# Research Papers — 2026-06 · bundle page 185 of 260 > Full-text markdown of papers #9201–9250 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/185 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/184.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/186.md) ### Overcoming Decoder Inconsistencies in Whisper for Dravidian and Low-Resource Languages Authors: Chowdam Venkata Kumar, Kumud Tripathi, Pankaj Wasnik Published: 2026-06-08 Categories: cs.CL, cs.SD arXiv: https://arxiv.org/abs/2606.09535v1 Multilingual ASR models such as Whisper perform well on high-resource languages but exhibit substantially higher Word Error Rates (WER) for Dravidian languages compared to Indo-Aryan ones. Through linguistic and dataset analysis, we show that Dravidian languages have longer words, higher vocabulary diversity, and lower repetition, resulting in sparse token distributions and frequent character-level substitution errors. Baseline fine-tuning further reveals decoder imbalance between self-attention (linguistic context) and cross-attention (acoustic cues). Although synthetic token-repetition experiments indicate potential gains, they are impractical. Motivated by these observations, we introduce two decoder-level enhancements: Weighted-Attention, which adaptively balances attention sources, and Self-Conditioning, which reinjects intermediate predictions to improve token consistency. Experiments demonstrate consistent WER reductions for low-resource and agglutinative languages. --- ### Interpretable Crisis Behavior Analysis Using Mobility and Social Media Data Authors: Muhammad Hamza Arshad Majeed, Sidahmed Benabderrahmane, Talal Rahwan Published: 2026-06-08 Categories: cs.CY, cs.CL arXiv: https://arxiv.org/abs/2606.09532v1 Crises alter both how people move and how they communicate. During emergencies such as wildfires and pandemics, changes in mobility patterns and online emotional discourse evolve jointly, yet they are typically studied in isolation. This paper presents a unified and interpretable pipeline that integrates mobility and social media data to identify cross-domain behavioral patterns in crisis settings. The framework is evaluated through two case studies: a short-horizon analysis of the January 2025 Los Angeles wildfires (prototype case) and a longitudinal analysis of UAE COVID-19 behavior from March 2020 to December 2021 (primary case, 671 days). The pipeline aligns heterogeneous daily signals, transforms them into binary behavioral states, applies Formal Concept Analysis (FCA) to extract co-occurrence structure, mines association rules, and validates rule stability through chronological holdout testing. A structured policy-translation layer renders robust rules as operational briefs specifying triggers, lead times, and action playbooks. Results reveal clear cross-domain behavioral structure in both crises. In the wildfire case, traffic stress, fear/anger sentiment, and governance discourse are tightly coupled within a 33-day window, with key rules reaching 100\% confidence and lift scores up to 2.5. In the COVID case, repeated mobility adaptation and sentiment volatility yield 8 stable same-day rules (88\% holdout pass rate) and 40 clean predictive rules with 2--7 day lead horizons. The work demonstrates that interpretable multimodal fusion can produce both scientifically credible and policy-actionable crisis intelligence. --- ### Hybrid Metaheuristic Combining the Dragonfly Algorithm and Tabu Search for the Traveling Salesman Problem Authors: Ammar Bouketta Published: 2026-06-08 Categories: cs.NE arXiv: https://arxiv.org/abs/2606.09529v1 The Traveling Salesman Problem (TSP) is a classical NP-hard combinatorial optimization problem that aims to find the shortest Hamiltonian cycle visiting each city exactly once and returning to the starting point. This paper proposes a hybrid metaheuristic for the TSP by combining the Dragonfly Algorithm (DA), a swarm-intelligence-based global search method, with Tabu Search (TS), a memory-based local search technique. The proposed method follows a High-Level Relay Hybridization (HRH) scheme, in which DA is first used to explore the solution space and generate a promising initial tour, while TS subsequently refines this solution through neighbourhood-based improvement and tabu memory. The hybrid approach is evaluated on standard TSPLIB benchmark instances, including burma14, att48, and ch150, and compared with standalone DA, standalone TS, and several classical metaheuristics such as Genetic Algorithm, Ant Colony Optimization, Particle Swarm Optimization, and Random Search. A systematic grid-search procedure is also conducted to study the influence of the main hyperparameters on solution quality and execution time. The experimental results indicate that the proposed hybrid can improve tour quality compared with the standalone DA and TS on the tested instances, highlighting the benefit of combining global exploration with local exploitation. However, the results also suggest that performance remains sensitive to parameter settings and problem size, motivating further validation on larger benchmarks and stronger TSP-specific baselines. --- ### Emergence of Context Characteristics Sensitivity in Large Language Models Authors: Nadya Yuki Wangsajaya, Haeun Yu, Isabelle Augenstein Published: 2026-06-08 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.09525v1 During instruction fine-tuning (IFT), large language models (LLMs) learn to follow instructions by using the provided context to answer a query. While prior work has studied how context characteristics correlate with context usage by the LLM, this analysis has been limited to inference time, leaving open how these relationships are acquired in the first place. Here, we measure how models' sensitivity to such characteristics shifts across successive IFT stages: supervised fine-tuning (SFT), direct preference optimization (DPO), and reinforcement learning with verifiable rewards (RLVR). Experiments across four models and three datasets show that SFT makes models more likely to use contexts that are easy to understand, such as containing high length, context-query similarity, and fluency. Post-SFT dynamics may either reinforce or resolve these preferences depending on the training dataset. Our findings reveal that context usage is actively reshaped at each IFT stage, and designing a balanced IFT dataset is important in ensuring robust context utilization of instruction-tuned models. --- ### Closing the Prior-Posterior Loop: Self-Reflective Molecular Design with Analysis-Driven LLM Iteration Authors: Junyi Gong, Zijie Qiu, Ben Zhong Tang Published: 2026-06-08 Categories: physics.chem-ph, cs.AI arXiv: https://arxiv.org/abs/2606.09520v2 Can a general-purpose large language model design molecules with the precision of a seasoned chemist? Current LLM-based frameworks answer this question with scalar feedback loops - generate, score, reject - that amount to informed trial-and-error. Here we show that replacing a single number with the full physicochemical rationale from first-principles calculations transforms the LLM from a stochastic sampler into a causal reasoner. Our system couples retrieval-augmented generation with a self-reflection module that feeds orbital energies, atomic charges, and electron densities - rather than compressed scores - back into the design loop. On HOMO-LUMO gap targets from 2.0 to 5.0 eV, this structure-property-relationship (SPR) reflection achieves a deviation as low as 0.0014 eV with a 100% success rate under the SPR+RAG configuration, consistently outperforming scalar-feedback and non-reflective baselines in median and mean deviation. The framework generalizes seamlessly to dipole-moment design, synthetic accessibility optimization, and molecular docking, and proves robust across 7 distinct LLM backbones. These results establish a new paradigm: when the model understands not only that a molecule fails, but why, iterative molecular design becomes genuinely mechanistic. --- ### Investigating Calibration Challenges in Probabilistic Electricity Price Forecasting Authors: Jan Niklas Lettner, Hadeer El Ashhab, Benjamin Schäfer Published: 2026-06-08 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.09517v1 As renewable energy integration increases market volatility, probabilistic electricity price forecasting has become essential for effective risk management. However, current-proper-scoring rules often prioritize forecast sharpness at the expense of calibration, leading to overconfident and statistically unreliable uncertainty estimates. This work highlights the critical gap between theoretical scoring and practical calibration, demonstrating that models can become mere proxies for deterministic forecasts when reliability is neglected. We conclude that future research must shift toward calibration-aware objectives and architectures to ensure the distributional integrity of energy market forecasts. --- ### SwiftVR: Real-Time One-Step Generative Video Restoration Authors: Jiaqi Yan, Xiangyu Chen, Xinlin Zhong, Haibin Huang, Chi Zhang, Jie Liu, Jiantao Zhou, Xuelong Li Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09516v1 Real-time video restoration (VR) for live streams requires high-resolution outputs under strict per-frame latency constraints. Existing one-step diffusion-based VR models remain difficult to deploy on consumer-grade GPUs due to two main bottlenecks: quadratic spatial attention at high resolutions and the latency-memory overhead of large video autoencoders. We present SwiftVR, a streaming one-step generative VR framework that reduces both bottlenecks under a causal chunk-wise protocol. For attention, mask-free shifted-window self-attention gathers each spatial window into a dense tensor via deterministic indexing, keeping all attention calls on the dense scaled dot-product attention path without masks, cyclic shifts, padding, or hardware-specific sparse kernels. Because SwiftVR uses only standard dense SDPA calls, the trained model transfers to consumer GPUs without retraining or custom kernels. For autoencoding, a lightweight Restoration-aware Autoencoder enables fast chunk-wise decoding while preserving reconstruction quality. On a single H100, SwiftVR sustains 31~FPS at 2560x1440 and 14~FPS at 3840x2160, whereas all compared diffusion-based VR baselines exceed the memory limit at 4K. On a consumer RTX~5090, SwiftVR reaches 26~FPS at 1920x1080. To our knowledge, SwiftVR is the first generative VR model to achieve real-time 1080p streaming