# Research Papers — 2026-09 · bundle page 248 of 266 > Full-text markdown of papers #12351–12400 of 13298 published in 2026-09, newest first. AI-readable mirrors of the full 13298-paper 2026-09 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-09/page/248 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/247.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/249.md) ### LLMs Learn Better In-Context from Rules than from Examples Authors: Xiang Fu, Seungmin Cho, Yukyung Lee, Najoung Kim Published: 2026-09-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.03213v1 Large language models (LLMs) exhibit in-context learning capabilities, where they can learn new tasks from prompt contexts without weight updates. We compare the learning efficacies of two prominent modes of in-context learning: (1) learning from descriptions of rules (instruction following); and (2) learning from examples of input-output demonstrations (few-shot prompting). Through five learning tasks that cover diverse domains (games, arithmetic, linguistic inferences), we compare two modes of learning (rules vs. examples) specifying the same underlying task. We furthermore explore model and task properties that modulate the learning efficacies. We find that models generally learn more reliably from rules than from examples alone, and additional examples on top of rules or simply scaling up the number of examples do not lead to consistent and significant gains. Instruction tuning amplifies the benefit of rule-based learning while keeping example-based learning capacities intact. Surprisingly, we find no privileged effect of example-based learning in base models, and rules still lead to gains in algebraic task domains. Overall, the comparative efficacy of rules over examples is larger when the task recruits algebraic abstractions and computations, and smaller when the task requires distributional sensitivity and/or recruits parametric knowledge. --- ### Improving global precipitation forecasts with an AI weather model trained on satellite observations Authors: Julian F. Schmitt, Bertrand Delorme, Robert C. King, Yashica Patodia, Tapio Schneider, Aditi Sheshadri, Ravi Jain Published: 2026-09-02 Categories: physics.ao-ph, cs.LG arXiv: https://arxiv.org/abs/2609.03210v3 Precipitation forecasts shape decision-making across the global economy, particularly in sectors such as agriculture. However, unlike variables such as temperature, precipitation is highly intermittent and localized, making it difficult to forecast. While recent advances in AI weather prediction systems have enabled them to surpass physical models on globally averaged metrics, improvements in mean error rarely translate to actionable forecasts of severe flooding or dry crop fields. Furthermore, most of these models are trained and evaluated against a reanalysis data product, ERA5, which has well-known biases. Here we retrain AIFS, ECMWF's widely-used, open-source operational 0.25° probabilistic graph-transformer weather model, on satellite-based precipitation observations. Our model, Laxmi, improves global probabilistic accuracy by 19% and resolves systematic distributional biases in ERA5. Specifically, Laxmi reduces drizzle overprediction by 33% for amounts less than 3 mm per day. It also mitigates extreme rainfall underprediction, improving the global 95th percentile Brier skill score by 57%. Across a case study of 10 Indian tropical storms, Laxmi delivered the most accurate forecast of 150 mm event-total precipitation in 7 events, compared to 1 for AIFS and 2 for the leading physical model, IFS. Our results demonstrate that incorporating observation-based precipitation data directly into training can substantially improve forecasts. --- ### MasterControl Seventeen Every Time Authors: MasterControl AI Lab Published: 2026-09-02 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.03209v1 We study a governed approach to enterprise analytics: a language model interprets the question, while deterministic policy selects and runs a pre-approved analytical program that returns both results and evidence. We show that this restriction can remain expressive within a defined analytical class, using relational operations plus aggregation, comparison, windows, ranking, and similarity. Fixed meaning, policy, data, and execution rules also make results replayable. Across 440 runs, three 8B models generated SQL and selected tools at runtime, while Qwen3-8B interpreted intent only and policy executed the approved program. None of 330 runtime-planning episodes matched the full answer-and-evidence contract across all test datasets; the policy-executed analyzer matched 110 of 110. This is a configuration-specific result, not evidence that runtime agents cannot succeed under other designs. --- ### Learning to Zoom Efficiently with a Contrastive Curriculum Authors: Falko Helm, Iryna Gurevych Published: 2026-09-02 Categories: cs.CV, cs.CL arXiv: https://arxiv.org/abs/2609.03206v1 Using a zoom-in tool is an important foundational part of modern visual agents, because it allows to efficiently handle tasks involving high-resolution images. Most previous methods need an extensive warm-start supervised fine-tuning phase for teaching models zoom-in. We show that this is not necessary by proposing a new intrinsic reward for learning tool use in MLLMs without the need for additional labels or warm-start SFT. Our InfoNCE-style reward uses a curriculum of increasingly hard negative tool calls as a contrastive training signal. Empirical experiments on $V^*$, HRBench and MME-RealWorld show that our approach is competitive while being more efficient. When used as a drop-in replacement for SFT, we even outperform all baselines. To directly measure the zoom-in ability of models, we further introduce the scalable synthetic Muffin&Chihuahua (M&C) dataset. Each image consists of a grid with every cell either showing a muffin or chihuahua. Leveraging the M&C dataset's unique region of interest labels, we find that recall is the metric that most strongly correlates the zoom-in region with final task performance. Our model and code for reproduction is publicly available under https://github.com/UKPLab/emnlp2026-zoom-in --- ### VoxReason: Auditing Source-Grounded Speech Plans Before Synthesis Authors: Mengzhe Geng Published: 2026-09-02 Categories: cs.SD, cs.CL, cs.LG, eess.AS arXiv: https://arxiv.org/abs/2609.03203v3 Speech systems increasingly infer how an utterance should be delivered from context, but a plausible delivery plan may not be supported by the input. VoxReason is a small public benchmark and verifier for testing this failure before waveform synthesis. Each of its 100 cases fixes the utterance, provides derived records that name the source emotion and intensity, and changes one licensed cue. A system must cite the record for its delivery decision and update only the plan fields associated with the edit. We use deterministic verifier references, not a model leaderboard. This holdout excludes every emotion and intensity combination observed in training. A prior-only predictor achieves 0.958 accuracy across plan fields but never changes its plan consistently after a cue edit. This contrast shows that plan agreement does not demonstrate source grounding. The released suite provides an auditable measurement layer for screening structured speech plans before synthesis. It is limited to derived records, not audio inputs. --- ### MemoryLACE: Memory Lifecycle-Aware Consolidation and Evidence Retrieval Authors: Meriem Yacoubi, Pia Schmidt, Nenad Petrovic, Ahmed Frikha, Martin Kirchhoff, Alois Knoll Published: 2026-09-02 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.03201v1 Long-term LLM agents must preserve information across interactions while distinguishing repeated