks, despite not being explicitly trained for such challenges. This inherent robustness is achieved through the use of fixation and foveation modules: the fixation module identifies fixation points to filter out irrelevant information, while the foveation module generates foveated embeddings with multi-scale information. At the 50% fixation-budget setting, FDT achieves higher accuracy than DeiT-S (81.9% vs. 80.9%) while reducing multiply-accumulate operations by 34.57%, highlighting one operating point on its accuracy-efficiency trade-off. These attributes position FDT as an HVS-inspired step toward artificial neural networks that combine adaptive computation with improved resilience. --- ### ProofCouncil: An LLM Agent for Solving Open Mathematical Problems Authors: Johannes Schmitt, Tim Gehrunger, Jasper Dekoninck, Gergely Bérczi, Uri Kreitner, Liam Price, David Holmes Published: 2026-07-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.09474v1 Large language models (LLMs) have shown increasing promise in solving open problems in mathematics. However, their performance can be further improved through agentic workflows tailored to real-world mathematical practice. To this end, we introduce ProofCouncil, a mathematical agent that is designed to tackle open problems using an author-critic architecture. ProofCouncil served as a submission to the second batch of FirstProof, a challenge consisting of 10 real-world mathematical problems that agents must solve autonomously. Its submissions for 6 of the 10 problems were judged by the referees to be correct up to at most minor revisions, showing the best performance among participating teams. We also evaluate ProofCouncil on 30 open problems collected from mathematical researchers. Among the 21 solutions that received human feedback, 5 were judged completely correct, 2 more were judged promising pending final verification, and a further 8 contained useful partial progress. In this short paper, we describe the development of ProofCouncil and the agent-building library used to create it, which we release as open source to the community. --- ### Deep Learning for Dynamic Programming with Recursive Utility Using First-order Conditions Authors: Xianhua Peng, Wu Guo, Songyan Wang, Jianfei Zhu Published: 2026-07-10 Categories: q-fin.CP, econ.EM, math.OC, stat.ML arXiv: https://arxiv.org/abs/2607.09461v1 This paper proposes the certainty-equivalent first-order learning (CEFOL) algorithm, a deep learning algorithm for solving discrete-time dynamic programming problems with recursive utility. Dynamic programming with recursive utility is challenging because nonlinear certainty equivalent appears in the Bellman equation and the first-order optimality conditions but is difficult to evaluate. By introducing a separate neural network to represent the certainty equivalent, CEFOL enables the exploitation of the Bellman and model-specific first-order optimality conditions. In addition to certainty equivalent, CEFOL also uses neural networks to learn the value functions, policy functions, and Lagrange multipliers by using model-specific first-order conditions to construct residuals for minimization. By using first-order and KKT residuals to learn the policy, CEFOL directly accommodates general equality and inequality constraints on the controls, including occasionally binding constraints, without requiring penalty functions or problem-specific reformulations. We apply the algorithm to risk-sensitive and Epstein--Zin consumption-saving problems, a small-noise robust-control problem, and a DSGE model with recursive preferences and stochastic volatility. Across these applications, out-of-sample Bellman diagnostics and model-specific optimality residuals, including Euler or first-order residuals where applicable, are generally of order 1.0e-4 to 1.0e-3 over the relevant state regions, with larger values mainly near binding constraints, and the learned value and policy functions closely match VFI benchmarks when available. The CEFOL algorithm also works for dynamic programming problems with expected utility, as expected utility is a special case of recursive utility. --- ### Active rejection enables reliable generalization of universal machine-learning interatomic potentials Authors: Mingxiang Luo, Xinnan Mao, Lu Wang, Lei Bai, Feng Ding, Yuqiang Li Published: 2026-07-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.09456v1 Universal machine learning interatomic potentials (uMLIPs) bridge quantum-mechanical accuracy and large-scale molecular dynamics, but the cost of high-accuracy calculations such as r$^2$SCAN limits training to datasets that remain small relative to the open materials space. Strong average benchmark performance also does not guarantee reliable energy--force predictions for every structure. We propose Adaptive Multi-Teacher Routing (ATR), which reformulates high-fidelity data construction as a structure-wise decision problem under uncertainty. Using a small set of real r$^2$SCAN labels, ATR calibrates multiple pretrained uMLIP teachers and combines structural descriptors, teacher identity, and inter-teacher disagreement to estimate the reliability of each structure--teacher pair. It selects high-confidence predictions for pseudo-label generation and rejects structures for which no teacher is sufficiently reliable. With real r$^2$SCAN labels for only 0.2\% of candidate structures, ATR distils 2.89 million traceable r$^2$SCAN-level pseudo-labels for pretraining. On held-out r$^2$SCAN structures and the MP-r$^2$SCAN benchmark, a lightweight CHGNet trained on the ATR-generated dataset consistently outperforms the baseline and non-routed controls. Finite-temperature molecular dynamics further shows that ATR improves dynamical robustness across multiple material systems, maintaining stable trajectories where baseline simulations undergo catastrophic structural collapse. These results establish active rejection as an effective mechanism for converting multiple pretrained uMLIPs into a scalable and reliable data-construction system for high-fidelity uMLIPs. --- ### Hydra++: Real-Time Hierarchical 3D Scene Graph Construction With Object-Level Shape Estimation Authors: Hyungtae Lim, Nathan Hughes, Xihang Yu, Ruihan Xu, Yun Chang, Jingnan Shi, Rajat Talak, Luca Carlone Published: 2026-07-10 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2607.09455v1 3D scene graphs provide a hierarchical abstraction of environments by encoding spatial entities, such as objects and places, and their relationships. However, existing scene graph systems model object geometry coarsely, relying on partial point clouds or class-level CAD templates, which limits instance-specific shape detail. This paper presents Hydra++, a system-level investigation into how learning-based object shape estimators can be integrated into a hierarchical 3D scene graph pipeline. Hydra++ incorporates category-agnostic shape estimation and a reprojection-mask consistency check to reject degenerate predictions from partial observations or imprecise segmentation. In its default CRISP-based configuration, Hydra++ performs online scene graph construction; slower estimators such as SAM3D are evaluated as modular alternatives to demonstrate generalization-latency trade-offs. Furthermore, to address the challenges of sparse and noisy depth measurements in outdoor environments, Hydra++ supports a hybrid LiDAR-camera configuration for large-scale operation, improving scene-level reconstruction quality. Experiments in both simulation and real-world outdoor campus scenarios demonstrate that Hydra++ improves object- and scene-level reconstruction quality. Project page is available at https://hydra-plusplus.github.io/. --- ### Practical Source Code Recovery from Binary Functions Using Anchor-Based Retrieval and LLM Reasoning Authors: Charles Edward Gagnon, Steven H. H. Ding, Philippe Charland, Benjamin C. M. Fung Published: 2026-07-10 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2607.09452v1 We present a practical pipeline for recovering source code from stripped binary functions by combining reverse engineering, anchor-based source code retrieval, and large language model reasoning. Our binary-to-source-code retrieval method attempts to identify the source function from a source code database, rather than generating approximate decompiled pseudocode. It extracts anchors such as strings, constants, external calls, and available function names using Ghidra, retrieves candidate files via an inverted-index search database, narrows candidates to likely function snippets, and re-ranks them with a large language model (LLM) based on disassembly, decompiled code, and source metadata. Confident matches can also serve as anchors in later passes. In an evaluation backed by our high-fidelity source code database on a stripped, optimized tcpdump binary, our proposed binary-to-source matching method achieves 95.2% assembly instruction coverage. Experiments on a GitHub-based retrieval database showed lower performance with 35.5% instruction coverage on average, mainly due to retrieval misses. These results show that source-level binary recovery excels with high-quality databases