on a consumer-grade GPU, while attaining strong no-reference perceptual quality with lower inference cost. Project is available at https://h-oliday.github.io/SwiftVR. --- ### BUDDY: BUdget-Driven DYnamic Depth Routing for Adaptive Large Language Model Inference Authors: Yuhua Zhou, Shaoqi Yu, Shichao Weng, Changhai Zhou, Mingze Yin, Fei Yang, Aimin Pan Published: 2026-06-08 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.09514v1 Large language models (LLMs) incur high inference cost due to their depth and parameter scale. Depth pruning can reduce latency by skipping redundant Transformer blocks, but existing methods (i) provide limited control under user-specific compute budgets and (ii) typically fix the routing path, failing to adapt as the context grows during decoding. We propose Buddy, a budget-driven dynamic depth routing framework. Buddy uses a lightweight Decision Module to score intermediate layers conditioned on the input and deterministically executes the top-k layers to satisfy a given budget. To support decode-time adaptation, Buddy reuses the first-layer KV cache as a low-overhead global context source and pools it together with the newest token representation before each routing decision. When no explicit budget is provided, an optional Budget Predictor estimates an input-dependent compute level to balance quality and efficiency. Experiments on Llama-family and Qwen models show that Buddy is competitive with strong static pruning baselines and often improves the accuracy-compute trade-off, while uniquely supporting strict budget control, decode-time rerouting, and multiple budgets within a single trained model. --- ### Securing Self-supervised Data Curation for Foundation Models Robustness Authors: Sandeep Gupta, Roberto Passerone Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09511v1 Self-supervised data curation provides a pathway to scaling and improving the generalization capabilities of machine learning models. By leveraging self-supervised learning (SSL) for data curation, the demand for massive training datasets required by foundation models can be effectively met. SSL greatly alleviates the costs associated with annotation and manual dataset curation while minimizing the need for human oversight. However, the integrity of SSL-curated datasets must be rigorously checked, as reliance on anonymous and unvetted external sources can substantially increase the risk of data poisoning. In this paper, we propose a Poisoned Data Detector (PDD), an active defense mechanism designed to ensure the integrity of SSL-curated datasets prior to foundation model training. PDDs are designed using a combination of the pretrained ImageBind model and traditional classifiers, including Random Forest (RF), k-Nearest Neighbors (KNN), Naive Bayes (NB), and Support Vector Machines (SVM). We rigorously evaluated PDDs using 176,200 images from three diverse datasets and three different adversarial attacks encompassing both in-distribution and out-of-distribution scenarios. Notably, SVM-PDD achieves superior performance for both in-distribution (Set3-Set5) and out-of-distribution (TrueFace and 140K RealFace) datasets. Our design demonstrates strong scalability and enables the rapid integration of new adversarial attack detectors through an ensemble approach. --- ### Local Search on Vertex Coloring for Bipartite Graphs Authors: Johanna Gasse Published: 2026-06-08 Categories: cs.NE arXiv: https://arxiv.org/abs/2606.09509v1 Local search is a well-known heuristic method used in optimization. In this thesis, we explore its capabilities on the vertex coloring problem, an $NP$-hard problem with relevance in both theoretical analysis and practical application. To recognize limitations in the applicability of local search of the vertex coloring problem, we analyze local search landscapes on differently-structured bipartite graphs. We identify structures that ensure only global optima can exist as well as ones that enable the existence of non-global local optima, showing that on general bipartite graphs, it is possible for local search to return arbitrarily bad results. Further, we analyze the capabilities of local search on graphs where a local optimum can be found. To do so, we introduce a gray-box local search mutation operator that removes less frequent colors with higher probability and prove that it finds an optimal coloring on complete bipartite graphs in an expected run time of $Θ(n \log n)$. This is a drastic improvement to the exponential tun time of the black-box Random Local Search, showing that gray-box mutation operators can improve the run time of local search. --- ### From Rigid to Dynamic: Entropy-Guided Adaptive Inference for Long-Context LLMs Authors: Zhanchao Xu, Haoyang Li, Qingfa Xiao, Fei Teng, Chen Jason Zhang, Lei Chen, Qing Li Published: 2026-06-08 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2606.09508v1 Existing sparse attention and KV cache compression methods for long-context LLM inference typically apply fixed sparsity patterns or uniform budgets across all attention heads, overlooking the substantial variation in attention behavior among heads and contexts. We observe two distinct entropy patterns among attention heads: Rigid Heads, whose entropy stays near zero across input segments, and Dynamic Heads, whose entropy fluctuates significantly. Crucially, the distribution of these types is context-dependent and cannot be predetermined offline. We therefore propose EntropyInfer, a training-free framework that uses attention entropy to adaptively allocate compute at the granularity of individual heads and segments during prefilling. For decoding, we introduce a latent KV cache compression scheme that leverages generated output tokens, rather than prefill tokens alone, to identify and retain the most critical cache entries. Extensive experiments on Llama, Qwen and openPangu model series show that EntropyInfer consistently outperforms baselines including SnapKV, AdaKV, and CritiPrefill, achieving up to 2.39$\times$ end-to-end speedup beyond 100k tokens with minimal quality degradation compared to full attention. The code is released in https://github.com/SHA-4096/EntropyInfer. --- ### Prisma-World: Camera-Controllable Multi-Agent Video World Model Authors: Huiqiang Sun, Zhan Peng, Size Wu, Kun Wang, Kang Liao, Dianyi Wang, Xingyu Zeng, Sheng Jin, Yangguang Li, Zhiguo Cao, Ziwei Liu, Wei Li Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09507v1 Video world models have made rapid progress in generating controllable visual experiences, but most of them still simulate the world from a single observer. Extending such models to multiple agents raises a central challenge: if each agent's future state is generated independently, overlapping views may instantiate different versions of the same scene, leading to inconsistent objects, layouts, and appearances across agents. Conventional camera conditioning controls individual trajectories, but it does not explicitly couple the generation of views that should agree under shared scene geometry. We introduce Prisma-World, a camera-controllable multi-agent world model that formulates multi-agent generation as a joint geometry-aware denoising process for cross-view consistency. Prisma-World processes all agent videos within one full-attention sequence, uses a multi-agent RoPE design to distinguish agent identities while preserving synchronized temporal coordinates, and injects relative camera geometry into attention to bias overlapping viewpoints toward shared scene evidence. To further strengthen multi-view consistency and enhance global spatial perception, we augment our framework with an overlap-decaying curriculum training paradigm alongside minimap-conditioned structural guidance. To facilitate the training and evaluation of multi-agent models, we introduce PrismaDataset, a large-scale UE5 dataset with panoramic acquisition across diverse scenes, composable multi-agent view groups with flexible agent counts and complex camera trajectories, and precise camera/action annotations for consistency training and evaluation. Experiments show that a single Prisma-World model can generate high-fidelity multi-agent videos with flexible agent numbers, camera controllability, improved cross-view consistency, and spatial grounding under minimap guidance. --- ### Deterministic Integrity Gates for LLM-Assisted Clinical Manuscript Preparation: An Auditable Biomedical Informatics Architecture Authors: Yoojin Nam, Jinhoon Jeong, Namkug Kim Published: 2026-06-08 Categories: cs.AI, cs.DL arXiv: https://arxiv.org/abs/2606.09500v4 As autonomous research agents and AI co-scientist systems push large language models (LLMs) from drafting toward end-to-end manuscript production, the bottleneck shifts from generation to verification. Fluent LLM output can hide fabricated citations, numbers that drift from source tables, and unmet reporting-guideline items; existing tools generate without verifying, and self-critique inherits the blind spots that produce confident fabrication. We describe an architecture pairing generation with verification, resting on three principles: decompose the workflow into self-contained skills, gate every stage transition with halt-on-failure, and resolve each integrity question with the cheapest sufficient mechanism, a deterministic, re-executable check where one suffices and a prose-level probe only where interpretation is unavoidable. This determinism-where-possible split, organized as an integrity-gate taxonomy, is the core contribution. It is realized as MedSci Skills, an open-source toolkit of 43 skills with a 21-detector deterministic tier, evaluated on three public-dataset pipelines (STARD, PRISMA, STROBE) and a seeded-defect ablation. Across the three pipelines every content-hash manifest verified clean and the gates surfaced real defects; on 27 identical injected defects the deterministic gates detected all 27 with no false positives on the matched clean fixtures, whereas a single-prompt LLM reviewer detected 11, its misses in code, bibliography, and style defects the prose hides. Determinism-where-possible verification yields an auditable, re-executable trail that exposes the evidence a human needs to check an LLM-assisted manuscript: feasibility and reproducibility evidence, not a claim of human-competitive quality, which a separate blinded study addresses. MedSci Skills is MIT-licensed and archived (v3.8.0). --- ### Targeting World Models to Compromise Robot Learning Pipelines Authors: Ethan Rathbun, Ahmed