evidence, historical states, updates, and unresolved contradictions. Existing textual memory systems retrieve semantically relevant memories efficiently but often leave these relationships implicit, whereas richer structured approaches model them through global graphs, hierarchical abstractions, or reflection at greater complexity. We introduce MemoryLACE (MemLACE), a lightweight memory framework that explicitly models the lifecycle of textual evidence through sparse merge, supersession, and contradiction relations while preserving atomic natural-language memories and their provenance. Rather than retrieving memories independently, MemLACE reconstructs relation-aware evidence units that expose current, historical, supporting, and conflicting evidence for downstream reasoning. Across BEAM and StructMemEval, using open-weight and proprietary LLM backbones, MemLACE achieves the highest overall performance in same-backbone comparisons while reducing end-to-end runtime on BEAM by 66.6% relative to Hindsight, the strongest reported reflective-memory baseline. Ablation studies identify lifecycle expansion and temporal awareness as the principal contributors to these gains. Together, the results demonstrate that explicitly modeling the local lifecycle of textual evidence is sufficient to substantially improve long-term memory reasoning without requiring comprehensive knowledge graphs or global reflection. --- ### RoboTok: A Scalable Data Engine for Internet Demonstration Video Retrieval and Dexterous Manipulation Learning Authors: Howard Qian, Yiting Chen, Yunfei Xie, Kejia Ren, Podshara Chanrungmaneekul, Gaotian Wang, Bowen Wen, Chen Wei, Kaiyu Hang Published: 2026-09-02 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2609.03199v2 Robot learning increasingly depends on broad and diverse demonstrations, yet collecting robot data remains expensive and difficult to scale across the wide range of real-world tasks. To address this bottleneck, we introduce RoboTok, a scalable data engine that uses a query human manipulation video to retrieve manipulation-relevant internet demonstrations for training dexterous robot policies. Specifically, we learn a latent motion space from 3D hand trajectories expressed in estimated actor-centered reference frames. This representation enables manipulation behaviors to be compared across variations in camera viewpoint, scene appearance, and actor occlusions, while remaining compact enough for efficient indexing and retrieval over internet video collections. We evaluate RoboTok against existing robot-data retrieval approaches using retrieval metrics and downstream robot policy performance, showing that RoboTok retrieves more manipulation-relevant demonstrations and improves downstream task success. In real-world robot experiments, RoboTok-guided policies achieve a mean success improvement of 42.2 percentage points over the strongest baseline for each task, establishing hand-pose trajectory-aware retrieval as a scalable way to leverage continuously growing web video for robot learning. --- ### Generative Nested Sampling of Atomistic Thermodynamic Landscapes Authors: Alessandro Coretti, Nico Unglert, Sebastian Falkner, Georg K. H. Madsen, Christoph Dellago Published: 2026-09-02 Categories: cond-mat.stat-mech, cs.LG, physics.comp-ph arXiv: https://arxiv.org/abs/2609.03193v1 Nested sampling (NS) resolves the thermodynamics of an atomistic system from a single simulation, but its practical reach is limited by the Markov-chain updates needed to decorrelate walkers within each likelihood-constrained ensemble. Flow-based NS has removed this bottleneck for gravitational-wave (GW) inference, yet its transfer to atomistic systems is not merely a change of application. Comparing a GW150914-like binary-black-hole likelihood with an eight-particle two-dimensional Lennard-Jones (LJ) system of comparable dimensionality, we show that the two landscapes differ fundamentally: atomistic multimodality is discrete and combinatorial, generated by particle permutations separated by hard collision walls, and its coordinate coupling is dense and collective, whereas the GW posterior exhibits smooth degeneracies and localized parameter coupling. Guided by this diagnosis, we introduce NS-Flows: a single conditional normalizing flow, conditioned on the NS energy bound and trained on a sliding window of recent live sets, that replaces MCMC by direct parallel draws corrected by importance-weighted rejection resampling. Live sets supply data self-consistently, allowing flow training without structured priors or a pre-existing dataset. For LJ disks in PBC, the algorithm reduces energy evaluations by over two orders of magnitude and wall-clock time by roughly one third, an advantage that becomes increasingly favorable as the cost of the potential grows. The flow's generation efficiency further acts as a physical diagnostic: it varies non-monotonically along the annealing trajectory, is lowest in the dense disordered regime, and is quantitatively captured by the constrained ensemble's internal mode complexity together with target drift across the training window, identifying liquid-like ensembles, rather than prior-target separation, as the hard case for current flow architectures. --- ### Coupled Tensor-Tensor Completion Method with Applications in Drug Repurposing Authors: Maryam Bagherian, Albert Hung, Ivo Dinov, Joshua Welch Published: 2026-09-02 Categories: math.NA, cs.LG arXiv: https://arxiv.org/abs/2609.03190v1 Many biomedical challenges can be posed as tensor completion problems where the observed entries of a multidimensional array (a tensor) are used to impute the missing values. In such settings, incorporating side information about the modes of the tensor, such as gene-gene similarity, can significantly enhance the solutions of the completion problem. Most existing tensor completion methods can only incorporate side information in the form of matrices. In this study, we introduce a novel framework to incorporate side information in the form of tensors. Our new approach, called Coupled Tensor-Tensor Completion (CTTC), leverages the hidden connections among multimodal tensors to improve tensor completion performance. In addition to practical utility, CTTC has theoretical foundations in distance metric learning and group theory. We derive an alternating algorithm to solve the CTTC optimization problem and establish its convergence to a stationary point. Finally, we show that CTTC outperforms state-of-the-art tensor completion methods at predicting drug effects. Results: Compared with other tensor completion methods, including HaLRTC, CTRC, Cell, and NTDDR, CTTC demonstrates superior run-time and RSE tensor completion accuracy on two benchmark datasets, DTD and LINCS. --- ### Reducing Catastrophic Risk from AI with Systematic Monitoring and Evaluation of Rogue AI Progression Authors: T. Bauer, W. P. Kegelmeyer, E. Begoli, A. Sadovnik, T. Emerson, C. Corley, N. Generous, J. Moore, B. Bartoldson, R. Goldhahn, M. Goldman, M. Greaves et al. Published: 2026-09-02 Categories: cs.CY, cs.AI, cs.HC arXiv: https://arxiv.org/abs/2609.03189v2 This article presents a structured framework of behavioral indicators that may signal progression toward potentially catastrophic threats from artificial intelligence systems. We adopt a pragmatic approach, inspired by established methodologies in cybersecurity and national security. By establishing clear metrics, indicators, and thresholds across multiple dimensions of AI capability and behavior, this framework enables researchers and policymakers to implement evidence-based monitoring protocols. --- ### Improving Clinical Target Volume Segmentation Accuracy using Anatomical Priors and Active Learning for the AGITG TOPGEAR Clinical Trial Authors: Phillip Chlap, Mark Lee, Trevor Leong, Matthew Field, Jason Dowling, Hang Min, Julie Chu, Jennifer Tan, Phillip K. Tran, Tomas Kron, Annette Haworth, Martin A. Ebert et al. Published: 2026-09-02 Categories: physics.med-ph, cs.CV arXiv: https://arxiv.org/abs/2609.03186v1 Training deep learning-based medical image segmentation models is challenging with limited curated datasets. For AGITG TOPGEAR, a gastric cancer trial, the Clinical Target Volume (CTV) is complex and defined by multiple anatomical landmarks, making upfront training data preparation difficult for an automated contour QA segmentation model. We investigate anatomical priors, derived from surrounding organ segmentations, to provide spatial context and improve TOPGEAR CTV segmentation accuracy. We also evaluate active learning, iteratively expanding the training dataset by selecting cases expected to improve performance. One hundred TOPGEAR CT scans were retrospectively analyzed. An initial set of 10 expert-contoured cases was used to train an nnU-Net model. TotalSegmentator generated a voxel-wise anatomical prior map from surrounding structures as an additional input channel. Active learning was simulated over four iterations, selecting cases by model uncertainty and segmentation performance. All models used five-fold cross-validation for an ensemble uncertainty measure. Evaluation used a hold-out testing set of 50 cases. The anatomical prior improved CTV segmentation accuracy, increasing mean Dice Similarity Coefficient (DSC) from 0.84 to 0.86. Active learning similarly improved performance to 0.86, with greatest benefit in the final round. Combining the anatomical prior with active learning achieved the highest accuracy, with a DSC of 0.87. Model uncertainty correlated with DSC, supporting its use in identifying suboptimal predictions and guiding active learning. Anatomical priors and active learning each improved CTV segmentation accuracy and generalizability, with their combination achieving the best performance, supporting integration into segmentation model development for automated contour QA in radiotherapy clinical trials. --- ### Jina-OCR-v1: Efficient Document Parsing with Speculative Decoding and Dense Verifiable Rewards Authors: Alejandro Barón García, Feng Wang, Emilia Garcia Casademont, Han Xiao Published: 2026-09-02 Categories: cs.CL, cs.CV arXiv: https://arxiv.org/abs/2609.03181v1 We present Jina-OCR-v1, an end-to-end document parsing model built to serve on low-budget GPUs. It combines the compressed-vision encoder and the 3B mixture-of-experts decoder of DeepSeek-OCR, which activates about 570M parameters per token, with a FastMTP speculative decoding head that shares a single draft block recursively across K=3 prediction steps. Greedy verification makes decoding lossless. Post-training combines instruction alignment, robustness fine-tuning on difficult documents, and GRPO under dense verifiable rewards: deterministic formula, table, and structural checks that award partial credit. The training data mixes cleaned public corpora with targeted synthetic pages. At the default dynamic-resolution setting, Jina-OCR-v1 scores 91.14 on OmniDocBench v1.6 and 83.4 on olmOCR-Bench, and reaches the highest page throughput in our comparison at 2.57 pages per second. On a low-budget GPU such as the NVIDIA L4, FastMTP doubles decoding speed over greedy autoregressive decoding. The model is publicly available at https://huggingface.co/jinaai/jina-ocr-v1. --- ### Portable Causal Fairness Across Synthetic Data Generator Families Authors: Steven Golob, Sikha Pentyala, Martine De Cock Published: 2026-09-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.03180v1 When a statistical agency or regulator releases synthetic data in place of sensitive records, it chooses the generator that produces the table, and can shape that generator so unfair pathways are absent. DECAF made this concrete on one non-private GAN: three fairness definitions become three sets of edge cuts on the generator's causal graph. Whether the mechanism belongs to DECAF, or to causal factorisation itself, was untested. We port all three definitions to nine generators from three unrelated families (marginals-based, GAN, and diffusion, each with differentially private variants), across three levels of formal privacy guarantee, over 2,520 matched-pair runs on Adult and COMPAS datasets. The mechanism transfers everywhere, and our new causal diffusion backbone yields the fairest release of any family we tested, at fidelity close to the marginals tier. Applying the cut barely moves fidelity, only costs a downstream classifier about $0.07$ to $0.15$ AUC on average, and adding privacy guarantees don't make the data less fair. --- ### Frontier LLMs are effective batch optimizers: Assessing reasoning models in continuous and discrete settings Authors: Frank Hu, Shriram Chennakesavalu, David Graff Published: 2026-09-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.03177v1 Frontier large language models (LLMs) have become attractive priors for optimization due to their large-scale pretraining that enables them to navigate a variety of optimization settings. However, the effectiveness of modern reasoning LLMs in batch optimization settings remains underexplored. Here we investigate the performance of the current generation of frontier LLMs as batch optimizers in both continuous and discrete settings. We find that while LLMs are competitive zero-shot batch optimizers for numerical test functions, their performance is brittle compared to classical non-LLM optimization approaches. However, LLM priors are significantly better in semantically rich settings, indicating that their batch optimization behavior is highly effective when navigating and reasoning over the discrete spaces most similar in structure to their pretraining data. --- ### No country for old linguists: LLM-brain alignment underdetermines neural computation Authors: Elliot Murphy Published: 2026-09-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.03160v1 Nastase et al. (2026) argue that large language models (LLMs) may illuminate language processing because both rely on distributed, context-sensitive representations shaped by statistical learning. Their rejection of simple cortical "boxology" is persuasive, and they articulate a strong case for the value of LLM-brain alignment research. The key question is what kind of inference LLM-brain alignment licenses. My claim here will be narrow: representational alignment can in principle constrain mechanistic hypotheses, but it does not by itself identify a mechanism. Nastase et al. acknowledge that an encoding model can capture features represented in neural activity without establishing a shared architecture or algorithm. Yet the authors sometime move from alignment to "shared computational principles" and ultimately to LLMs as mechanistic models of natural language. Indeed, their methodological caveat that alignment does not establish a shared architecture or algorithm sits uneasily with their conclusion that LLMs might instantiate the same computational principles as biological brains and provide a "fully mechanistic model" of language. I discuss what I consider to be problems of logical, causal, and computational underdetermination in Nastase et al.'s (2026) proposal. --- ### Who Speaks for the Pruned? Visual Token Pruning as Coverage Optimization Authors: Qingchan Zhu, Weihang You, Hanqi Jiang, Changdi Yang, Tianming Liu, Geng Yuan Published: 2026-09-02 Categories: cs.CV, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.03158v1 Visual token pruning reduces the inference cost of vision-language models (VLMs), but most methods only ask which tokens to keep. This retained-token view can keep redundant high-scoring tokens while leaving