and remains a useful tool in noisy environments. --- ### Robustifying Vision-Language Models via Test-Time Prompt Adaptation Authors: Xingyu Zhu, Huanshen Wu, Shuo Wang, Beier Zhu, Jiannan Ge, Jiaheng Zhang, Long Chen Published: 2026-07-10 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2607.09450v1 Pre-trained Vision-Language Models (VLMs) such as CLIP achieve strong zero-shot generalization, but their performance degrades sharply under adversarial perturbations. Existing test-time adaptation methods typically rely on sample-level confidence heuristics, overlooking the intrinsic distributional structure of the data. This sample-centric approach limits robustness, as it fails to distinguish confident adversarial mispredictions from true semantic consistency. In this work, we observe that adversarial distortion is structurally brittle: while holistic representations are corrupted, semantic integrity is often preserved in the distribution of augmented views. Motivated by this insight, we propose RITA, a Robust test-tIme prompt-TAdaptation framework that shifts from sample-level estimates to distribution-level alignment. Specifically, RITA employs optimal transport to align the distribution of augmented visual features with textual prototypes, mitigating adversarial outliers and rectifying cross-modal semantic misalignment. Furthermore, we introduce a dynamic cache to progressively accumulate reliable cues from the test stream for online refinement. Extensive experiments demonstrate that RITA significantly improves adversarial robustness without compromising clean accuracy. --- ### How Does Bayesian Causal Discovery Fail? Characterising Structural Consequences in Linear Gaussian Networks under Latent Confounding Authors: Debargha Ghosh, Silja Renooij, Anna V. Kononova Published: 2026-07-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.09449v2 Bayesian causal discovery is widely used for its ability to quantify epistemic uncertainty over directed acyclic graphs (DAGs) through posterior inference. However, its behaviour under latent confounding remains poorly understood, as existing work typically notes that confounding breaks identifiability without characterising how the posterior distribution over DAGs responds. In this work, we analyse posterior behaviour under latent confounding in linear Gaussian causal models, focusing on additive latent confounding between exactly two observed variables. We derive a critical correlation threshold above which the score function favours graphs with a spurious edge between the confounded variables, and show that this threshold decreases with sample size -- more data lowers the correlation required for the spurious edge to be favoured. Beyond this threshold, we characterize two distinct posterior failure regimes determined by the local structure around the confounded variables. Our findings are supported by exact posterior computations on multiple graph structures, demonstrating both the predicted failure regimes. --- ### Parameter-Efficient Vision-Language Adaptation with Continuous Metadata Conditioning for Animal Re-Identification Authors: Anil Osman Tur, Tonje Knutsen Sordalen, Kim Tallaksen Halvorsen, Cigdem Beyan Published: 2026-07-10 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.09443v1 Long-term animal re-identification (ReID) must remain robust to gradual morphological evolution and seasonal appearance shifts. Although recent vision-language models provide strong pretrained visual representations, adapting them to longitudinal ecological settings remains challenging, particularly under identity and temporal distribution shifts. We present a parameter-efficient CLIP adaptation framework for animal ReID and introduce a continuous metadata-conditioning mechanism that incorporates numerical attributes directly into the prompt representation during training. While low-rank visual adaptation, prompt-based supervision, and cross-modal alignment provide the adaptation framework, the proposed metadata-conditioning strategy constitutes the primary methodological contribution. By preserving the continuous structure of numerical metadata rather than discretizing it into textual categories, the proposed approach enables smooth modulation of the embedding space during training while maintaining a purely visual inference pipeline. Experiments on a seven-year longitudinal fish dataset and multiple wildlife benchmarks demonstrate improved performance under closed-set, open-set, and time-aware evaluation protocols. The results demonstrate that continuous metadata conditioning improves robustness to longitudinal appearance variation and temporal distribution shifts, while parameter-efficient adaptation enables a purely visual inference pipeline without requiring metadata at test time. Code and evaluation splits can be found at: https://github.com/AnilOsmanTur/MetaPrompt-ReID. --- ### Near-optimal node-private community estimation in polynomial-time Authors: Laurentiu Marchis, Olga Klopp, Po-Ling Loh, Ilias Zadik Published: 2026-07-10 Categories: math.ST, stat.ML arXiv: https://arxiv.org/abs/2607.09441v1 In this paper, we resolve an open question of Klopp & Zadik (2026) by providing a high-probability polynomial-time, node-private algorithm which nearly matches the performance of their exponential-time node-private algorithm for exact recovery in stochastic block models. Our result involves an explicitly constructed Lipschitz surrogate for the penalized likelihood function, as well as a carefully devised accept-reject algorithm that samples community labels from the corresponding exponential mechanism in polynomial-time. We rigorously analyze the privacy, runtime, and utility of our proposed algorithm, showing that even when the number of communities K grows logarithmically with the number of nodes n, we can achieve the minimax rates for exact recovery with the privacy parameter epsilon growing as log(n), thus matching known lower bounds on the cost of privacy for this setting. --- ### Test-Time Scaling for Small VLMs on Multilingual Visual MCQ Authors: Spiros Baxevanakis, Peng-Jian Yang Published: 2026-07-10 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2607.09438v1 Test-time scaling (TTS) reliably improves reasoning in large language models, but whether it transfers to small open vision-language models remains unclear. We examine this on EXAMS-V, a multilingual visual multiple-choice benchmark, comparing self-consistency, describe-then-reason with PRM-guided beam search, and two post-hoc selectors across Qwen2.5-VL-7B-Instruct and Qwen3.5-4B. What matters is the conditions under which TTS runs, not the search or verification machinery. The largest factor is parseability: an early prompt format left many chains reasoning correctly yet never committing to an answer letter, which a standard answer cue and a guided repair step largely remove. A larger decoding budget removes the rest: raising the per-chain token limit from 1k to 2k recovers 3.7 pp, whereas sampling more chains (8 to 16) adds only 0.15 pp. Once chains have room to finish, elaborate methods contribute little: PRM-guided beam search trails plain self-consistency by 0.39 pp at over eight times the cost, and neither a training-free generative critic nor a trained multimodal PRM beats majority vote across both policies. The largest gain comes instead from the policy model itself (+11.4 pp). Our best configuration reaches 84.1% on the held-out ImageCLEF 2026 test split, ranking first on the Visual MCQ leaderboard. --- ### comprisk: A scikit-learn-compatible Python toolkit for competing-risks survival analysis Authors: Sunny Yang, Weiyan Zhao, Wanqi Zhao Published: 2026-07-10 Categories: stat.CO, stat.ML arXiv: https://arxiv.org/abs/2607.09431v2 Medical time-to-event data are frequently subject to competing risks, where the occurrence of one terminal event precludes the others and standard survival methods that treat competing events as censoring yield biased absolute-risk estimates. Valid analysis instead targets the cause-specific cumulative incidence function (CIF). This methodology has been available to applied researchers almost exclusively through R packages, forcing Python-based machine-learning workflows into a Python-to-R round trip. We present comprisk, a scikit-learn-compatible Python toolkit that puts the canonical competing-risks methods behind one API: a scalable competing-risks random survival forest, Fine-Gray subdistribution-hazard regression and a penalized variant, cause-specific Cox regression, the Aalen-Johansen CIF estimator, and Gray's K-sample test, together with competing-risks-aware model evaluation. Every estimator is validated numerically against its R reference implementation. The forest uses a histogram-based, numba-compiled split kernel that fits 10-22x faster than randomForestSRC at comparable discrimination on real clinical cohorts and scales to n = 10^6 on a consumer CPU. comprisk is distributed on PyPI and lets applied researchers run correct, scalable competing-risks analysis without leaving the Python scientific stack. --- ### comprisk: A scikit-learn-compatible Python toolkit for competing-risks survival analysis