Agha, Saaduddin Mahmud, Christopher Amato, Alina Oprea, Eugene Bagdasarian Published: 2026-06-08 Categories: cs.RO, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2606.09499v1 World models have recently seen a rapid growth in both their popularity and capability as more data efficient tools for generating robot training data or simulating real world environments, with many works proposing their integration into the robot learning pipeline. While highly practical, in this work we demonstrate that world models introduce a uniquely stealthy and effective data poisoning entry point into the robot learning supply chain that can result in the deployment of unsafe or otherwise compromised robotic policies despite training on seemingly safe ground truth training data. In contrast to traditional data poisoning techniques which directly implant dangerous trajectories into sold or uploaded datasets, our novel attack methods inject malicious prompts or compromising transition dynamics into visibly safe teleoperated datasets which are only activated once fed through a world model as input. This can result in the generation of synthetic, dangerous robot training trajectories and subsequently unsafe or compromised robot policies. We demonstrate the effectiveness of our attacks against both state of the art action conditioned and text conditioned world models, showing a full end-to-end backdoor on a downstream DRL policy and a proof-of-concept for the VLA setting. Overall these findings necessitate research into more secure world models and reevaluating their position within the robot learning supply chain. --- ### Self-Harness: Harnesses That Improve Themselves Authors: Hangfan Zhang, Shao Zhang, Kangcong Li, Chen Zhang, Yang Chen, Yiqun Zhang, Lei Bai, Shuyue Hu Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09498v3 The performance of LLM-based agents is jointly shaped by their base models and the harnesses that mediate their interaction with the environment. Because different models exhibit distinct behaviors, effective harness design is inherently model-specific. Yet agent harnesses are still largely engineered by human experts, a paradigm that scales poorly as modern LLMs become increasingly diverse and rapidly evolving. In this paper, we introduce Self-Harness, a new paradigm in which an LLM-based agent improves its own operating harness, without relying on human engineers or stronger external agents. We operationalize Self-Harness as an iterative loop with three stages: Weakness Mining, which identifies model-specific failure patterns from execution traces; Harness Proposal, which generates diverse yet minimal harness modifications tied to these failures; and Proposal Validation, which accepts candidate edits only after regression testing. We instantiate Self-Harness across Terminal-Bench-2.0, SWE-bench Verified, and AppWorld using a minimal initial harness and three base models from diverse families: MiniMax M2.5, Qwen3.5-35B-A3B, and GLM-5. Across all nine model--benchmark combinations, every final harness improves both held-in and held-out pass rates, with overall relative gains of up to 132%. Qualitative analyses further show that the retained mechanisms address benchmark-specific bottlenecks in artifact handling and runtime control, software-patch verification, and application-state retrieval. These results suggest a path toward LLM-based agents that are not merely shaped by their harnesses, but can also participate in reshaping them. --- ### Self-Harness: Harnesses That Improve Themselves Authors: Hangfan Zhang, Shao Zhang, Kangcong Li, Chen Zhang, Yang Chen, Yiqun Zhang, Lei Bai, Shuyue Hu Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09498v1 The performance of LLM-based agents is jointly shaped by their base models and the harnesses that mediate their interaction with the environment. Because different models exhibit distinct behaviors, effective harness design is inherently model-specific. Yet agent harnesses are still largely engineered by human experts, a paradigm that scales poorly as modern LLMs become increasingly diverse and rapidly evolving. In this paper, we introduce Self-Harness, a new paradigm in which an LLM-based agent improves its own operating harness, without relying on human engineers or stronger external agents. We operationalize Self-Harness as an iterative loop with three stages: Weakness Mining, which identifies model-specific failure patterns from execution traces; Harness Proposal, which generates diverse yet minimal harness modifications tied to these failures; and Proposal Validation, which accepts candidate edits only after regression testing. We instantiate Self-Harness on Terminal-Bench-2.0 using a minimal initial harness and three base models from diverse families: MiniMax M2.5, Qwen3.5-35B-A3B, and GLM-5. Across all three models, Self-Harness consistently improves performance, with held-out pass rates increasing from 40.5% to 61.9%, 23.8% to 38.1%, and 42.9% to 57.1%, respectively. Qualitative analyses further show that Self-Harness does not simply add generic instructions, but effectively turns model-specific weaknesses into concrete, executable harness changes. These results suggest a path toward LLM-based agents that are not merely shaped by their harnesses, but can also participate in reshaping them. --- ### ContextShift: A Controlled Benchmark for Context Dependence in Object Detection Authors: Dan Zlotnikov, Alex Lazarovich, Ohad Ben-Shahar Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09495v1 Modern object detectors achieve strong performance on standard benchmarks, yet their robustness to contextual variation remains insufficiently understood. Prior evaluations largely rely on aggregate metrics such as AP on uncontrolled distribution shifts, which can obscure how performance degrades under context change. We introduce ContextShift, a controlled benchmark that systematically manipulates object--context relationships while preserving object appearance. Built on COCO 2017, it isolates context as an independent variable through geometric transformations and synthetic and natural background substitutions, including a continuous compatibility axis based on normalized pointwise mutual information (NPMI). Across diverse detector architectures, we observe a consistent degradation pattern: false negatives increase by up to 227% and prediction volume decreases by up to 44%, while false positives remain stable or decline. This suppression behavior is not captured by aggregate metrics such as AP, which can mask substantial recall loss and changes in prediction dynamics. Further analysis suggests that degradation is driven less by reduced confidence than by a reduced formation of valid detection candidates. Moreover, performance along the statistical compatibility axis is non-monotonic, peaking at intermediate NPMI and degrading toward both extremes, indicating that statistical co-occurrence does not correlate linearly with effective visual context. Finally, we show that context-aware augmentation improves robustness: every augmented variant outperforms the dataset-only baseline on both original and manipulated test images, partially recovering performance lost to prediction-suppression failures by exposing models to object--context decoupling during training. --- ### LLM-Orchestrated Conformance Checking in Stroke Care Without Computer-Interpretable Guidelines Authors: Giorgio Leonardi, Stefania Montani, Manuel Striani, Alessandro Canessa, Delfina Ferrandi Published: 2026-06-08 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.09489v1 Objective: Conformance checking in healthcare seeks to assess whether patient care pathways adhere to clinical guidelines. However, its practical application often depends on the availability of formal, machine-interpretable representations of guidelines, such as Computer-Interpretable Guidelines (CIGs), which are seldom available in real-world clinical settings. Methods: This work introduces a modular framework based on the orchestration of Large Language Models (LLMs) to support medical conformance checking directly from unstructured clinical and guideline texts, without requiring predefined CIGs. The proposed architecture integrates multiple LLMs and supporting components to extract patient traces from clinical discharge letters, identify normative rules from textual clinical guidelines, translate these rules into executable scripts, and compute a Trace Conformance Indicator to quantify compliance within the event log. Results: The framework was implemented and evaluated in the stroke care domain at the neurological ward of Alessandria Hospital. Hundreds of patient traces were automatically extracted from hospital data and assessed against 50 rules derived from the reference guideline. The analysis showed that more than 86\% of the available traces were conformant. Conclusion: The results demonstrate the feasibility of using orchestrated LLMs for practical healthcare conformance analysis. At the same time, the study provides evidence of a high level of adherence to stroke care guidelines at Alessandria Hospital. --- ### Detecting Differences Is Not Understanding Structure: Large Language Models Fail at Graph Isomorphism Authors: Kumar Thushalika, Sukumar Kishanthan, Asela Hevapathige Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09484v1 Large language models (LLMs) have shown impressive performance on diverse reasoning tasks, yet their capacity for structural reasoning in graphs remains unclear. We investigate whether LLMs can genuinely understand graph isomorphism -a fundamental problem in graph theory. While LLMs achieve near-perfect accuracy on isomorphism detection, we show this performance is illusory. When identical graphs are presented with permuted node labels, LLMs fail to identify their isomorphism. This finding suggests that LLMs exploit patterns rather than reasoning about abstract graph structure. Since permutation invariance is a fundamental requirement for valid structural reasoning, these results indicate that success on graph reasoning benchmarks should not be interpreted as evidence of genuine topological understanding. --- ### Memory Beyond Recall: A Dual-Process Cognitive Memory System for Self-Evolving LLM Agents Authors: Tianxiang Fei, Mingyang Song, Mao Zheng, Xiang Yu Published: 2026-06-08 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.09483v1 Long-term memory for an LLM agent is more than retrieving the right passage at the right time. Current memory systems