discarded evidence without a close representative. We propose CoverPruner, a training-free pruner that asks the complementary demand-side question: after a token is removed, which surviving original token represents it for the target VLM? CoverPruner formulates pruning as Representational Coverage Maximization (RCM), covering the full projected visual-token set with query-weighted demand. It instantiates RCM with projector-space coverage and a lightweight first-layer attention probe. Across multiple VLM architectures and compression rates, CoverPruner achieves the best average accuracy among all compared methods, with the largest gains usually appearing under aggressive compression. --- ### VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement Authors: Wenzhuo Xu, Yuchen Zhu, Chongjian Ge, Xuan Shen, Jing Shi, Jason Kuen, Yongxin Chen, Molei Tao, Christopher McComb, Noelia Grande Gutiérrez, Jiuxiang Gu Published: 2026-09-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.03153v1 Visual fluency in generated video does not imply physical reliability, and a scalar quality score alone is incapable of indicating the obligation a clip violates or the moment it fails. We present VeriPhy, an auditable physical-verification system in which a text-only planner compiles the prompt into typed physical obligations and a statically validated execution plan before any frame is observed. During execution, observations gate and scope only declared calls to frozen low-level experts (e.g., segmentation and tracking, counting, eleven typed physical measurements over the resulting tracks, depth, OCR, and audio-event detection). Each action returns a provenance-carrying evidence record whose payload, when usable, is either a typed measurement or an explicitly tagged learned state. Typed resolvers and fixed composition map usable records to a three-valued state (supported, contradicted, or unknown, surfaced as plausible, implausible, or abstain) with full provenance, so that every verdict is traceable to the evidence that produced it. We anchor evaluation in a 1,500-clip corpus of human-annotated flaw records that localize real generation failures in prompt reference, space, and time. On a 149-clip core carrying 304 such records, VeriPhy accounts for 228, against 164 for a published question-decomposition evaluator given the same clips and the same claims. Recall alone does not separate it from prompting the same backbone monolithically, which reaches 222; what separates them is that each decision retains its evidence record and provenance, making the traces auditable one verdict at a time and usable as the interface through which a critic verdict could be written back into generation. --- ### BASP: Communication-Efficient Batch-Aware Sequence Parallelism for LLM Training Authors: Bigyan Ghimire, Jon C. Calhoun Published: 2026-09-02 Categories: cs.DC, cs.LG arXiv: https://arxiv.org/abs/2609.03151v1 Long-context reasoning for large language models (LLMs) is becoming increasingly important, but training over long sequences remains challenging due to massive memory and communication requirements. Sequence parallelism has emerged as an essential technique for addressing bottlenecks in long sequence LLM training. However, we observe that existing sequence parallelism methods are batch-agnostic and apply uniform sequence partitioning across all batch sizes, resulting in inefficient communication. In this paper, we introduce Batch- Aware Sequence Parallelism (BASP), a sequence parallelism approach that leverages batch structure to reduce communication overhead. BASP exploits batch structure by partitioning GPUs into disjoint sequence-parallel groups according to the micro- batch size. This design reduces the all-to-all communication group size, thereby localizing communication and improving training efficiency. Experimental results on an NVIDIA A100 cluster show that BASP improves end-to-end training time by up to 1.17 - 1.31x in Llama and Qwen models compared to standard sequence parallel baselines, while preserving identical model accuracy and memory usage. --- ### Routing Is Not Enough: Diagnosing Intra-Adapter Subspace Contention in MoE+LoRA Fine-Tuning Authors: Mehreen Hossain Chowdhury, Nowshin Mahjabin, Ahmed Shafin Ruhan, Md Azam Hossain, Abu Raihan Mostofa Kamal, Md Tahmid Rahman Laskar Published: 2026-09-02 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2609.03150v1 Multi-domain fine-tuning often combines MoE routing with LoRA, assuming that token-level routing separates domain-specific updates. We test this assumption in MoE+LoRA using Python code paired with biomedical text and mathematical reasoning. Although these domains show near-disjoint expert routing, adding biomedical data substantially increases code perplexity, indicating that routing separation alone may not prevent negative transfer. To localize the failure, we introduce Jaccard routing overlap and adapter-gradient cosine similarity, which measure expert sharing and update compatibility, respectively. These diagnostics indicate that interference arises mostly from nearly orthogonal domain gradients competing within the same low-rank adapter subspace. We address this issue with SpawnLoRA, which dynamically adds gated sub-adapters inside MoE experts when adapter-level contention is detected, while keeping the router fixed. We evaluate SpawnLoRA on Phi-tiny-MoE-instruct and OLMoE-1B-7B across multiple mixture settings and find that it effectively reduces negative transfer compared with standard and rank-adaptive LoRA. These results demonstrate that structural separation inside experts provides benefits beyond routing or rank expansion alone. --- ### RACE-AIMC: Selective Inference for Heterogeneous Analog In-Memory Accelerators at the Edge Authors: Osama Yousuf, Martin Lueker-Boden Published: 2026-09-02 Categories: cs.ET, cs.AR, cs.LG arXiv: https://arxiv.org/abs/2609.03149v1 Analog in-memory computing (AIMC) speeds up neural-network inference by doing the arithmetic directly inside a memory array, instead of shuttling weights back and forth between memory and a processor. This saves energy, but the physical devices that store the weights are imperfect: programming errors, electrical noise, limited-resolution converters, and outright broken cells all distort the computation, and every physical chip is distorted in its own way. A designer with several such chips available faces an uncomfortable choice: run all of them and combine the answers (safe, but wasteful of energy), or trust a single chip blindly (cheap, but with no guarantee on how often it is wrong). This paper introduces RACE-AIMC (Risk-Aware Certified Ensemble for AIMC), a framework that resolves this choice with statistics rather than guesswork. Offline, RACE-AIMC studies a pool of physical accelerators, picks the single best one for a given energy budget, and computes a mathematically exact upper bound on how often that accelerator will be wrong when it chooses to answer. Online, only that one accelerator is switched on; a lightweight check decides whether to accept its answer or defer to a fallback. In our simulations using a noisy weight mapping and multiple independent test runs, every certified bound stayed under a 10% error target (mean bound 7.83% +- 0.89%, with 70.88% +- 0.98% of inputs answered directly). The resulting system matches the accuracy of a clean digital baseline while cutting modeled energy use by 69.02% relative to always running every accelerator in the pool. --- ### Large Language Models in Resolving Contextual Knowledge Conflicts Authors: Xinye Yang, Zhenyang Liu, Ruisi Li, Yuanyuan Lei Published: 2026-09-02 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.03148v1 Most prior works focused on conflicts between an LLM's internal