Authors: Sunny Yang, Weiyan Zhao, Wanqi Zhao Published: 2026-07-10 Categories: stat.CO, stat.ML arXiv: https://arxiv.org/abs/2607.09431v1 Medical time-to-event data are frequently subject to competing risks, where the occurrence of one terminal event precludes the others and standard survival methods that treat competing events as censoring yield biased absolute-risk estimates. Correct analysis instead targets the cause-specific cumulative incidence function (CIF). This methodology has been available to applied researchers almost exclusively through R packages, forcing Python-based machine-learning workflows into a Python-to-R round trip. We present comprisk, a scikit-learn-compatible Python toolkit that consolidates the canonical competing-risks methods (a scalable competing-risks random survival forest together with Fine-Gray subdistribution-hazard regression including a penalized variant, cause-specific Cox regression, the Aalen-Johansen CIF estimator, and Gray's K-sample test) behind a single, consistent API, and adds competing-risks-aware model evaluation (inverse probability of censoring weighted time-dependent AUC and Brier score, cause-specific concordance indices with closed-form confidence intervals, and calibration curves). Every estimator is validated numerically against the established R reference implementations. The forest uses a histogram-based, numba-compiled split kernel that fits 10-22x faster than randomForestSRC at comparable discrimination on real electronic-health-record cohorts and scales to n = 10^6 on a consumer CPU. comprisk is distributed on PyPI and lets applied researchers perform correct and scalable competing-risks analysis entirely within the Python scientific stack. --- ### Multimodal Scenario Similarity Search for Autonomous Driving Authors: Tamás Matuszka, András Tamásy, Balázs Szolár Published: 2026-07-10 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2607.09428v1 Large-scale autonomous-driving datasets contain vast numbers of recorded scenarios, creating a need for efficient retrieval methods that can identify situations similar to a given query. Existing approaches typically rely on either visual representations or motion-based descriptions, making it difficult to understand their relative strengths and limitations for scenario retrieval. In this work, we present a multimodal framework for autonomous-driving scenario retrieval that combines visual and trajectory-based representations within a unified retrieval pipeline. We investigate two trajectory-based approaches: Exo-Trajectory, an explicit matching method based on surrounding-agent motion, and ScenarioFormer, a transformer-based representation learned from object trajectories using contrastive learning. We compare these approaches against strong vision-based baselines and analyze their behavior across a diverse set of driving scenarios. Experimental results show that trajectory representations provide strong retrieval performance for motion-centric events such as cut-ins, turning maneuvers, and traffic queueing, while visual embeddings excel when appearance cues are informative. Most importantly, combining visual and trajectory information consistently improves retrieval quality, yielding the best overall performance. These findings demonstrate that appearance and motion capture are complementary notions of scenario similarity and motivate multimodal retrieval systems for autonomous-driving data mining, dataset curation, and scenario-based validation. --- ### A Sovereign, Open-Source Foundation Model for German and English Authors: Soofi-Team, :, Benedikt Droste, David Fitzek, Ruben Härle, Lukas Helff, Maximilian Idahl, Alex Jude, Abbas Goher Khan, Maurice Kraus, Timm Ruland, Richard Rutmann et al. Published: 2026-07-10 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2607.09424v3 We present Soofi S 30B-A3B, a sovereign, open-source Mixture-of-Experts (MoE) hybrid Mamba Transformer foundation model for German and English. Its hybrid design activates only 3B of 30B parameters per token and keeps the inference cache near-constant as context grows, giving it a decisive throughput advantage over dense models for long-context, high-concurrency deployment. Pretrained on roughly 27 trillion tokens with deliberately up-weighted German, Soofi S matches dense 14 to 27B models on aggregate English and German benchmarks while achieving the best code aggregates in both languages among 17 open base models, and outperforms every European sovereign baseline in our comparison, including ones far larger in active parameters. Among fully open models, Soofi S obtains the highest English and German evaluation scores, ahead of Olmo 3 32B and Apertus 70B. Soofi S was built end-to-end on the German Industrial AI Cloud, a sovereign HPC scale AI infrastructure operated by Deutsche Telekom in Munich. Soofi S will be released under highly permissive, open-access terms: weights, selected intermediate checkpoints, full per-source data accounting, hyperparameters, and training and evaluation code. Where source licenses permit, data-construction artifacts are released under permissive licenses; commercially licensed sources are documented with aggregate statistics and exact mixture accounting. --- ### Action-Factored Multi-Agent Reinforcement Learning for Scalable Quantum Device Tuning Authors: Edwin De Nicolo, Rahul Marchand, Cornelius Carlsson, Pranav Vaidhyanathan, Natalia Ares Published: 2026-07-10 Categories: cs.LG, cond-mat.mes-hall arXiv: https://arxiv.org/abs/2607.09422v1 Cooperative multi-agent reinforcement learning is well suited to problems with large parameter spaces and exploitable local structure, such as the tuning of electrostatically-defined quantum-dot arrays. However, if parameter cross-talk is strong, a non-stationary environment from the perspective of any individual agent can destabilize learning - the same effect that plagues manual tuning of such systems. We propose using a factored representation of the action space, learned online, to decouple agents and minimize their interference. Our framework, QADAPT, uses this factorization to efficiently learn shared policies based on local measurements and rewards. With this modular strategy, we achieve zero-shot generalization to unseen quantum device sizes and maintain an approximately constant number of convergence steps to reach target regimes. This work provides a scalable route toward the rapid calibration of large-scale quantum processors. --- ### SVF-CR: Synchronized Visual-Facial Cross-Refinement for Multimodal Ambivalence and Hesitancy Recognition Authors: Hyein Park, Namho Kim, Junhwa Kim Published: 2026-07-10 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.09417v1 Ambivalence and hesitancy are subtle behavioral states that are expressed through a combination of verbal content, facial behavior, visual context, and acoustic cues. Effective recognition therefore requires not only extracting informative unimodal representations, but also modeling how temporally aligned behavioral evidence interacts across modalities. In this paper, we propose a synchronized visual-facial cross-refinement framework (SVF-CR) with pairwise multimodal evidence fusion for ambivalence and hesitancy recognition. The proposed method first extracts whole-video segment tokens and cropped-face segment tokens using the same temporal partition. The synchronized visual and facial tokens are refined through intra-modal self-attention and bidirectional visual-facial cross-attention, allowing whole-video context and local facial behavior to mutually refine each other before evidence construction. We then construct segment-level visual-facial evidence using consistency and discrepancy modeling, followed by temporal self-attention and attention pooling. Textual and acoustic features are lightly refined through context self-attention and are fused with the enhanced visual-facial evidence at the final decision stage using pairwise evidence fusion. Experiments on the BAH (Behavioral Ambivalence/Hesitancy) public evaluation split show that the proposed synchronized visual-facial cross-refinement improves public macro-F1 over both global visual-face token fusion and synchronized evidence baselines, achieving a public macro-F1 of 0.7156. Code is available at : https://github.com/hiinnnii/BAH-Challenge-ECCV2026\_SVF-CR. --- ### Self-Guided Test-Time Training for Long-Context LLMs Authors: Xinyu Zhu, Zhe Xu, Xiaohan Wei, Yunchen Pu, Fei Tian, Chonglin Sun, Kaushik Rangadurai, Hua Zhi, Frank Shyu, Sandeep Pandey, Luke Simon, Yu Meng et al. Published: 2026-07-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2607.09415v1 Long-context processing has become increasingly important for large language models (LLMs), but simply extending the context window does not guarantee effective utilization of long inputs. As input length grows, accuracy often degrades, indicating that models still struggle to identify and use the evidence most relevant to a question. A promising way to improve long-context utilization is test-time training (TTT), which treats the test context as a training example for instance-specific parameter adaptation. However, applying