collapse belief revision, causal coupling, and cross-domain abstraction into a single retrieval surface tuned for surface recall, and consequently struggle on implicit personalisation that requires reasoning over how a user has evolved. We propose DCPM, which reorganises agent memory along a cognitive capability hierarchy ascending from raw inputs and atomic facts, through diachronic belief trajectories and identity, to domain schemas, latent intentions and cross-domain patterns. The hierarchy is driven by two processes inheriting the architectural split of dual-process theory: a synchronous daytime writer (System1) that records belief revisions as doubly linked supersedes chains, and an asynchronous nighttime engine (System2) that induces schemas and intentions and sweeps for cross-domain collisions abstracted into higher-level core schemas. On LongMemEval, PersonaMem and PersonaMem-v2, enabling System2 contributes most where the benchmark rewards implicit cross-session inference (up to +5.20 on PersonaMem-v2) and least on span recall, matching the architectural prediction. --- ### Loss-Guided Adaptive Scale Refinement for Molecular Force Prediction Authors: Limin Yu Published: 2026-06-08 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.09480v1 Molecular systems involve interactions across multiple spatial scales, from local coordination and short-range perturbations to long-range electrostatic and solvent-mediated effects. However, most molecular representation learning methods rely on manually predefined scales, and the task-optimal modeling scale may not coincide with these fixed levels. This study introduces a loss-guided adaptive scale refinement framework for molecular force prediction, treating predefined scales as initial anchors and discovering task-effective resolutions through interpolation, routing, differentiable scale updates, and scale pool refinement. Using a NaCl aqueous ionic system as a minimal testbed, this study constructs short-scale and long-range force prediction branches and analyzes their complementarity. Oracle hard routing reduces the overall force MAE from 399.65 to 382.67, while continuous oracle interpolation further reduces it to 380.96. In close-contact regimes with nearest-ion distance below 0.6 nm, the close-contact MAE decreases from 327.22 to 260.51. A minimal scale pool update experiment shows that starting from endpoint anchors {0,1}, loss-guided updates automatically generate intermediate scales and recover most of the continuous oracle performance. The final updated scale pool {0,0.125,0.25,0.375,0.5,0.75,1} achieves an overall MAE of 381.23. These results support adaptive scale refinement as a promising direction for molecular representation learning, especially when fixed-scale modeling is insufficient. --- ### Optical Music Recognition for Real-World Manuscripts with Synthetic Data Authors: Jiří Mayer, Martina Dvořáková, Vojtěch Dvořák, Markéta Herzánová Vlková, Filip Bím, Pavel Pecina, Samuel Šomorjai, Petr Žabička, Jan Hajič Published: 2026-06-08 Categories: cs.CV, cs.DL arXiv: https://arxiv.org/abs/2606.09479v1 Optical Music Recognition (OMR) has seen major progress in model design, with end-to-end methods now capable of recognising notation at all levels of complexity. However, the impact of this progress has been limited by the visual domains of available training datasets, which are largely born-digital. Existing large collections of sheet music in libraries and other heritage institutions contain predominantly manuscripts, whose visual domains are highly diverse and different, so existing OMR systems fail when applied in the real world. These institutions are often resource-constrained, so large in-domain datasets cannot be expected. We provide a first baseline on real-world manuscripts with complex piano notation in the resource-constrained scenario. Using fine-grained music notation graph (MuNG) annotations and the Smashcima synthesis tool, we then show that while some direct transcriptions of in-domain data remain essential, domain adaptation using synthetic musical manuscript images brings significant improvement. Furthermore, the symbols used do not need to be in-domain, so the expensive fine-grained annotation can be avoided. We thus bring OMR closer to one of its stated goals: preserving and promoting musical cultural heritage. --- ### Efficient Minimal Solvers for Visual-Inertial Relative Pose Estimation in Multi-Camera Systems Authors: Tao Li, Zhenbao Yu, Banglei Guan, Jianli Han, Weimin Lv Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09477v1 Estimating the relative poses of multi-camera systems is a fundamental problem in computer vision, with critical applications in autonomous vehicles, mobile devices, and unmanned aerial vehicles (UAVs). However, existing solutions often suffer from high computational complexity or rely on an excessive number of point correspondences, limiting their real-world applicability. To address these limitations, we propose two efficient minimal solvers for estimating the relative poses of multi-camera systems using a novel parameterization. The first solver leverages the vertical direction prior provided by Inertial Measurement Units (IMUs), while the second utilizes the rotation axis direction prior from IMUs. Our methods require only four point correspondences and reduce the problem of multi-camera relative pose estimation to solving a univariate 6th-degree polynomial, a significant improvement over existing approaches, which typically involve 8th-degree polynomials. This reduction in computational complexity and correspondence requirements makes our solvers particularly effective when integrated into RANSAC frameworks, demonstrating strong potential for visual odometry applications. Through rigorous evaluations on synthetic data and the KITTI benchmark, our methods achieved superior computational efficiency and competitive accuracy compared to state-of-the-art algorithms. --- ### Emergent alignment and the projectability of ethical personas Authors: Guillermo Del Pinal, Youngchan Lee, Calum McNamara, Alejandro Perez Carballo Published: 2026-06-08 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.09475v2 Work on `emergent misalignment' shows that finetuning LLMs on narrow tasks can induce broadly misaligned behavior. This supports the `persona selection' (PSM) hypothesis: during pre-training, LLMs learn to simulate different characters and perspectives, which can be elicited and refined during post-training. This paper investigates the converse phenomenon, `emergent alignment', and uses it to support and refine the PSM and motivate a novel desideratum for alignment. We finetune a helpful-only model on broad and narrow safety tasks. To create SFT samples, we follow the `Constitutional AI' (CAI) approach and use four constitutions which encode reasonable alignment strategies: deontology, consequentialism, virtue ethics, and aligning AIs as subordinate to human authority. For each of those models, we show that finetuning on two narrow safety sub-categories reliably induces emergent alignment over a representative set of general safety categories, and on safety subcategories that we directly filtered-out of the data sets used for narrow alignment. To test the `PSM' using a more fine-grained evaluation, we used a multidimensional `ethical persona' diagnostic. For each constitutionally finetuned (broad/narrow) model, we evaluate how well their behavior matches their expected signature profile. Our results show that our CAI models acquire their expected ``ethical persona'' -- e.g., the model narrowly fine-tuned on SFT samples created using the consequentialist constitution agrees significantly more with utilitarian than deontological beliefs. Yet our coarse and fine-grained evaluations show that there are significant differences across our (broad/narrow) finetuned CAI models in how well they project. We conclude that alignment strategies should be evaluated, not just on their (in-distribution) general safety performance, but also specifically on their degree of projectability. --- ### Training-Free Generalized Few-Shot Segmentation through Open-Vocabulary Semantic Arbitration Authors: Silas Kwabla Gah, Ebenezer Owusu Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09474v1 Generalized Few-Shot Semantic Segmentation (GFSS) has traditionally been approached as a representation-learning problem, requiring task-specific adaptation to incorporate novel classes from limited support examples. Recent foundation models, however, already exhibit strong open-vocabulary recognition and segmentation capabilities, raising a different question: can GFSS be solved through inference-time coordination of frozen semantic priors rather than parameter adaptation? We answer this question with Open-V, a training-free GFSS framework that combines Segment Anything (SAM3) Promptable Concept Segmentation (PCS) with a K-shot CLIP support centroid through calibrated per-pixel semantic arbitration. OpenV introduces no trainable components and supports arbitrary semantic categories at inference time. Beyond segmentation performance, our study contributes three broader findings. First, we show that support information can be incorporated through inference-time semantic grounding, and that its contribution increases as foundation-model text priors weaken on label-disjoint vocabularies. Second, we identify a reproducibility confound in foundationmodel segmentation, demonstrating that preprocessing and evaluation-space mismatches can silently distort reported performance. Finally, we validate Open-V across PASCAL5i, COCO-20i, and ADE-OW, showing that training-free coordination of foundation-model priors generalizes across both conventional GFSS and open-vocabulary evaluation settings. On PASCAL-5i (1-shot), Open-V attains base/novel/harmonic mIoU of 78.4/77.5/77.9, without GFSS-specific training surpassing the strongest trained baseline by +17.7 HM. --- ### Report the Floor: A Training-Free Conformal Interval Is a Mandatory Baseline for Probabilistic Time-Series Forecasting Authors: Valery Manokhin Published: 2026-06-08 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2606.09473v1 Probabilistic forecasters are increasingly learned, yet the baselines they are compared against are often weak or omitted. We show that the simplest possible conformal interval - a last-value point forecast wrapped in a finite-sample split-conformal residual quantile, with no parameters and no training - is a far stronger baseline