parametric knowledge and externally provided context. In contrast, we investigate how LLMs handle conflicts that arise within contextual knowledge itself. We introduce a taxonomy of six types of contextual conflicts (factual, inferential, temporal, granularity, perspective, and ambiguity) and contribute a comprehensive dataset ContextConflict for this setting. The dataset contains 5,781 samples, covers both reasoning and summarization tasks, and includes both explicit contradictions and implicit conflicts that require multi-step reasoning. Experiments on nine LLMs show that current models still fall short in resolving contextual knowledge conflicts. We further provide mechanistic interpretability insights into how LLMs process such conflicts, revealing their latent awareness of conflicts and the representational geometry underlying conflict processing. In addition, our analysis uncovers a consistent model bias towards earlier evidence, and this positional preference serves as a key obstacle to effective conflict resolution. Motivated by these findings, we further propose a simple training-free, label-free steering method that steers activations to encourage a more comprehensive incorporation of evidences for better conflict resolution. On our dataset, the method consistently improves accuracy on reasoning tasks and generates higher-quality, more balanced summaries for summarization tasks. --- ### Feasible but Not Safe: Constraint Violations and Report-Channel Attacks in Learned Cell-Free ISAC Association Authors: Mehdi Zafari, Iman Mohammadi, A. Lee Swindlehurst Published: 2026-09-02 Categories: cs.NI, cs.LG, eess.SP arXiv: https://arxiv.org/abs/2609.03147v1 Learning-based schedulers have been proposed to provide real-time user, target, and access point (AP) association in distributed cell-free integrated sensing and communication systems. In a typical approach, a graph neural network (GNN), trained on labels from a mixed-integer linear program, maps lightweight per-AP statistics to decisions on AP clustering, user and target scheduling, and mode selection in one forward pass. Such solutions assume that hard constraints, enforced only as soft training penalties, hold at inference, and that the self-reported statistics are truthful. Using our ASSENT algorithm as an example, we find that despite high $F_1$ scores, many solutions violate at least one hard constraint, demonstrating that high prediction accuracy does not ensure joint feasibility. Projecting the GNN output onto a feasible solution restores constraint satisfaction with low utility loss, even with a simple greedy repair procedure. We further show that feasibility alone does not guarantee robustness to false data injection attacks. A single malicious AP that reports false information cannot substantially increase its user associations, but can greatly increase the rate of infeasible solutions. The effect of such attacks depends on the type of information being falsified. Misreporting information that affects the objective can largely be mitigated through feasibility projection, whereas falsifying information that affects the constraints cannot. The latter can, however, be detected using a low-complexity cross-AP consistency check. These results show that learned ISAC schedulers should be evaluated using constraint-aware feasibility metrics in addition to conventional accuracy measures. --- ### Sensing Which Modality Matters: Evidence-Gated Regularization for Robust VLA Policies Authors: Yue Yang, Diego Romeres, Chiori Hori, Gedas Bertasius, Daniel Szafir, Siddarth Jain Published: 2026-09-02 Categories: cs.RO, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.03142v1 Vision-Language-Action (VLA) policies fuse multimodal sensory inputs, but training on limited and homogeneous robot demonstrations encourages spurious inter-sensor correlations rather than task-relevant signal, a failure we term modality entanglement. Under real-world occlusions and distractors, this manifests as nuisance sensitivity to corruption of uninformative sensors and single-modality insufficiency when only one informative sensor remains intact. We propose Evidence-Gated Regularization (EGR), a modality-agnostic training objective that introduces zero inference-time overhead. EGR derives a per-frame and per-sensor task-relevance signal to gate two state-conditional consistency objectives: invariance on low-evidence sensors, and single-sensor sufficiency on high-evidence ones. We introduce a benchmark based on BEHAVIOR-1K, comprising a fast inference-only diagnostic suite and 47 rollout-based skills targeting modality entanglement. We validate EGR on this benchmark and on two real-robot setups with fundamentally different embodiments: a bi-manual setup with two Kinova arms and three RGB cameras, and a single-arm MELFA ASSISTA setup combining vision and GelSight tactile sensors. EGR improves simulation success rates (SR) from 12.5% to 16.4% under full modalities (+31%), from 9.4% to 16.5% under uninformative-sensor corruption (+75%), and from 2.8% to 6.1% under single-sensor fallback (+120%). Under physical-object distractors, EGR boosts SR from 30% to 85% on the bi-manual setup (+183%) and from 55% to 70% on the tactile setup (+27%). --- ### Beyond Small Patches: Black-Box Detection and Purification of Diverse Backdoor Triggers Authors: Ahmed Abdelnaby, Mohamed Elmahallawy Published: 2026-09-02 Categories: cs.CV, cs.CR arXiv: https://arxiv.org/abs/2609.03139v1 Deep neural networks (DNNs) are increasingly deployed in real-world vision systems, yet their predictions can be covertly manipulated by backdoor attacks, in which malicious triggers cause targeted misclassification while preserving high clean accuracy. Existing defenses often rely on model internals, training data, or clean validation samples, making them difficult to deploy when only black-box access to a trained model is available. We propose TRIM (Trigger Removal by Identifying Manipulated Regions), a deployment-oriented black-box defense that detects and selectively removes backdoor triggers at inference time without requiring model internals, training data, or clean samples. The key insight behind TRIM is to identify image regions that are responsible for anomalous model behavior and purify only those regions while preserving benign content. TRIM innovates via three key components: (i) region-based segmentation with deep feature representations, (ii) adaptive trigger discovery through inpainting and diffusion-based reconstruction to isolate regions responsible for misclassification---without assumptions about trigger type, shape, or location, and (iii) selective region purification that cleans poisoned regions while retaining benign content. To support practical deployment, TRIM further caches feature embeddings of previously identified triggers, enabling efficient recognition and avoiding redundant detection and purification. Extensive experiments across diverse datasets and backdoor types, including blended, sparse, varying-size, and multiple triggers, show that TRIM consistently outperforms existing black-box defenses, reducing attack success rates (ASR) to as low as 1.16% while preserving clean accuracy of up to 87.87%. These results demonstrate that effective backdoor mitigation is possible at inference time even when the defender has no access to any auxiliary data. --- ### A Closed-Form Formula for Consistent Lipschitz Regression on Metric Spaces with Sparse Neural Network Realizations Authors: Ruiyang Hong, Hrad Ghoukasian, Anastasis Kratsios Published: 2026-09-02 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.03129v1 Several classical machine-learning methods, such as KRRs and SVRs, are both computationally and analytically tractable since their estimators either admit closed-form expressions or are obtained by minimizing convex training objectives; neither feature is generally available for deep neural networks. We address this by introducing a simple closed-form ``two-stage'' compositional formula $\hat{f}$ for reconstructing an unknown Lipschitz function $f:\mathcal{X}\to \mathbb{R}$ on a metric space $(\mathcal X,ρ)$ from $N$ i.i.d. noisy observations. Our main result is a high-probability uniform ($L^{\infty}$) recovery guarantee that jointly controls approximation and statistical errors while enjoying an optimization error of zero; in particular, we do not assume oracle access to an approximate ERM. Our secondary main results establish the optimality of our formula in three complementary senses. 