TTT to the entire long context is prohibitively expensive, while adapting on randomly sampled spans introduces severe noise. Because most spans in a long context are irrelevant to the specific question, training on them may even degrade the base model's performance. Our preliminary study shows that TTT is highly sensitive to training-span quality: on LongBench-v2, TTT on randomly sampled spans hurts performance, whereas TTT on oracle spans substantially improves it. Motivated by this, we propose a simple method, Self-Guided TTT (S-TTT): before adaptation, the model identifies the evidence spans it should learn from, and the standard language-modeling training objective is applied only to those selected spans. On two challenging long-context reasoning benchmarks, LongBench-v2 and LongBench-Pro, S-TTT improves accuracy for both Qwen3-4B-Thinking-2507 and Llama-3.1-8B-Instruct, achieving up to a 15% relative improvement. --- ### Similarity search generalisation in contrastive learning with InfoNCE loss Authors: Nick Whiteley Published: 2026-07-10 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2607.09405v1 Similarity search is a primary application of embedding models trained by contrastive learning. For one of the most popular contrastive learning loss functions, InfoNCE, we show that the population risk with $k$ negative samples is $O(1/k)$ close to an expected cross-entropy which quantifies deviation between i) a softmax similarity search over unseen data using the learned embedding function, and ii) an idealised softmax search over the same data but using similarity implicitly represented in the positive sample generator. This complements existing interpretations of InfoNCE in the $k\to\infty$ limit which are phrased in terms of mutual information, and alignment versus uniformity in embeddings. To quantify generalisation performance, we introduce a new continuity bound for the InfoNCE loss, obtained via Gâteaux differentiation. The bound preserves the structure of averaging over negative samples present in the loss function and features an ``inverse temperature'' parameter which can be tuned to account for the algorithmic temperature. For embedding functions which are Lipschitz in a parameter, this yields a simple demonstration that the averaging effect of $k$ negative samples in the InfoNCE loss carries over to stabilisation of the generalisation error as $k$ grows. --- ### SYNRARE: Synthetic Rare Disease EHR Generation for ML Benchmarking Authors: Nicolai Dinh Khang Truong, Richard Röttger Published: 2026-07-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.09404v1 Motivation: Rare disease (RD) diagnosis is frequently delayed due to the similarities in symptoms to common disease variants. Machine Learning Algorithms applied to Electronic Health Records show promise for accelerating the diagnosis; however, legal and privacy concerns pose significant barriers. To address these issues, Synthetic Data Generation is an alternative method for obtaining Electronic Health Records and can be applied with any Machine Learning algorithm for benchmarking and development purposes. Despite the availability of Synthetic Data Generation algorithms, support for generating a subset of patients that differ in a definable degree from the majority to simulate patients with RD is often lacking. Results: We present SYNRARE, a graphical user interface based on the Synthea framework that enables easier modification and generation of synthetic Electronic Health Records of RD patients, which differ only to a definable degree from patients with common diseases, thereby enabling the benchmarking and testing of algorithms under controlled technical conditions. SYNRARE enables researchers to rapidly benchmark their Machine Learning algorithms across any scenario. Availability and implementation: SYNRARE, including detailed instructions for installing, is available at https://gitlab.sdu.dk/screen4care/synrare. --- ### Fictional Worldbuilding: Multi-Agent LLM Collaboration with Hierarchical Context Compression and Iterative Review Authors: Jingbo Chen, He Wang, Wei Yuan, Yuqiao Lai, Zhenyan Lu Published: 2026-07-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.09403v1 Worldbuilding, the construction of coherent fictional worlds, is a foundational task in game design and literary creation. Large Language Models (LLMs) offer new possibilities for automated content generation, but their application to worldbuilding faces three challenges: context explosion that grows linearly with the building process, the tension between creative diversity and content consistency, and the absence of automated quality assurance. This paper presents AutoWorldBuilder, a multi-agent collaborative system that addresses these challenges through five integrated components: a structured concept network with conflict detection; a DAG-based hybrid batch scheduler that groups tasks by semantic locality; a four-layer context compression mechanism achieving approximately 90% token reduction; an iterative review system with specialized Auditor agents that improves proposal pass rates from 42% to over 85%; and a skill-driven agent architecture supporting zero-code extension with differentiated temperature configuration. Two experiments across 20 diverse worldbuilding tasks, using GPT-OSS 120B and DeepSeek v3.2 as LLM backends, demonstrate a 95.0% success rate. The system generated 56-103 self-consistent concepts per world in 18-31 minutes with zero-conflict delivery. The architectural patterns validated here, including layer-as-budget compression, semantic-locality scheduling, and separation of generation and review, transfer to the broader class of knowledge-intensive, multi-agent LLM applications. --- ### Data-Efficient Deep Learning: Empirical Guidelines for Training Set Size Estimation in Inertial Sensor Classification Authors: Ofir Kruzel, Itzik Klien Published: 2026-07-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.09402v1 Deep learning models dependency on large-scale inertial datasets presents a significant bottleneck in inertial sensor-based classification tasks, such as human activity recognition and smartphone location recognition. In these domains, data collection requires massive recording campaigns that are complex, time-consuming, and difficult to scale. Currently, data-driven guidelines for determining the minimum sample size required to reach a desired accuracy level do not exist. To address this gap, this study presents a systematic empirical evaluation of learning curve convergence rates in inertial classification. We introduce a unified framework that analyzes classification performance under both binary and multi-class scenarios, and derive an empirical formula to estimate performance relative to dataset size. Testing across six diverse, real-world datasets totaling 102.7 hours of inertial measurements demonstrates that accuracy follows a consistent logarithmic growth pattern, regardless of task complexity. Leveraging this finding, we propose a quantitative stability point metric, defined as the sample size required for the learning curve to stabilize within a predefined mean absolute percentage deviation of its asymptotic maximum. Our analysis reveals that models often reach practical stability with substantially fewer samples than traditional heuristics suggest. Ultimately, we offer a generalizable framework to extrapolate total data requirements from small-scale pilot studies, optimizing the tradeoff between recording effort and model reliability. These findings shift the prevailing paradigm from maximizing data volume toward optimizing data efficiency, offering concrete, data-backed guidelines for planning recording campaigns in inertial sensing applications. --- ### On-Device Adaptive Battery Power Prediction for Electric Vehicles Authors: Avik Bhatnagar, Anton Paule, Tobias Schuermann, Sebastian Reiter, Oliver Bringmann Published: 2026-07-10 Categories: cs.LG, cs.AI, cs.AR, cs.PF arXiv: https://arxiv.org/abs/2607.09400v1 Adaptive power management in Electric Vehicles (EVs) requires accurate power prediction. Although deep learning models have emerged as highly effective for time-series forecasting in this domain, their performance is prone to degradation when exposed to data with distributions different from the training data. We introduce a novel approach that enables on-device learning in resource-constrained EV systems to continuously adapt pretrained battery prediction models to new, unseen data. We leverage existing pretrained models by transforming them into adaptable versions that retain critical hyperparameter knowledge from their initial training. We comprehensively investigate both online and offline model adaptation strategies. Our results demonstrate significant improvements in forecasting performance across various models and time horizons, achieving mean absolute error reductions of up to 7.49\% and 14.88\% with online and offline adaptation techniques, respectively. This study highlights the substantial benefit of on-device adaptation, resulting in enhanced battery power predictions than unadapted model deployments in real-world EV scenarios. --- ### Fully