than its near-total absence from recent learned-forecasting and conformal-time-series comparisons would suggest. In one-step-ahead online forecasting across 2,217 real series from nine public sources (Monash, LOTSA, the LTSF traffic/electricity/weather suites, METR-LA, BOOM, nips/probts), this ConformalNaive interval decisively beats the naive value-quantile baselines, the entire NPTS family (NPTS 73%, SeasonalNPTS 64% of series), and the published Conformal Seasonal Pools (CSP) method (71% of series, bootstrap 95% CI (69,73), paired Wilcoxon p approx 7.6e-135); it is on par with the simpler learned conformal predictors (RCI, quantile regression; median relative Winkler within 2%) and is beaten only by the adaptive-online and ensemble methods (SPCI, ACI, AgACI), which track distribution shift and lead by 9-33% relative Winkler. It is also better calibrated than a trained neural forecaster: on the six datasets that introduced DeepNPTS, the trivial floors cover the truth 84-85% of the time at a nominal 95%, versus DeepNPTS's 66%. At multi-step seasonal horizons the picture inverts: the random-walk floor is the weakest method and the seasonal pool (CSP) wins - a boundary we map. Finally we give ConformalNaive+, a one-line, training-free, horizon-adaptive selector that attains the better of two complementary floors at every horizon with restored coverage. We argue the matching conformal naive floor must be a mandatory baseline whenever a learned probabilistic forecaster claims gains. --- ### Escaping the KL Agreement Trap in On-Policy Distillation Authors: Haoran Xin, Anhao Zhao, Ying Sun, Jin Li, Xiaoyu Shen, Hui Xiong Published: 2026-06-08 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2606.09471v1 On-policy distillation (OPD) provides dense token-level supervision by asking a teacher to score student-generated rollouts. However, when the student drifts into an unrecoverable prefix, the teacher may locally agree with the degraded state, producing low reverse KL but little corrective training signal. We identify this persistent regime as a low-KL agreement trap. Further analyses show that tokens during and after such traps produce less useful supervision signals. We propose KAT (KL Agreement Trap Termination), an online OPD termination rule that detects persistent low-KL agreement with a dynamic training-adaptive threshold. By filtering weak supervision from degenerate agreement, KAT improves avg@k accuracy by 2.66% and pass@k by 3.43% across four mathematical benchmarks, while reducing average rollout length by 59.73%. --- ### A Finetuned SpeechLLM for Joint Multi-Granular L2 Assessment and Natural-Language Rationales Authors: Aditya Kamlesh Parikh, Cristian Tejedor-Garcia, Catia Cucchiarini, Helmer Strik Published: 2026-06-08 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.09470v1 Automated L2 speech assessment can assign proficiency labels, but often lacks interpretability. We propose a rubric-guided SpeechLLM for multi-aspect, multi-granular assessment, trained with a hybrid objective combining supervised fine-tuning and Bounded Direct Preference Optimization. The model jointly predicts ordinal labels at the sentence-level (accuracy, fluency, prosody), word/phoneme-level accuracy, and generates a natural-language rationale in the same response. On SpeechOcean762, our approach matches or outperforms single-granularity models while remaining competitive with prior approaches. We analyze rationale reliability along two axes: self-consistency with model predictions and alignment with ground-truth labels, using sentiment consistency (plausibility) and mention-based agreement (faithfulness). Rationales are plausible at the sentence level, but faithfulness degrades at the word/phoneme level: references are sparse and weakly aligned with token-level labels. --- ### DECSELFMASK: Leveraging Unlabeled Text via Self-Relevance-Guided Masking for Decoder-Only Classification Authors: Pietro Ferrazzi, Matteo Merler, Giovanni Bonetta, Alberto Lavelli, Bernardo Magnini Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09466v3 Classification tasks require annotated data, which can often be expensive, time-consuming, or even unfeasible to collect. This is the case of the medical domain, where large datasets often have few annotated examples. To address this, we propose DecSelfMask (Decoder Self-learning by Masking), an approach to enhance decoder-only performance on classification tasks. We build on common self-learning approaches by leveraging a model to create training examples from unlabeled data, and propose a novel relevance-guided masking strategy. We use relevance attribution methods to determine what portions of unannotated texts are relevant for a task. We then create self-supervised training examples by masking out those portions, training the model to reconstruct them via next-token-prediction. We hypothesize that those examples convey knowledge about the structure and semantics of unannotated data that can be useful for downstream performance. We test our approach on 136 tasks from a collection of 1.9M clinical notes from an Italian hospital. We quantify DecSelfMask's impact on downstream tasks on 5 models of different scales and families, including a probing analysis. Experiments show consistent gains, outperforming the base models (+9.1 points in Macro F1), continual pretraining (+6.3), synthetic label generation (+12.5), as well as common baselines. The results show that relevance attribution can serve as a powerful source of supervision for decoder learning from unlabeled data, beyond existing methods. --- ### DECSELFMASK: Leveraging Unlabeled Text via Self-Relevance-Guided Masking for Decoder-Only Classification Authors: Pietro Ferrazzi, Matteo Merler, Giovanni Bonetta, Alberto Lavelli, Bernardo Magnini Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09466v2 Classification tasks require annotated data, which can often be expensive, time-consuming, or even unfeasible to collect. This is the case of the medical domain, where large datasets often have few annotated examples. To address this, we propose DecSelfMask (Decoder Self-learning by Masking), an approach to enhance decoder-only performance on classification tasks. We build on common self-learning approaches by leveraging a model to create training examples from unlabeled data to propose a novel relevance-guided masking strategy. We use relevance attribution methods to determine what portions of unannotated texts are relevant for a task. We then create self-supervised training examples by masking out those portions, training the model to reconstruct them via next-token-prediction. We hypothesize that those examples convey knowledge about the structure and semantics of unannotated data that can be useful for downstream performance. We test our approach on 136 tasks from a collection of 1.9M clinical notes from an Italian hospital. We quantify DecSelfMask's impact on downstream tasks on 5 models of different scales and families, including a probing analysis. Experiments show consistent gains, outperforming standard supervised fine-tuning approaches (+19.9 points in Macro F1), synthetic label generation (+12.5), and continual pretraining (+6.3), as well as common baselines. --- ### H2HMem: A Multimodal Memory Benchmark for Agents in Human-Human Interactions Authors: Shiping Zhu, Yibo Yang, Zhengyang Wang, Tiancheng Shen, Dandan Guo, Ming-Hsuan Yang Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09461v1 Large language model agents are increasingly deployed in human-human interaction settings, such as meeting assistants and clinical documentation systems, where they must observe conversations and retain information for downstream queries. Unlike traditional human-assistant settings, these environments are inherently multimodal, involve complex discourse phenomena such as anaphora and deixis, and contain asynchronous or conflicting information from multiple participants. However, existing memory benchmarks largely focus on single-user, text-only interactions, failing to capture these challenges. To address this gap, we introduce H2HMem, a Human-to-Human Multimodal Memory Benchmark for evaluating memory capabilities in complex human-human interactions. H2HMem includes both dyadic and multi-party conversations with multimodal information streams, and evaluates agents along three dimensions: memory recall, reasoning, and application. Experiments with advanced agents reveal substantial limitations in constructing, retaining, and utilizing memories across modalities, participants, and sessions, highlighting substantial room for improvement in next-generation LLM agents. --- ### AbstRAG: Learning to Abstract for Retrieval Problems Authors: Lei Xu, Xin Quan, Daniel Pedronette, André Freitas Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09459v1 Retrieval-augmented generation often fails when the query, the document evidence, and the user's intent are expressed at different levels of abstraction. A query may ask about a class, a relation, or an event, while the document only states specific instances, indirect framings, or scoped formulations. We define this mismatch as an abstraction gap: the minimal set of typed assumptions required to align query intent with the available evidence. To close this gap, we introduce AbstRAG, which treats abstraction as an explicit retrieval object. AbstRAG decomposes the query--evidence gap into expression, conceptual, intent--evidence, and event-type components, and scores relevance by combining match quality, a query-independent utility prior, and the cost of the required bridges. Its central mechanism is reflective refinement: a critic diagnoses retrieval failures, localizes the failed abstraction operator, proposes a minimal stage-specific patch, and accepts the patch only under sufficiency and compression controls. Across three within-document retrieval benchmarks against seven baselines, AbstRAG outperforms on nDCG@10 in 18 of 21 paired-bootstrap contrasts and improves generation accuracy by 1.9%, 5.2%, and 4.0% across the three benchmarks; ablations confirm that reflective refinement drives most of the retrieval gain and the compression control alone reduces over-expansion false positives from 73.7% to 0% on a stress slice. --- ### Breaking the Tokenizer Barrier: On-Policy Distillation across Model Families Authors: Yifan Niu, Han Xiao, Dongyi Liu, Zelong Wang, Dihong Gong, Yasheng Wang, Jia Li Published: 2026-06-08 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.09456v1 