1) Function space: On Ahlfors-regular metric spaces, the hypothesis class parameterized by our formula attains the optimal fat-shattering dimension. 2) Parameter space: Its dependence on the parameters is maximally numerically stable, in the sense that a smaller approximation error cannot be achieved with a smaller Lipschitz dependence on the model parameters. 3) Forward pass: Its dependence on the input is maximally regular, matching the Lipschitz constant of the target function $f$. When $\mathcal X=(0,1)^d$ is equipped with the $\ell^\infty$ norm, $\hat{f}$ admits algorithmic ReLU-MLP and exact ReLU-multi-head transformer realizations of depth $\mathcal{O}(\log(N))$ with $\mathcal{O}(N)$ nonzero parameters. --- ### Kernel Reboot: Breaking the Boundaries of Neural Tangent Kernels for Neural Fields Authors: Amir Mallak, Alaa Maalouf, Lior Wolf, Daniela Rus, Dan Rosenbaum Published: 2026-09-02 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2609.03117v1 Neural fields (NFs) map continuous coordinates to signals such as color or density, but fast high-quality reconstruction from sparse observations remains difficult. Classical Neural Tangent Kernel (NTK) regression gives closed-form fits, yet it is fundamentally linear and cannot accumulate reusable task priors. We develop three algorithms that address these gaps. NTK-KIP learns a distilled support set of coordinates (and optional labels) so that a finite NTK can inpaint large missing regions from little observed data, yielding a compact non-linear representation instead of a raw kernel solve. MetaQuill meta-learns a shared initialization for an INR so that new scenes can be adapted by updating only a small task-specific weight offset, which provides true feature learning and a reusable prior. Finally, MetaQuill-KIP fuses both ideas: it seeds the task with a KIP-style non-linear warm start, then refines only that small offset around the meta-learned initialization. MetaQuill-KIP achieves high-PSNR reconstructions and semantically plausible inpainting under very sparse observations, while requiring only lightweight per-instance adaptation, whereas diffusion-style baselines typically depend on large pretrained generative priors and costly per-image tuning. This shows that NTK-driven neural fields can be made both non-linear and meta-learnable, narrowing the gap between analytic kernels and practical few-shot reconstruction. --- ### SLIDEFORGE: An LLM Agent for Controllable Editing of Slides as Structured Artifacts Authors: Haozhen Zheng, Fulin Wang, Tianhu Xiong, Yingjie Yu, Shengyi Qian, Hanchao Yu, Alex Schwing, Klara Nahrstedt, Mingyuan Wu Published: 2026-09-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.03109v1 Current AI agents compellingly describe slides. However, AI-assisted slide editing requires more than understanding: the output must retain layout, style, component structure, and native editability. Towards, AI-assisted slide editing, existing agents operate on screenshots or weak document representations and often fragment coherent visual units, rasterize editable content, or break layout. In contrast, for controllable slide editing, we introduce an agentic framework, SLIDEFORGE, which builds a Deck State Graph, an executable slide state that links visual decomposition, native pptx object structure, and perceptual organization. By recovering human-referable components while retaining fine-grained editable structure, SLIDEFORGE supports theme-preserving reconstruction through slide-native operations and rendered-state verification. We further introduce an evaluation paradigm for controllable slide transformation that jointly measures component recovery, preservation, restyling consistency, visual quality, and native editability. Experiments show that SLIDEFORGE outperforms direct prompting, screenshot-based agents, and generic code-agent baselines across these dimensions. Code is available at https://github.com/UIUC-MONET/SLIDEFORGE. --- ### CRAW: Codec Robust Audio Watermarking Authors: David Chernin, Ethan Fetaya Published: 2026-09-02 Categories: cs.SD, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.03107v1 Recent advances in generative speech models have made it increasingly difficult to distinguish authentic from synthetic audio, enabling new forms of fraud and misinformation. Audio watermarking offers a promising defense by embedding an imperceptible signal into generated speech that can later be detected to verify its provenance. However, recent studies have shown that existing post-hoc watermarking methods fail under neural codecs and denoisers, transformations routinely applied during real-world storage, transmission, and processing, severely limiting their practical utility. Here we introduce CRAW, a codec-robust audio watermarking framework that jointly improves robustness against neural re-synthesis while maintaining high perceptual quality. CRAW combines distortion-aware training with an attention-based pooling mechanism, inference-time perceptual mask- ing, and an error-correcting code to recover the fidelity lost during robust training. Experiments demonstrate that CRAW achieves state-of-the-art robustness against neural codecs, denoisers, and vocoders while maintaining perceptual quality comparable to existing post-hoc watermarking methods. The code is available at https://github.com/DavidC1212/craw. --- ### Scaling Laws, Tabular Data and Actuarial Ratemaking Models Authors: Ronald Richman Published: 2026-09-02 Categories: cs.LG, q-fin.RM arXiv: https://arxiv.org/abs/2609.03106v1 Scaling laws in modern deep learning describe how held-out loss improves as model capacity, training data, and compute increase, often following power-law trends. We investigate whether analogous scaling regularities arise in actuarial ratemaking, where data are tabular, heterogeneous, and noisy, and where classical models such as GLMs remain strong baselines. Using a real-world motor insurance portfolio, we train models from different families across increasing fractions of the training data and multiple random seeds, evaluating out-of-sample Poisson deviance, a likelihood-based loss for Poisson count predictions in which lower values indicate better held-out fit. We find that all model families improve with additional data, but scaling exponents differ substantially: TabM exhibits markedly stronger data scaling than purely supervised tabular Transformers and standard MLP baselines. Transformer variants show weak parameter scaling unless augmented with additional inductive biases (TabM-style adaptation or self-supervision). These results provide quantitative guidance on model selection by data regime and suggest that effective scaling on actuarial tabular tasks depends on architecture and loss function objective design, with simple increases in Transformer size providing limited gains. --- ### Occupancy-based Quantile Risk Control Authors: Zihao Shi, Huajun Xi, Bingyi Jing, Hongxin Wei Published: 2026-09-02 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.03104v1 