Trainable Deep Differentiable Logic Gate Networks and Lookup Table Networks Authors: Wout Mommen, Lars Keuninckx, Matthias Hartmann, Werner Van Leekwijck, Piet Wambacq Published: 2026-07-10 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.09399v1 We introduce a novel method for both partial and full optimization of the connections in deep differentiable logic gate networks (LGNs) and lookup table networks (LUTNs). Our training method utilizes a probability distribution over a set of connections per gate/lookup table (LUT) input pin, selecting the connection with highest merit, all whilst the optimal gate types or LUT-entries are learned in parallel. We show that the connection-optimized LGNs outperform standard fixed-connection LGNs on the Yin-Yang, MNIST Handwritten Digits and Fashion-MNIST benchmarks, while requiring only a fraction of the number of logic gates. We achieve 98.92% on the MNIST dataset with two layers of 8000 gates. With only one layer of 8000 gates, we obtain 98.45%, showing that our method requires almost 50 times fewer gates compared to fixed-connection LGNs. Training stability up to ten layers has been ensured by employing a high learning rate, straight-through estimators and trimming constant-output gate types. Additionally, we present a LUT neuron description that enables stable training with backpropagation, tested up to 6-layer deep networks. The model requires four times fewer trainable parameters and still achieves a higher accuracy compared to the fixed-connection LGN training algorithm. Our connection-training algorithm also works well for the LUTNs, achieving an accuracy of 98.88% for two layers of 2000 6-input LUTs. --- ### STEEL: Sparsity-Aware Fused Attention for Energy-Efficient Long-Sequence Inference on AMD's XDNA NPU Authors: Victor J. B. Jung, Gagandeep Singh, Joseph Melber, Kristof Denolf, Francesco Conti, Luca Benini Published: 2026-07-10 Categories: cs.DC, cs.AI, cs.PF arXiv: https://arxiv.org/abs/2607.09385v1 The growing adoption of large language model-based agents within operating system workflows has increased the importance of energy-efficient inference on laptop-class systems-on-chip (SoCs). While cloud offloading remains common, it introduces reliability and privacy concerns that are particularly problematic for agentic workloads. Recent laptop SoCs, therefore, incorporate neural processing engines (NPUs) optimized for energy efficiency; however, effectively mapping attention mechanisms onto NPUs remains challenging due to architectural diversity and explicit data-movement programming models. In this work, we present STEEL, the first open-source implementation of FlashAttention targeting XDNA-like NPUs. STEEL introduces a dataflow formulation of prefill attention, enabling efficient exploitation of spatial parallelism and on-chip memory. Furthermore, STEEL addresses the load imbalance induced by the causal mask by leveraging a sparsity-aware pipeline placement onto the NPU array, reducing synchronization overhead and improving utilization. We evaluate STEEL on the AMD Ryzen AI 9 HX 370 SoC and compare its performance against optimized CPU and GPU implementations. Experimental results show that STEEL reduces energy consumption by an average of 9.17x and 1.75x relative to CPU and GPU baselines, respectively. On XDNA 1, STEEL achieves an average 9.6x latency reduction over the prior state of the art, and delivers a 22.8x speedup on average compared to a layer-by-layer attention implementation on XDNA 2. --- ### Learning Physics-Informed Surrogate Model of Linear Elastic Displacement Fields from Geometry Authors: Rodolphe Barlogis, Ferhat Tamssaouet, Quentin Falcoz, Stéphane Grieu Published: 2026-07-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.09382v1 This work aims to develop a fast and physically consistent surrogate model for real-time structural health monitoring of fractured elastic domains. We propose a physics-informed DeepONet framework that predicts displacement fields from both boundary conditions and fracture geometry, using a dedicated encoding strategy for the latter and without relying on finite-element-generated training data. The traction-free condition on the fracture boundary is imposed weakly through a localized penalty term. The presented numerical example focuses on one representative fracture geometry, demonstrating the feasibility of the formulation and laying the groundwork for extensions to surrogate modeling across diverse fracture geometries. --- ### When Routes Run Out: Adversarial Co-Learning and Explainable Robustness in Quantum Repeater Networks Authors: Brennan Bell, Inti Gabriel Mendoza Estrada, Andreas Trügler, Paul Erker Published: 2026-07-10 Categories: quant-ph, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2607.09378v1 We study an adversarial bandit problem for entanglement-based quantum-network routing over a modest graph corpus. Alice selects an end-to-end repeater route for an Ekert-91 protocol (E91) representing her move, while Eve selects an attack surface, either edge intercept--resend or repeater memory degradation. Payoffs are drawn from cached SeQUeNCe-simulated E91 transcripts, and Alice accepts a turn when the finite-sample statistic violates the Clauser-Horne-Shimony-Holt (CHSH) bound. Performing adversarial co-learning across 50 structured topologies, we find that learned retention tracks a full-matrix minimax reference closely (Pearson $r=0.99$): under a one-surface Eve action model, bottleneck families have zero retention, while non-bottleneck families follow a $1-1/N$ coverage principle. We then fit decision-tree explanation models to graph-, attack-, and route-level topology-corpus targets and report their faithfulness. Finally, we construct prompt records for local language models to summarize the tree evidence, resulting in an open-source explanation workflow for quantum-repeater network games. --- ### Mach-Mind-4-Flash Technical Report Authors: Foundation Model Team Published: 2026-07-10 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2607.09375v1 We present Mach-Mind-4-Flash, a 35B-parameter Mixture-of-Experts (MoE) agentic model with 3B activated parameters. Through post-training optimization alone without scaling pre-training compute, the model achieves performance on par with or surpassing that of 100B-parameter-class models. By introducing scalable agentic interaction environments for large-scale reinforcement learning, the model attains significant performance gains on real-world application tasks. Our pipeline comprises three stages: (1) a unified RL/OPD training infrastructure with dynamic multi-teacher scheduling and operator-level acceleration, delivering 17\% end-to-end training speedup; (2) multiple domain-specific RL experts trained in parallel across Reasoning, General, and Agent tracks, then fused into a single generalist via Multi-Teacher On-Policy Distillation (MOPD) -- a routed reverse-KL objective that eliminates the see-saw degradation of mixed-reward RL; (3) Hybrid Median-length Policy Optimization (HMPO), a single-stage token-efficiency method that compresses reasoning chains by 19--46\% with $\le$0.7 percentage-point accuracy loss. Mach-Mind-4-Flash scores 92.70 on AIME'26, 82.82 on IFBench, 80.74 on Behavioral-SafetyBench, 75.80 on BFCL-v4, 72.31 on BrowseComp-zh, and 84.20 on ClawBench -- leading or matching models with 10--30$\times$ its activated size at a fraction of the inference cost. --- ### Graph Neural Networks for Scalable and Transferable Node Centrality Approximation Authors: Samra Sana, Giorgio Mantica, Saul Imbrici Published: 2026-07-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.09372v1 Graph Neural Networks (GNNs) provide a learning-based framework for approximating graph quantities that are expensive to compute exactly. This paper investigates GNNs for scalable approximation of betweenness and closeness centrality, formulated as a node-ranking problem. Exact centrality values are used as supervision, and ranking quality is evaluated using Kendall's tau rank correlation. We study whether message-passing GNNs can learn transferable structural representations across different graph topologies rather than only fitting the distribution used during training. On unseen Erdos renyi graphs, the proposed models achieve tau = 0.851 for betweenness and tau = 0.894 for closeness. A large-scale betweenness model trained on graphs with N = 5,000 nodes achieves tau = 0.938, demonstrating scalability. Mixed-distribution training on Erdos renyi, Barabasi-Albert, and Gaussian Random Partition graphs improves betweenness transfer across graph families. In contrast, closeness centrality remains more sensitive to community-structured graphs and shows reduced transfer to real-world topologies. Finally, GNN inference achieves up to a 97.7x speedup over exact computation. These results show that mixed-distribution training can improve structural transfer in GNN-based centrality approximation, while identifying closeness centrality's sensitivity to topology as an open challenge. --- ### Spectrally Deconfounded Gradient Boosting Authors: Andrea Nava, Peter Bühlmann, Fabio Sigrist Published: 2026-07-10 