On-Policy Distillation (OPD) has become a core technique in the post-training of Large Language Models (LLMs) for transferring knowledge from domain experts to student models. However, existing OPD distillation methods require teacher and student models to share the same tokenizer, restricting the applicability of OPD within the model series. Current mainstream practice typically employs Supervised Fine-Tuning (SFT) on teacher-generated responses for cross-tokenizer distillation, which fails to capture the rich knowledge embedded in the teacher's probability distribution. In this work, we enable the standard on-policy distillation method to operate across model families, ensuring that high-fidelity token-level signals can propagate across different tokenizers with a precise token-mapping algorithm. Extensive experiments show that cross-tokenizer OPD is significantly more compute-efficient than baselines on various benchmarks. Our results unlock a broader range of teacher-student pairs for OPD, opening up new avenues for adapting and enhancing interactions between LLMs. --- ### GD-MIL: Grade-Disentangled Multiple Instance Learning for Multimodal Biochemical Recurrence Prediction in Prostate Cancer Authors: Dasari Naga Raju Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09453v1 Biochemical recurrence (BCR) after radical prostatectomy is a critical endpoint in prostate cancer, yet risk stratification relies almost entirely on variables dominated by Gleason grade. Whether H&E whole slide images (WSIs) carry prognostic signal beyond grade, and whether multiple instance learning (MIL) can recover it, remains unsettled. A key obstacle is that many pipelines select model checkpoints on the evaluation fold, artificially inflating concordance. We construct a rigorous benchmark on TCGA-PRAD (487 patients, 101 BCR events) using strict out-of-fold scoring over five-fold cross-validation repeated across five seeds. The choice of MIL aggregator (ABMIL, CLAM, TransMIL, PatchGCN) has little effect (C-index 0.61-0.64 with UNI2-h), while the feature extractor is the dominant factor (ResNet50 0.566 versus pathology foundation models up to 0.639). A clinical Cox model on grade, stage, and age reaches 0.687; no imaging-only model significantly outperforms it (p > 0.10). We introduce Grade-Disentangled MIL (GD-MIL), a gated-attention MIL encoder trained with a gradient-reversal grade adversary that encourages the slide representation to be invariant to Gleason grade before late fusion with clinical variables. GD-MIL achieves C-index 0.704, significantly outperforming both the clinical baseline (delta-c = +0.029, p = 0.0005) and the best imaging-only model (delta-c = +0.062, p = 0.039), suggesting H&E morphology contains prognostic information complementary to grade. A median risk split yields log-rank p < 0.0001 separation in BCR-free survival (~20% vs ~70% at five years). --- ### Dense Force Estimation with an Event-based Optical Tactile Sensor Authors: Agis Politis, René Zurbrügg, Valentina Cavinato Published: 2026-06-08 Categories: cs.RO, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.09451v1 Humans rely on spatially dense, geometry and force-aware tactile feedback at high temporal resolution for dexterous manipulation. While vision-based tactile sensors enable dense force estimation, they are limited by camera frame rates, motion blur, and data bandwidth. Event-based optical tactile sensors offer an attractive alternative with microsecond temporal resolution and low motion blur, but existing methods are restricted to predicting only net forces. We introduce the first framework for dense 3D force field reconstruction using event-based optical tactile sensors. Our approach estimates 3D surface displacements from event data and maps them to forces via the inverse Finite Elements Method (iFEM). Shear displacements are recovered through the proposed event-based marker tracking algorithm, while normal displacements are predicted by a convolutional neural network trained on a collected dataset of synchronized force-displacement-event data. Experiments demonstrate accurate reconstruction of physically grounded forces, achieving a mean absolute error of (0.14 N, 0.10 N, 0.93 N) over force ranges up to (4 N, 4 N, 20 N), while operating at an average of 100 Hz. This work constitutes a first step toward enabling dense force feedback for high-frequency control in robotic grasping and dexterous manipulation. --- ### TheoremBench: Evaluating LLMs on Theorem Proving in Formal Mathematics Authors: QuocViet Pham, Elvir Karimov, Andrey Galichin, Ivan Oseledets Published: 2026-06-08 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.09450v1 LLMs have recently achieved strong results on formal proving benchmarks. However, existing evaluations remain heavily concentrated on competition-style problems and often fail to capture how models behave on longer, more dependency-rich mathematical developments. We introduce TheoremBench, a Lean4 benchmark designed to evaluate theorem provers beyond contest settings. The benchmark is built from nearly one hundred classical theorems and is released in two complementary forms: a plain main version containing one target theorem per instance, and a premised version that expands each theorem into a structured family of related proving tasks consisting of the main theorem together with automatically extracted supporting subtheorems. This design enables evaluation of not only whether the final theorem was proved from scratch, but also of partial progress through the internal proof structure of a theorem. Our experiments show that explicit premises substantially improve performance for Lean4-capable prover models. To provide a comprehensive evaluation, we introduce theorem-level coverage and token-efficiency metrics that expose qualitative differences in proof behavior. The results show that current provers remain strongly biased toward easy subtheorems and often solve theorems through long and inefficient tactic traces rather than compact proof plans. TheoremBench therefore provides a more fine-grained view of formal reasoning ability and highlights the importance of structural benchmark design for evaluating Lean4 theorem provers. --- ### Reasoning without Gold Standards: A Proxy-Judge Theory of Autoformalization Authors: Lei Xu, Xin Quan, André Freitas Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09449v1 Complex reasoning tasks increasingly require systems to produce outputs whose correctness cannot be judged by exact match against a single reference. Autoformalization (AF) is a representative example; it asks a model to translate informal mathematical or logical reasoning into a formally checkable object, yet expert-validated formalizations do not scale beyond toy cases and a single informal argument can admit many valid formal renderings. Progress therefore depends on whether partial, structured proxies can substitute for exact references. We introduce a reference-free proxy-judge framework for AF that replaces gold-standard matching with a vector of per-axis property checks. The framework organizes the proxy along three structural scopes that cover global properties of the elicited object, per-module properties internal to its sub-components, and cross-domain properties that re-align it to the informal source, and aggregates each axis into a verdict vector. The vector drives a reflective refinement loop in which a violated coordinate routes the controller to a matching repair target, so each iteration changes only what is judged wrong. Under bounded judge noise, the expected intrinsic gap contracts geometrically to a noise-dependent plateau. Across seven formalization backbones on miniF2F, ProofNet, e-SNLI, and ProntoQA, refinement consistently lifts Pass Rate over the single-shot ICL baseline, and the per-axis proxy outperforms a matched scalar proxy on benchmarks where the baseline has room to improve. Structured proxy judgments therefore provide both a practical refinement signal and a theoretical handle on convergence when exact references are unavailable. --- ### AliyunConsoleAgent: Training Web Agents in Real-World Cloud Environments via Distillation and Reinforcement Learning Authors: Bojie Rong, Zheyu Shen, Qiaoping Wang, Pengfei Kang, Yang Xu, Yawen Wei, Hanyu Wu, Zhi Zhao, Leihao Pei, Linquan Jiang Published: 2026-06-08 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.09447v1 We present AliyunConsoleAgent, a web agent framework for automated documentation verification in real-world cloud consoles. Major cloud platforms encompass hundreds of products with rapid feature iteration, causing console UIs to frequently diverge from their corresponding documentation. Verifying that documented procedures accurately reflect the current console and can be executed end-to-end demands an estimated 4 million recurring inspections annually, yet manual coverage remains below 1%. While agent systems built on frontier proprietary models achieve high success rates, their prohibitive cost and data privacy constraints preclude large-scale deployment. We propose a two-stage training paradigm: supervised fine-tuning (SFT) on distilled frontier-model trajectories, followed by reinforcement learning using Group Relative Policy Optimization (GRPO) and a dual-channel outcome reward model in real cloud environments. To support large-scale RL training, we construct a high-determinism rollout system featuring Terraform-based resource pre-provisioning and LLM-driven on-demand provisioning, which effectively isolates environment noise from the training signal. We further introduce a rule-based reward evaluation protocol grounded in backend audit logs, providing objective, reward-hacking-resistant outcome judgment. Our model evolves from mechanical instruction following to autonomous decision-making with cloud console and product-specific understanding. Experiments on a challenging 278-task benchmark where the best frontier model achieves only 65.34% demonstrate that AliyunConsoleAgent-32B achieves a 63.52% mean success rate -- a 20.24 percentage-point improvement over the base model, narrowing the gap to the best frontier proprietary model to 1.82 pp (bootstrap 95% CI (-1.27, 7.39)) -- at 92% lower inference cost. --- ### Leveraging Morphology for Historical Script Metrological Analysis Authors: Malamatenia Vlachou Efstathiou, Raphaël Baena, Dominique Stutzmann, Mathieu Aubry Published: 2026-06-08 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.09446v1 Advances in handwritten text