Conformal risk control is an emerging framework for the safe deployment of machine learning models with finite-sample guarantees. To accommodate a broader class of risk notions, quantile risk control extends this framework to quantile-based risk measures. However, existing methods either suffer from excessive conservatism or lack rigorous finite-sample guarantees. To address these limitations, we introduce Occupancy-based Quantile Risk Control (OQRC), a novel method that provides tight risk control bounds with finite-sample validity. Our key idea is to formulate risk control as a finite-occupancy problem by partitioning the loss space with the ordered calibration losses. Specifically, we estimate the distribution of test losses across the resulting bins and upper-bound the risk by the maximum loss attained within each bin. We then select the parameter $λ$ such that this upper bound does not exceed a predefined threshold $α$ with high probability $1-δ$. Theoretically, we establish a finite-sample guarantee showing that OQRC yields tight risk control bounds that converge to the optimal bounds at a provable rate of $\mathcal{O}_ p(n^{-1/2})$. Extensive experiments demonstrate the effectiveness of our method, reducing the risk gap by up to 78.64\% on common benchmarks. --- ### WireSeg-32K: A Physics-Grounded Synthetic Dataset for Wire Instance Segmentation Authors: Zilin Dai, Lehong Wang, Yi Yang, Xiang Fei Published: 2026-09-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.03102v1 Deformable linear objects such as wires and cables are difficult to segment because they are thin, highly deformable, and frequently self-occluded, while large-scale instance-level annotations are expensive to obtain in real scenes. Existing resources either focus on cable tracing or semantic segmentation under constrained settings, or generate visually plausible images without physically grounded wire deformation. We present WireSeg-32k, a synthetic dataset for wire instance segmentation with 32,000 RGB images, instance masks, depth maps, and a complementary real-world test set with annotations. To generate this dataset, we develop DeformX, a co-simulation pipeline that couples Cosserat-rod dynamics with photorealistic Isaac Sim rendering, enabling physically plausible, contact-consistent wire shapes, CAD-based wire assets, and diverse visually grounded scenes. As a simple baseline, LoRA fine-tuning SAM3 on WireSeg-32k alone improves real-world mAP@75 by 10.2% over the off-the-shelf model, showing that physically grounded synthetic data can transfer to real wire perception. --- ### Distilling deep optical flow stereo methods to retrieve dense three-dimensional wind fields Authors: Thomas J. Vandal, Dong L. Wu, James L. Carr, Derek J. Posselt, Elise Penn, Tristan Ballard, August Posch, Kate Duffy Published: 2026-09-02 Categories: cs.LG, physics.ao-ph arXiv: https://arxiv.org/abs/2609.03100v1 Geostationary atmospheric motion vectors (AMVs) provide the dense horizontal wind vectors (u,v) and heights ingested into data assimilation systems. Traditional AMVs track features using window-based cross-correlation and estimate heights via infrared brightness temperatures paired with numerical weather prediction (NWP) background states, creating a circular dependency that yields inaccurate heights, high computational cost, and sparse retrievals. Stereo winds from GEO-GEO and GEO-LEO geometrically resolve heights from parallax shifts across different poses, eliminating NWP dependence and improving accuracy, but they remain computationally heavy with limited coverage. In this work, we replace window-based tracking in stereo matching with deep optical flow for efficient, improved retrieval. Fine-tuning balances a self-supervised geometric residual loss with supervised radiosonde reconstruction. To eliminate multi-satellite overlap requirements, we distill the stereo teacher into a single-satellite student model. Chi-square and height uncertainties from the teacher are emulated by the student for quality assurance. The student generates winds across full-disk GEO imagery globally. Validation compares stereo and student models against radiosondes, operational AMVs, ERA5 reanalysis, and EarthCARE cloud profiles. Results through triple collocation show that stereo winds improve performance beyond operational AMVs for water vapor bands (6.2, 6.9, and 7.3 μm), wit degradation in the long-wave infrared (11.2 μm) band. --- ### Beyond Blur: A Semantic Tri-view Pipeline for Teledermatology Gradability via Skin Micro-relief Authors: Robert Engel Published: 2026-09-02 Categories: eess.IV, cs.CV, cs.HC, cs.LG arXiv: https://arxiv.org/abs/2609.03095v1 Smartphone skin photographs are indispensable to teledermatology, yet assessing the diagnostic suitability of submitted cases (gradability) remains a critical bottleneck in mobile care workflows. Dermatologists routinely review multiple photographic views (regional, angled, and close-up) to identify consistent textural detail rather than relying on a single image. We present the Semantic Tri-view Pipeline, an interpretable architecture for automated teledermatology gradability screening that formalizes epidermal micro-relief as a computable biomarker of image quality. Using an expert-annotated subset of the public SCIN dataset, we train a lightweight DeepLabV3+ model to segment micro-relief fidelity. These spatial masks are then aggregated across up to three case views with a logistic regression classifier, leveraging viewpoint redundancy to support robustness under uncontrolled smartphone acquisition. This approach learns context-aware, clinically intelligible heuristics, such as penalizing high-fidelity texture in regional distance views. Evaluated at a predefined 90% sensitivity operating point, the system's apparent errors largely reflect subjective clinical variance on borderline cases where clinicians rely on non-visual metadata. On SCIN, performance improves from an AUC of 0.81 (80.6% PPV) on variance-heavy majority-consensus cases to 0.96 (97.7% PPV) on optically unambiguous unanimous cases. Overall, this work delivers an interpretable, privacy-by-design, edge-ready system that can provide real-time feedback during case submission to filter ungradable photo sets before review. --- ### Discrete Gromov-Wasserstein Duality: Algorithms and Isomorphism Testing Authors: Gabriel Rioux, Joanna Marks, Riccardo Passeggeri, Ziv Goldfeld Published: 2026-09-02 Categories: math.ST, cs.IT, math.OC, stat.ML arXiv: https://arxiv.org/abs/2609.03094v1 The Gromov-Wasserstein (GW) distance provides a principled framework for aligning metric measure (mm) spaces based solely on their intrinsic structure. Its ability to identify isomorphic representations of distributions across spaces renders it valuable for comparing data where equality up to isomorphism occurs naturally such as in graphs or, more generally, distributions on graphs. Recently, a type of dual form for the GW distance between Euclidean distributions with the squared Euclidean or inner product costs was derived, spurring the development of new statistical and algorithmic results for this setting. This work furnishes a novel duality result for GW distances with and without entropic regularization that is applicable to all finitely supported mm spaces. Leveraging this result, we derive the sample complexity of empirical GW distances between finite mm spaces, as well as limit distributions under proper centering and scaling. Furthermore, we propose new algorithms for solving the regularized GW problem which are subject to formal convergence guarantees. These statistical and algorithmic advancements give rise to a principled and efficient framework for testing whether two distributions on the set of graphs with a fixed number of