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2607.09371v1 Flexible machine-learning methods can be sensitive to hidden confounding: they may learn associations induced by unobserved confounders rather than stable signals. Spectral deconfounding mitigates this problem by shrinking high-variance directions of the covariate matrix that, under dense confounding, carry latent confounder information. Existing work has largely focused on linear models. We develop a nonlinear spectral deconfounding framework for gradient boosting. Our approach replaces the ordinary squared-error loss by a spectral loss, which alters the boosting dynamics by slowing down learning in confounding-aligned directions. We show that deconfounding is not achieved by the spectral loss alone, but by the interaction between spectral shrinkage and regularization, especially in terms of early stopping. Moreover, we provide a mixed-model interpretation that connects LAVA-type shrinkage to random-effects adjustment and yields an empirical-Bayes procedure for tuning the spectral loss. We also extend the method to general likelihoods and nonlinear confounding using Laplace approximations and kernel random effects. Across synthetic and real-world experiments, spectrally deconfounded boosting improves estimation of the target function under hidden confounding and is substantially more scalable than existing nonlinear spectral deconfounding baselines. --- ### Diversifying to Verify: When Task-Equivalent Programs Differ in Verifiability Authors: Shirley Yu, Ruben Martins Published: 2026-07-10 Categories: cs.SE, cs.AI, cs.LO arXiv: https://arxiv.org/abs/2607.09366v1 Program verification is crucial for software correctness, but producing fully verified programs remains difficult in practice. This paper studies whether implementation structure affects automated verifiability when multiple generated programs are intended to satisfy the same task-level semantics. We present Diversify2Verify, a staged LLM-based pipeline for Why3 that infers representation-specific contracts, generates and tests diverse recursive and imperative array/list implementations, and attempts verification with bounded verifier-guided annotation repair. We also construct a verification-oriented benchmark of 73 tasks over integers, arrays, and lists, yielding 292 implementation variants. Diversify2Verify verifies 96 artifacts initially and 154 after two repair passes, improving artifact-level verification from 32.9% to 52.7%. At the task level, at least one variant verifies for 49 of 73 tasks, a 67.1% success rate. These results show that task-equivalent implementations can differ substantially in verifiability and that implementation diversity helps find verification-friendly artifacts. --- ### CtrlVTON: Controllable Virtual Try-On via Visual-Instance-Prompt Segmentation Authors: Seungyong Lee, Hyun Jun Jang, Sangoh Kim, Sungjoon Park Published: 2026-07-10 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.09362v1 Virtual try-on (VTO) has made significant progress in realistically transferring garments onto a target person. Yet most systems give the user little control over how a garment should be worn -- its size (loose or fitted), style (e.g., tucked in or untucked, open or closed), and spatial placement on the body. We address this gap with two complementary contributions. First, we define and solve Visual-Instance-Prompt Segmentation via VIP-SAM: given a flatlay image of a garment, segment that specific instance in a photograph of a person wearing it. This is an instance-level task, distinct from the typically studied category-level segmentation. Second, we introduce CtrlVTON, a controllable VTO framework that recasts try-on as an image editing problem and adds segmentation masks as pixel-level control over garment layout, including style, size, and spatial placement on the body. VIP-SAM and CtrlVTON each achieve state-of-the-art results on their respective tasks. In particular, CtrlVTON generates images that follow user-provided layouts far more faithfully than the strongest proprietary editing systems while matching them on garment fidelity. --- ### Simon-SR: Spatially Adaptive Modulation and Visual Prompt Adaptation for Text-Reinforced Super-Resolution Authors: Haotong Cheng, Yuxuan Li, Zijie Cui, Rongling Tan, Chenyuan Wang Published: 2026-07-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.09351v1 Single Image Super-Resolution (SISR) reconstructs high-quality images from low-resolution inputs. While recent multi-modal methods improve perceptual quality, they remain sensitive to erroneous priors and require expensive annotations. To address these issues, we propose Simon-SR, a multi-modal SISR framework leveraging learnable prompts for efficient semantic mining and robust text-image fusion. Our approach combines Contrastive Prompt Learning with Prompt-Guided Spatially Adaptive Refinement to enhance multi-modal alignment. Experiments demonstrate that Simon-SR surpasses state-of-the-art methods, achieving maximum improvements of 0.50 dB in PSNR, 0.0133 in SSIM, and 0.0695 in LPIPS. Code will be released. --- ### Deceptive Grounding: Entity Attribution Failure in Clinical Retrieval-Augmented Generation Authors: Cedric Caruzzo, Donggeun Yoo, Tae Soo Kim Published: 2026-07-10 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2607.09349v1 Retrieval-augmented generation evaluation checks whether model claims are factually grounded in retrieved documents. It does not check whether retrieved evidence is attributed to the correct entity. A clinical RAG response can pass every automated check (zero hallucinations, near-perfect faithfulness, real citations) while presenting drug Y's clinical evidence as evidence about queried drug X. We term this deceptive grounding (DG): a failure invisible to faithfulness, hallucination, and citation checks because every claim is sourced from a real document, about the wrong entity. Using a controlled factorial benchmark across 13 models, we find DG rates spanning 8-87% at peak adversarial conditions. Medical and biomedical fine-tuned models reach up to 86.7%; domain specialization amplifies the failure rather than mitigating it. A controlled ablation identifies the mechanism: removing entity-specific clinical evidence from retrieved documents eliminates entity-attribution failure entirely, shifting all failures to confabulation. The two failure modes respond to the same trigger, taking different paths. Production measurement across 740 drug-disease pairs finds 7.8% overall DG in a deployed RAG system, rising to 13.6% for recently approved drugs. Entity-attribution verification (checking that cited evidence applies to the queried entity) detects DG at 97.0% precision and 98.7% DG recall (IPW-adjusted human gold standard); no existing framework implements it. --- ### DKCD: Domain Knowledge-Enhanced Causal Discovery from Unstructured Data Authors: Xin Li, Jin Li, Shoujin Wang, Kun Yu, Fang Chen Published: 2026-07-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2607.09348v1 Causal discovery from unstructured data is a challenging yet underexplored task in high-expertise domains such as healthcare, finance, and education. Existing methods typically leverage the general knowledge of large language models (LLMs) to identify causal factors from unstructured data and annotate them into structured data for causal graph construction. However, they remain limited by two key challenges (CHs): (CH1) insufficient identification of latent factors, which are implicit in the data yet essential for causal discovery, due to the lack of domain-specific knowledge; and (CH2) unreliable factor annotation, caused by the lack of domain-grounded reasoning, which propagates errors to the resulting causal graphs. To address these challenges, we introduce a novel Domain Knowledge-enhanced Causal Discovery framework (DKCD) for causal discovery from unstructured data in high-expertise domains with three interconnected components: (1) Knowledge Mining: It retrieves relevant domain knowledge based on observable factors to support subsequent causal reasoning. (2) Knowledge-guided Causal Reasoning: Reasoning with relevant knowledge, it discovers latent causal factors to address CH1 and generates key causal clues for more accurate data annotation to address CH2. (3) Causal Structure Discovery: It constructs the final causal graphs based on a more complete factor set and accurate annotations. Experiments on two domain-specific datasets show that DKCD significantly improves both causal factor identification and causal graph construction. --- ### Towards Detecting Inconsistencies in End-to-end Generated TODs Authors: Tiziano Labruna, Giovanni Bonetta, Bernardo Magnini Published: 2026-07-10 Categories: cs.CL, cs.SC arXiv: https://arxiv.org/abs/2607.09338v1 Generative AI is profoundly transforming the core technologies behind conversational systems, shifting from component-based to end-to-end approaches. However, Large Language Models (LLMs) may still generate inconsistencies, a critical issue particularly in Task-Oriented Dialogues (TODs), where system responses must strictly adhere to information from a domain knowledge