recognition have enabled large-scale transcription of historical documents, but still provide limited access to interpretable visual measurements for paleography, the study of historical scripts. In this paper, our main insight is that morphological script analysis, in particular the capacity to learn character prototypes from line-level transcriptions, enables the definition of scalable, meaningful, and stable paleographic measurements. More precisely, we leverage a transformer-based detection architecture together with a prototype-based line reconstruction module to learn prototypical characters and their occurrence, deformation, and positioning. Our contributions are twofold. First, we introduce a deep architecture and learning methodology that enables efficient character modeling with only line-level transcription supervision, significantly improving over the Learnable Typewriter baseline and enabling accurate character bounding box prediction, unlocking its potential for paleographic measurements. Second, we introduce and demonstrate the paleographical relevance of automatic measurements enabled by our architecture for characters, bi-grams, and spaces between graphical units. For this demonstration, we extend the annotations of the codex Paris, BnF, fr. 2813, commissioned in the late fourteenth century by Charles V and copied by four hands, to 160 pages. We visualize our measurements over these pages, showing how they enable us not only to differentiate graphical profiles, but also to discover and analyze subtle variations. This case study outlines the scalability of our approach and its frugality in terms of required training data, since a single column of text is sufficient to compute our measurements on each of the 160 pages. Data and code are publicly available at: https://malamatenia.github.io/morphology4metrology-analysis. --- ### SIFT: Selective-Index For Fast Compute of RAG Prefill by Exploiting Attention Invariance Authors: Rya Sanovar, Srikant Bharadwaj, Hritvik Taneja, Moinuddin Qureshi Published: 2026-06-08 Categories: cs.AI, cs.AR arXiv: https://arxiv.org/abs/2606.09441v1 Retrieval-Augmented Generation (RAG) injects LLM queries with relevant documents to improve response quality. This injection increases prompt length and slows time to first token (TTFT). Unlike standard queries, RAG queries have a unique property of context reuse where the same documents recur across user queries. Thus, fully recomputing documents for every RAG query does redundant compute and increases TTFT. Prior works precompute KV tensors of RAG documents offline and coarsely recompute some tokens during online prefill. However, such KV reuse is often slower than full recomputation on modern GPUs due to high-latency disk transfers. Further, such a coarse-grained recomputation degrades accuracy. To address these limitations, this paper proposes SIFT: Selective-Index For Fast Compute of RAG Prefill by Exploiting Attention Invariance. SIFT processes documents offline and extracts fine-grained locations of high attention scores for each document. Next, we identify the following attention invariance insights that enable us to exploit the extracted locations during runtime: (1) Local-Attention Invariance: The location of high attention scores within a document remain invariant to surrounding documents. This helps us predict the location of high scores where the document attends to itself. (2) Cross-Attention Consistency: Keys with high intra-document attention also attract cross-attention from subsequent documents. This helps us predict the location of high scores where the document attends to future documents. Critically, SIFT stores no KV data and only stores locations of high scores in the form of two compact bit vectors. SIFT's storage is up to 24,000x smaller than KV tensors, obviating costly disk transfers. During prefill, SIFT computes the attention only for the marked locations and improves TTFT by 1.71x while holding accuracy within 1% of full recompute. --- ### MUDIDI: A Two-Stage Framework for Multilingual Dictionary Digitization with Language Models Authors: David Setiawan, Temuulen Khishigsuren, Milind Agarwal, Pagnarith Pit, Aso Mahmudi, Ekaterina Vylomova Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09435v2 Multilingual dictionaries are among the most valuable documentary resources for low-resource and endangered languages, yet many remain available only as scans. For many decades, their digitization and conversion into a machine-readable format was nearly impossible due to language-specific scripts, complex multi-column layouts full of entries with abbreviations and cross-references. Recent vision-language models offer a promising solution, but it is unclear how well they preserve characters, markup, and process lexicographic structure. We introduce MUDIDI, a two-stage framework for multi-lingual dictionary digitization. Stage One evaluates the quality of character recognition and markup preservation; Stage Two focuses on dictionary entry segmentation with subsequent mapping into a machine-readable lexicographic schema, SIL's Multi-Dictionary Formatter. We also release a dataset that consists of human-annotated lexicographic entries collected from 30 public-domain dictionaries featuring diverse writing systems, language families, and formats. We benchmark OCR systems, general-purpose Large Language Models (LLMs), and Vision Language Models (VLMs) on the dataset, demonstrating superior performance of LLMs across most writing systems and languages in both stages, and provide practical guidelines on improving the results for more challenging scenarios. Finally, we show that supplementing additional information, such as dictionary introduction, to the LLMs can improve the quality of the digitized dictionary. Github: https://github.com/naarm-nlp/mudidi --- ### MUDIDI: A Two-Stage Framework for Multilingual Dictionary Digitization with Language Models Authors: David Setiawan, Temuulen Khishigsuren, Milind Agarwal, Pagnarith Pit, Aso Mahmudi, Ekaterina Vylomova Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09435v1 Multilingual dictionaries are among the most valuable documentary resources for low-resource and endangered languages, yet many remain available only as scans. For many decades, their digitization and conversion into a machine-readable format was nearly impossible due to language-specific scripts, complex multi-column layouts full of entries with abbreviations and cross-references. Recent vision-language models offer a promising solution, but it is unclear how well they preserve characters, markup, and process lexicographic structure. We introduce MUDIDI, a two-stage framework for multi-lingual dictionary digitization. Stage One evaluates the quality of character recognition and markup preservation; Stage Two focuses on dictionary entry segmentation with subsequent mapping into a machine-readable lexicographic schema, SIL's Multi-Dictionary Formatter. We also release a dataset that consists of human-annotated lexicographic entries collected from 30 public-domain dictionaries featuring diverse writing systems, language families, and formats. We benchmark OCR systems, general-purpose Large Language Models (LLMs), and Vision Language Models (VLMs) on the dataset, demonstrating superior performance of LLMs across most writing systems and languages in both stages, and provide practical guidelines on improving the results for more challenging scenarios. Finally, we show that supplementing additional information, such as dictionary introduction, to the LLMs can improve the quality of the digitized dictionary. Github: https://github.com/DavidSamuell/MUDIDI-Pipeline-for-Digitization-of-Multilingual-Dictionary/ --- ### A transition-density-based operator learning method for Fokker-Planck equations with various initial conditions Authors: Li Zeng, Xiaoliang Wan, Yaobin Wang, Fabio Nobile, Tao Zhou Published: 2026-06-08 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.09434v2 Solving Fokker-Planck equations (FPEs) for multiple initial conditions typically requires repeated computations, leading to substantial computational costs. In this work, we propose a transition-density-based operator learning method to efficiently approximate the solution operator of FPEs with various initial conditions. The core idea is to learn the transition probability density function (PDF) of the underlying stochastic differential equation (SDE), from which the solution associated with a new initial distribution can be obtained through the Chapman-Kolmogorov equation without retraining the model. A major challenge in learning the transition PDF lies in the singular behavior induced by the Dirac initial condition. To address it, we introduce a conditional normalizing flow whose base distribution is given by the explicit transition PDF of a linearized SDE. This base distribution captures the short-time behavior of the target transition PDF and allows the normalizing flow to learn a near-identity transformation at small times. We further incorporate a time-weighted loss function to stabilize training near the initial time and develop an importance-sampling strategy for evaluating solutions associated with general initial conditions. A variety of numerical experiments are presented to illustrate the effectiveness and robustness of the proposed method. --- ### Bayesian Selective Latent Inference for Wastewater-First Influenza Monitoring Authors: Yixuan Zhang, Yang Song, Hao Wang, Samir Bhatt, Hengguan Huang Published: 2026-06-08 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.09433v1 Wastewater influenza surveillance can reveal community circulation before clinical reporting, but wastewater alone is not a fully identifiable proxy for human burden. Existing wastewater models assume a fixed evidence set, while generic evidence-acquisition methods treat official surveillance streams as interchangeable costly features. We cast wastewater-first influenza monitoring as a selective decision problem: starting from mandatory wastewater evidence, the system must decide whether wastewater is sufficient, which delayed official stream to query next, and when abstention is the only scientifically defensible action under source ambiguity. We propose Bayesian Selective Latent Inference (BSLI), a principled Bayesian method that maintains a posterior over latent burden and identifiability, certifies answerability through explicit scientific gates, and optimizes query-stop decisions with an exact cost-calibrated Bellman policy. We prove the key variational, answerability, Bellman-optimality, and one-dimensional cost-calibration properties. On a fixed public-data benchmark with 5,933 forecasting episodes and 3,102 source-ambiguity episodes, BSLI improves the matched-budget cost-performance frontier while preserving conservative abstention under source ambiguity. --- ### Graph Mamba Operator: A Latent Simulator for Interacting Particle Systems Authors: Karn Tiwari, Niladri Dutta, N M Anoop Krishnan, Prathosh A P Published: 2026-06-08 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.09432v1 Modeling interacting dynamical systems requires capturing spatial interactions alongside long-range temporal dependencies. Graph neural networks (GNNs) provide a natural representation but typically rely on autoregressive rollouts and treat spatial and temporal dynamics separately, leading to error accumulation over long horizons. Existing approaches also focus on local interactions and short temporal contexts, limiting their ability to capture multi-hop dependencies and global structure. We introduce the Graph Mamba Operator (GraMO), a latent-space simulator that integrates state-space models with graph-based interaction learning. In contrast to prior work that sequences nodes or applies spatial and temporal updates in separate stages, GraMO couples graph-based interactions and temporal state updates within a single recurrence. The update is linear in the latent state, with input-dependent coefficients that adapt across regimes. We evaluate GraMO on N-body systems, motion capture, and robotics datasets, achieving the lowest error across benchmarks and the largest gains in long-horizon prediction. --- ### LargeMonitor: Monitoring Online Task-Free Continual Learning via Large Pretrained Models Authors: Mingqi Yuan, Xiaoquan Sun, Shihao Luo, Jiayu Chen Published: 2026-06-08 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.09430v2 Online task-free continual learning (TFCL) requires intelligent agents to sequentially accumulate knowledge from an unbounded, non-stationary data stream under strict single-pass constraints and without any explicit task identifiers. Existing online TFCL paradigms primarily rely on parameter-efficient prompt tuning or dynamic structure expansion driven by training-coupled optimization dynamics, such as empirical loss fluctuations or evolving latent distances. As a result, these training-coupled solvers remain agnostic to the structural origins of distribution drift, mechanically enforcing a fixed strategy across fundamentally distinct streaming variations. To address this gap, we propose LargeMonitor, a framework that leverages large pretrained foundation models to autonomously orchestrate task-free continuous adaptation. Specifically, LargeMonitor introduces a decoupled detection module utilizing the frozen, stable representation space of large vision models (LVMs) to achieve robust, zero-shot drift detection without training-dependent interference or brittle threshold tuning. Upon a confirmed drift, the framework activates a context-aware diagnostic module driven by large multimodal models (LMMs) to interpret the precise semantic etiologies of the stream variation (e.g., novel class emergence vs. environmental domain shift). This dual-stage capability empowers the continuous learner to dynamically deploy adaptive and shift-specific optimization strategies. Extensive experiments across multiple TFCL settings and benchmarks demonstrate that LargeMonitor achieves precise, robust detection and diagnosis of complex data streams while consistently improving the performance of existing online TFCL algorithms. --- ### LargeMonitor: Monitoring Online Task-Free Continual Learning via Large Pretrained Models Authors: Mingqi Yuan, Xiaoquan Sun, Shihao Luo, Jiayu Chen Published: 2026-06-08 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.09430v1 Online task-free continual learning (TFCL) requires intelligent agents to sequentially accumulate knowledge from an unbounded, non-stationary data stream under strict single-pass constraints and without any explicit task identifiers. Existing online TFCL paradigms primarily rely on parameter-efficient prompt tuning or dynamic structure expansion driven by training-coupled optimization dynamics, such as empirical loss fluctuations or evolving latent distances. As a result, these training-coupled solvers remain agnostic to the structural origins of distribution drift, mechanically enforcing a fixed strategy across fundamentally distinct streaming variations. To address this gap, we propose LargeMonitor, a framework that leverages large pretrained foundation models to autonomously orchestrate task-free continuous adaptation. Specifically, LargeMonitor introduces a decoupled detection module utilizing the frozen, stable representation space of large vision models (LVMs) to achieve robust, zero-shot drift detection without training-dependent interference or brittle threshold tuning. Upon a confirmed drift, the framework activates a context-aware diagnostic module driven by large multimodal models (LMMs) to interpret the precise semantic etiologies of the stream variation (e.g., novel class emergence vs. environmental domain shift). This dual-stage capability empowers the continuous learner to dynamically deploy adaptive and shift-specific optimization strategies. Extensive experiments across multiple TFCL settings and benchmarks demonstrate that LargeMonitor achieves precise, robust detection and diagnosis of complex data streams while consistently improving the performance of existing online TFCL algorithms. --- ### Guide Me Out: A Framework to Benchmark VLM Operators Communication in Crisis Scenarios Authors: Giacomo Gonella, Stefano Menini, Marco Guerini Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09428v1 Effective crisis response requires spatially grounded communication that bridges linguistic guidance of civilians with the physical environment, accounting for structural bottlenecks, evolving threats, and agent-specific contexts. Yet, current NLP research in crisis communication remains mainly limited to static, text-only classification settings, overlooking the critical communicative role of AI operators in dynamic, embodied scenarios. We address this gap with a novel benchmarking framework for evaluating Vision-Language Models (VLMs) tasked with guiding civilian agents through simulated evacuations. We test two communication strategies (narrowcast vs. broadcast), two environment representations (visual vs. graph-based), and two threat behaviors (static vs. moving) across nine maps of varying structural complexity. Our results show that Narrowcast consistently reduces civilian Fail rates compared to Broadcast across all difficulty levels. Guidance quality depends heavily on how the VLM operator represents the world: the visual modality drives performance, while adding an adjacency graph is model-dependent and often harmful. Moving threats raise Fail rates across all conditions as communication must continuously adapt over time. Together, these findings show that deploying VLMs as AI operators in evacuation scenarios remains a non-trivial challenge, where the choice of communication strategy and input representation can directly determine the success or failure of the intervention. --- ### WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces Authors: Wanli Li, Bowen Zhou, Yunyao Yu, Zhou Xu, Yifan Yang, Dongsheng Li, Caihua Shan Published: 2026-06-08 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.09426v3 Computer-use agents (CUAs) increasingly operate in runtimes that combine visual desktop control, command-line execution, code editing, browsers, and external tools. Existing benchmarks, however, often evaluate these interfaces as separable capabilities, leaving long-horizon cross-interface orchestration under-tested. Thus, we introduce WeaveBench, a long-horizon hybrid-interface benchmark with 114 tasks across 8 real-world work domains, grounded in real user requests and publicly verifiable artifacts. Each task requires agents to combine GUI observations/actions with CLI/code operations within a single trajectory. We evaluate these tasks on a real Ubuntu desktop inside deployed CLI-agent runtimes, augmented with a minimal desktop-control plugin. We also propose a companion trajectory-aware judge that inspects deliverables, files, screenshots, logs, and action traces, while detecting shortcut behaviors such as fabricated visual evidence or hard-coded metrics. Across frontier model-runtime pairings, the best PassRate reaches only 41.2%, showing the benchmark remains far from saturated. The trajectory-aware judge further reveals that outcome-only grading substantially overestimates agent performance. Overall, WeaveBench exposes a critical gap in CUA evaluation and provides an effective testbed to measure whether agents can orchestrate GUI, CLI, and code operations across long-horizon real-world tasks. --- ### Toward Signing Activity Projection in Sign Language Interaction Authors: Takao Obi, Wang Yusong, Koji Inoue, Kotaro Funakoshi Published: 2026-06-08 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.09424v1 Social robots must interact robustly not only with users assumed by speech-centered systems but also with diverse users whose communication relies on different modalities, e.g., sign language. One important capability gap is predictive turn-taking with signing users. Although Voice Activity Projection (VAP) has been successfully used to model future voice activity in spoken interaction, it remains unclear whether the framework transfers to sign language interaction. This paper presents an initial transfer study of adapting a VAP architecture to dyadic sign language interaction. Using interaction recordings from the Public DGS Corpus, we derive binary signing activity streams from lexical sign annotations and formulate proxy tasks for turn-taking prediction. The model uses pose-derived hand, eye-region, and mouth-region features extracted for each signer. The results show that SHIFT/HOLD prediction is promising, especially with hand cues, while SHIFT-prediction remains difficult. These findings provide initial evidence for both the promise and the current limitations of transferring predictive turn-taking models from spoken interaction to sign language interaction. Predictive modeling of sign language interaction still requires sign-language-specific event definitions that go beyond speech-derived categories. ---