base (e.g., restaurants in a city). A single hallucination (e.g., suggesting a non-existent restaurant) can lead to severe task failures. We investigate a method for automatically detecting inconsistencies by conceptualizing TODs as a Constraint Satisfaction Problem (CSP), where variables represent dialogue segments referencing the conversational domain, and constraints among variables capture dialogue properties such as turn coherence and adherence to domain knowledge. We propose a pipeline that first identifies variables in a target dialogue and then applies a CSP solver to identify valid solutions. By comparing the target dialogue with valid variable assignments, we can detect inconsistencies and suggest minimal changes to ensure dialogue consistency. We demonstrate the high accuracy of the CSP-based approach in detecting inconsistencies, and provide a detailed analysis of our findings. --- ### Shortcut Trajectory Planning for Efficient Offline Reinforcement Learning Authors: Guanquan Wang, Yoshimasa Tsuruoka Published: 2026-07-10 Categories: cs.LG, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2607.09336v1 Diffusion-based trajectory planners have shown strong performance in offline reinforcement learning, but their iterative denoising process often incurs high inference cost. Consistency-based planners reduce the number of sampling steps, yet they typically rely on a two-stage teacher--student distillation pipeline that increases training cost and may introduce instability. We propose Shortcut Trajectory Planning (STP), an offline model-based reinforcement learning framework that incorporates shortcut models as efficient trajectory generators. STP trains a conditional shortcut trajectory model in a single stage, supports adjustable one-step and few-step inference through step-size conditioning, and selects candidate plans using a critic augmented with feasibility-aware correction. Across standard D4RL benchmarks, including locomotion, navigation, manipulation, and dexterous control tasks, STP achieves strong performance while simplifying the training pipeline for fast generative planning. --- ### Communication-Efficient Digital-Twin Coordination for Heterogeneous LLM Embodied Agents over Computing Power Networks Authors: Nuocheng Yang, Sihua Wang, Zihan Chen, Tony Q. S. Quek, Changchuan Yin Published: 2026-07-10 Categories: cs.AI, cs.MA arXiv: https://arxiv.org/abs/2607.09330v1 Embodied agent teams powered by heterogeneous large language models (LLMs) are being widely deployed in physical artificial intelligence such as smart factories, warehouses, and service robotics. To enable collaboration among such an agent team, efficient coordination mechanisms that operate reliably under limited network resources are required. However, existing heterogeneous LLM-agent coordination frameworks that rely on multi-round natural-language-based conversations introduce three coupled challenges. First, inter-agent dialogue incurs communication overhead that grows rapidly with team size. Second, the quality of coordination is constrained by the heterogeneous capabilities of the agent team's LLMs. Third, agents may suffer from action delays due to iterative negotiation. To address these challenges, we propose LDT-Coord, a networked coordination framework built upon a lightweight digital twin (DT). Specifically, each agent independently selects its intended action and reports both the action decision and a structured temporal constraint over shared resources to the DT server, thereby decoupling coordination performance from natural-language reasoning ability. Then, DT executes a training-free, rule-based orchestrator algorithm to resolve cross-agent conflicts and returns coordination instructions to prevent such conflicts. To further reduce communication overhead, we formulate agent reporting control as a constrained partially observable Markov decision process (C-POMDP) and solve it with the PPO-Lagrangian algorithm. Simulation results show that LDT-Coord achieves a task success rate comparable to conventional coordination methods while reducing communication overhead by more than 70x and maintaining robustness under LLM heterogeneity. --- ### Dynamic Inverse Rendering for Enhanced Material-Lighting Decomposition Authors: Raza Yunus, Benjamin Ummenhofer, Jan Eric Lenssen, Eddy Ilg Published: 2026-07-10 Categories: cs.CV, cs.GR arXiv: https://arxiv.org/abs/2607.09329v1 Decomposing outgoing surface radiance into material and illumination during inverse rendering is essential for applications such as relighting and augmented reality, yet it is severely ill-posed since multiple combinations can result in the same observed colour. Capturing an object under multiple lighting conditions usually helps resolve this ambiguity as it constrains the optimization towards correct solutions. In this work, we explore the potential of reconstructing rigidly moving objects -- which provides observations of diverse light-surface interactions -- to resolve the material-lighting ambiguity in inverse rendering. For this purpose, we introduce a relightable approach that marries object tracking and reconstruction with inverse rendering for general rigidly moving objects. Our experimental analysis on synthetic data demonstrates that motion can be an advantage for disentangling material and lighting: the reconstructed material is significantly more accurate when the object is observed under rigid motion than when it is static. Moreover, results on RGB videos of real hand-held objects show that our pipeline preserves this advantage even under noisy real-world conditions. --- ### WildTrace: Benchmarking Natural Evidence Trails in Long-Context Reasoning Authors: Zixin Chen, Peng Liu, Haobo Li, Rui Sheng, Jianhong Tu, Xiaodong Deng, Fei Huang, Kashun Shum, Dayiheng Liu, Huamin Qu Published: 2026-07-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2607.09328v2 Answering complex questions over long documents frequently requires integrating evidence that the source itself disperses naturally across distant passages. In an incident report, the operating condition, design flaw, and missed safety check that jointly explain a disaster may appear dozens of sections apart; in a novel, a character's true motive may surface only through scenes far removed from the moment it becomes relevant. This source-internal evidence integration is central to real-world long-document analysis, yet existing benchmarks largely sidestep it. Needle probes, planted facts, and reverse-engineered multi-hop chains embed evidence that may differ from the host text in distribution, placement, or register, making it unclear whether strong performance reflects genuine source reasoning or distributional artifacts. We introduce WILDTRACE, a benchmark of 481 tasks over 214 naturally occurring long-form sources such as technical incident reports and lesser-known literary narratives, where all evidence trails arise from the document's own causal, temporal, and narrative logic. Drawing on Pearl's causal hierarchy and prior multi-hop reasoning typologies, we define seven source-internal evidence geometries that characterize the distinct relational demands of analytical reading in long documents. A source-first construction pipeline mines candidate trails from document structure before writing questions; each item then undergoes multi-stage validation covering clue necessity, answer groundedness, rubric fidelity, contamination resistance and answerability. As models are increasingly entrusted with real-world high-stakes analytical tasks, this gap between accessing information and reasoning over naturally dispersed evidence emerges as a defining challenge for the next stage of long-context research. --- ### Letting the Data Speak: Extracting Keywords from Crowdsourced Collections with AI Authors: Miguel Arana-Catania, Catherine Conisbee, Matthew Kidd Published: 2026-07-10 Categories: cs.CL, cs.AI, cs.DL, cs.IR, cs.LG arXiv: https://arxiv.org/abs/2607.09324v1 Identifying and assigning keywords at scale is a technical, practical, and ethical challenge for crowdsourced collections. This article reports the findings of the "Extracting Keywords from Crowdsourced Collections" project, which used the Their Finest Hour Online Archive, a crowdsourced Second World War digital collection hosted by the University of Oxford, as a case study. The project evaluated three Natural Language Processing approaches to automate keyword extraction: Named Entity Recognition, Keyword Extraction, and Topic Modelling. It tested these approaches across a range of artificial intelligence techniques, from traditional statistical methods to modern GenAI neural networks. Our quantitative and qualitative findings indicate that Natural Language Processing approaches offer real potential for keyword extraction at scale in crowdsourced collections, but that no single method offers a complete solution and that model choice significantly shapes results. We argue that in crowdsourced collections, where metadata is the direct product of engagement with living contributors, automated keyword extraction raises distinct stewardship responsibilities that must be addressed alongside technical performance. Open-weight, extractive models emerge from our evaluation as best placed to support responsible deployment, while generative AI, despite its abstractive potential, introduces accountability risks that anyone managing crowdsourced collections should weigh carefully. --- ### LongMedBench: Benchmarking Medical Agents for Long-Horizon Clinical Decision-Making Authors: Zihan Xu, Yanzhen Chen, Xiaocheng Zhang, Zhiting Fan, Weiqi Zhai, Hongxia Xu, Zuozhu Liu Published: 2026-07-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.09322v2 In this work, we introduce LongMedBench, a real-world EHR-based benchmark for long-horizon clinical decision-making. Prior evaluations of LLM-based medical agents have largely emphasized short-context knowledge QA and tool use. However, real-world medical care is inherently longitudinal, and clinicians must aggregate evidence across repeated visits, tests, and evolving treatments. Therefore, long-horizon interaction is essential for realistic assessment. LongMedBench is constructed via a reproducible pipeline that integrates MIMIC-IV admission records and clinical notes into time-series event streams and long-context memory datasets, enabling long-horizon, multi-session interactions between agents and a clinical environment. It comprises 335 patients, with 19.72 inpatient visits per patient on average and 44.91 medical events per visit. Guided by the long-horizon decision process, we propose an evaluation taxonomy with three suites: fact-based QA, temporal reasoning, and long-horizon decision-making. This taxonomy measures how agents understand and leverage historical patient information over extended horizons. Our experiments show that while recent LLMs can make good use of explicit timestamps, they have challenges in implicit time inference; The RAG and agent memory system can improve the performance of information retrieval tasks, but the performance of decision-making tasks is highly dependent on the model's immediate context. --- ### Automatic Thematic Indexing of Large Literary Corpora: A Machine Learning Approach to Voltaire's Complete Works Authors: Miguel Arana-Catania, Gillian Pink, Glenn Roe Published: 2026-07-10 Categories: cs.CL, cs.AI, cs.DL, cs.IR, cs.LG arXiv: https://arxiv.org/abs/2607.09316v1 Thematic indexing -- the practice of assigning structured conceptual labels to sections of text -- is essential to scholarly access in large-scale literary and historical editions, yet it remains a largely manual, labour-intensive process. This paper explores the application of machine learning to automatic thematic indexing, using two substantial sub-corpora of the Complete Works of Voltaire as a test case: the Essai sur les mœurs et l'esprit des nations and the Questions sur l'Encyclopédie. The task is framed as a multi-label classification problem, in which a model must assign the set of index entries that a professional indexer would apply to a given page of text. We compare a range of approaches -- from encoder-based models with classification heads to generative large language models (LLMs) fine-tuned via Low-Rank Adaptation (LoRA) -- spanning model sizes from approximately 3 to 120 billion parameters. Our best-performing model, from the Mistral family in a 4-bit quantised configuration, achieves F1 scores of up to 0.67; we argue that these figures represent lower bounds, given the inherent subjectivity of professional indexing and the frequency with which model predictions prove semantically valid despite diverging from the print index. We further evaluate cross-corpus generalisation and conduct a detailed qualitative analysis of model behaviour on literary and rhetorical features of the source texts that prove particularly resistant to automated treatment. Our findings have implications for the broader challenge of providing structured thematic access to large-scale literary and historical corpora. --- ### Exposure is not manifestation: measurement target and output resolution jointly determine which behavioural-faithfulness evaluator wins Authors: Kwan Soo Shin Published: 2026-07-10 Categories: cs.CL, cs.AI, cs.HC, cs.LG arXiv: https://arxiv.org/abs/2607.09306v3 Behavioural auditing asks whether a language model behaves as it claims, but detection scores are reported without separating two targets: whether a reply was produced under a behaviour-inducing condition (exposure) and whether the behaviour surfaced in it (manifestation). Scoring a compact 146-million-parameter auditor's frozen-representation read-out and a frontier judge against each label on the identical 720 replies, the gap between the instruments moves by roughly 0.2 AUROC when the target changes. Under the judge's deployed interface, a single verdict, the ranking reverses: the auditor leads on exposure, 0.804 against 0.718, and trails on manifestation, 0.690 against 0.811. Matching the output resolution from either direction, by asking the judge a target-specific question answered with a continuous confidence score or by thresholding the auditor's read-out, removes the reversal but not the interaction, which excludes zero at all three resolutions (0.207, 0.237 and 0.169). The target governs how far apart the instruments are; the interface governs whether that distance changes their order. The auditor's hyperbolic geometry confers no advantage here. A single behavioural-detection AUROC is under-specified: such claims are comparable only when they state the estimand, the evaluator, and its output interface. --- ### From Classification to Localization and Clinical Validation: Large-Scale Development of a Deep Learning System for Thoracic Disease Detection on Chest Radiographs in Thailand Authors: Isarun Chamveha, Tretap Promwiset, Napat Wanchaitanawong, Trongtum Tongdee, Pairash Saiviroonporn, Warasinee Chaisangmongkon Published: 2026-07-10 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2607.09305v2 Chest radiography (CXR) remains the most widely used thoracic imaging modality, yet expert interpretation is constrained by a severe shortage of radiologists in Thailand and across Southeast Asia. Local adaptation of deep learning models to Thai data has been shown to substantially improve accuracy on Thai populations. Here we present the development and comprehensive validation of the chest radiograph analysis model in Inspectra CXR version 5, a deep learning system that performs multi-label thoracic disease classification and weakly supervised lesion localization within a single model. The architecture couples a DenseNet-121 backbone with Attend-and-Compare Modules (ACM) and a Probabilistic Class Activation Map (PCAM) aggregation layer, producing a per-condition classification score and heatmap simultaneously. The model was developed on 874,858 frontal chest radiographs with paired radiologist reports from Siriraj Hospital, Bangkok. On a held-out, radiologist-verified in-domain test set of 19,871 cases, it achieved a mean AUROC of 0.994 (mean sensitivity 92.4%, specificity 98.6%) across nine clinically important conditions. On an independent generalization set of 5,992 cases from 13 hospitals across Thailand, the mean AUROC was 0.970, indicating robust transfer across sites. For localization, evaluated on 4,549 radiologist-annotated cases, the model attained a mean lesion-localization fraction (LLF) of 77.9% at 0.59 non-lesion localizations per image. In a usability evaluation with five thoracic radiologists, the system reached a classification concordance of 93.6%, a localization concordance of 94.7%, and a mean System Usability Scale (SUS) score of 89. These results indicate that a locally developed, localization-capable CXR system can deliver high accuracy, generalize across heterogeneous Thai hospitals, and earn the trust of practicing radiologists. --- ### TextileNet: Towards Zero-shot Text-style Segmentation of Manuscripts Authors: Anguelos Nicolaou, Antonella Ambrosio, Desiree Di Donato, Georg Vogeler Published: 2026-07-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.09299v1 Automatic writer identification systems have progressed remarkably in recent years, yet their deployment in archival paleography remains limited by the scarcity of labeled training data, open scribe sets, and degraded image quality. We present TextileNet, a fully convolutional multi-task network trained exclusively on synthetic data to produce dense pixel-level texture embeddings, which we transfer zeroshot to historical manuscript analysis. As an original contribution to evaluation methodology, we designed a paleographic visual quiz of 80 pair and triplet questions and administered it to a range from lay participants to senior paleographers under strict anonymity, establishing to our knowledge for the first time a human baseline for script-style discrimination on late medieval text. We employ TextileNet embeddings to perform zero-shot retrieval on sub-word granularity for hand and gender identification. Our experimental results help in building the credibility of TextileNet in the paleographic domain, but more than that demonstrate in experimental terms that the question of gender in handwriting needs to be treated with caution. ---