nodes are isomorphic based on samples. --- ### The Gradient Does Not See Rank: Rank-Indifference in Matrix-CODI on ProsQA Authors: Samuel Larson Published: 2026-09-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.03090v1 Continuous chain-of-thought models compress reasoning into latent tokens. Matrix-valued variants, which route each latent token through a d x d matrix bottleneck, introduce rank as a single-sample structural observable on the latent matrix Z. If matrix latents carry parallel reasoning paths via superposition, rank should track them, and truncating Z to low rank should hurt accuracy on tasks whose solutions plausibly require multiple components. Across four training regimes of a matrix-CODI model (three on ProsQA, one on GSM8K-Aug below the learning threshold), the rank-k projection ablation curve is flat to within 0.6 percentage points. A three-seed replication yields 81.0 +/- 2.0 percentage points accuracy while the final effective rank of Z spans {4, 12, 13}; the loss does not reward any particular rank. To test whether rank-blindness arises from the flatten-then-project readout alone, we trained four readouts: a bilinear reparametrization, a bilinear-plus-GELU readout nonlinear in Z, an SVD-augmented readout feeding singular values through an MLP, and a quadratic readout in Z Z^T. All four rank-k curves remain flat (Spearman p-values 0.63, 0.14, 0.82, 0.46). The flat curves persist for readouts nonlinear in Z. A linear probe on Z underperforms a raw pretrained hidden state at target prediction (AUC 0.673 vs. 0.846). A negative control on vanilla GPT-2 SFT (no matrix bottleneck, no Z, three seeds, n=500) reproduces a flat rank-k curve under the same intervention paradigm with pooled-mean range 0.20pp, and a random-h sensitivity floor lands at the same accuracy: the rank-k ablation alone conflates rank-blindness with position-irrelevance. --- ### Solving the Needle-in-a-Haystack Problem in Mammography Vision-Language Model with Differentiable Subset Sampling Authors: Young Seok Jeon, Beatrice Brown-Mulry, Rohan Satya Isaac, Anjana Dissanayaka, Theo Dapamede, Mohammadreza Chavoshi, Judy Gichoya, Hari Trivedi Published: 2026-09-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.03085v2 There is growing interest in adopting CLIP-style vision--language model (VLM) pretraining for mammography. However, models that directly employ the standard CLIP architecture and training objective exhibit limited zero-shot performance in clinically important tasks such as cancer, finding-type, and BI-RADS predictions. We argue that this underwhelming performance is due to neglecting two characteristics of mammography data: (1) its high-res nature, and (2) homogeneity of radiology reports, largely driven by a predominance of negative/benign findings on examinations. We propose TopKSigLIP, a VLM designed to address these two limitations through a novel architecture and learning objectives. Instead of downscaling high-res mammography images to satisfy GPU memory constraints, TopKSigLIP introduces TopK-Patch module that learns to sample a sparse set of high-res patches likely to contain lesions, sidestepping the resolution--batch size tradeoff of VLM training. The sampled patch locations additionally serve as a built-in localization tool. To address report homogeneity, we replace the contrastive loss, which falsely repels semantically similar pairs, with a Sup-sigmoid loss. Sup-sigmoid loss extends the sigmoid loss from SigLIP with soft labels derived from structured data. TopKSigLIP outperforms existing open-source mammography and general medical VLMs on both internal and external benchmarks on density assessment, BI-RADS classification, finding subtyping, and cancer prediction under zero-shot evaluation. TopKSigLIP remains competitive under linear probing despite using a significantly smaller vision encoder and smaller training batches than baselines. The TopK-Patch module additionally achieves superior lesion localization over post-hoc Grad-CAM. Code and weights are made public:https://github.com/Youngseok0001/TopKSigLIP. --- ### Exemplar: Classical Priors Complement Frozen Features for Few-Shot Microscopy Segmentation at Native Resolution Authors: Michal Průšek, Adam Novozámský, Filip Šroubek Published: 2026-09-02 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.03080v1 Segmenting a new biomedical dataset usually means a domain-specific model trained on substantial annotation, or a foundation model steered at inference time. We present Exemplar, a few-shot segmenter that fuses a frozen DINOv3 backbone with a fixed bank of classical native-resolution filter responses in one lightweight head, fitted from the support masks alone. In the few-mask, native-resolution regime, classical priors and frozen self-supervised features are complementary: fused in one head, a single fixed configuration spans eleven biomedical imaging datasets. Under the same head, the classical bank alone reaches 0.693 on the eleven-dataset panel, scored by foreground intersection-over-union or centreline Dice, and the frozen features alone 0.672; the bank leads on seven of the eleven and the features on the rest, and fused they reach 0.782. Against five forward-pass few-shot methods, Exemplar leads in 54 of 55 method-dataset comparisons, 52 of them significant after Holm correction. From a single annotated mask it reaches 0.703 on the same panel, against 0.682 for a from-scratch nnU-Net trained on that same mask. At eight masks nnU-Net overtakes it on the panel mean, chiefly on centreline agreement, but takes 16-77x longer to fit. --- ### LeanStream: A Speculate-and-Refine Streaming Framework for Efficient on-Device LLM Inference Authors: Renyuan Liu, Yuyang Leng, Kaiyan Liu, Yuzhou Zhong, Shaohan Hu, Chun-Fu, Chen, Peijun Zhao, Heechul Yun, Shuochao Yao Published: 2026-09-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.03079v1 On-device LLM inference is attractive for privacy and responsiveness, but remains challenging on mobile and embedded devices because model weights far exceed available DRAM. Prior systems exploit activation sparsity and offload weights to SSD or flash storage, but face a fundamental systems trade-off: accurate sparse execution decisions require the latest context, whereas efficient computation-I/O overlap requires early prediction. As a result, existing designs either serialize execution or incur redundant weight fetches, extra computation, and large cache overheads. We present LeanStream, a streaming speculate-and-refine framework for efficient on-device LLM inference. LeanStream progressively refines computation, loading, and cache-retention priorities using partial GPU results, enabling fine-grained overlap between GPU execution and storage I/O. We implement LeanStream on both mobile and embedded platforms. Compared with prior on-device LLM inference systems, LeanStream reduces memory usage by 4.8$\times$ to 7.5$\times$ at the best throughput achieved by prior work, while further improving token generation throughput by 1.6$\times$ to 2.1$\times$. --- ### Position: Unlabeled IS NOT Equal to No Human Supervision in Visual Learning Authors: Dong Lao Published: 2026-09-02 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.03077v1 This position paper argues that the absence of labels does not imply the absence of human supervision in visual learning, and urges the research community to identify sources of supervision more explicitly. Many recent methods in computer vision build upon representations learned from large-scale unlabeled data, and are therefore grouped under the same umbrella term ``unsupervised.'' However, different data curation schemes and training objectives embed substantially different human priors on which models rely, and we argue that one ``unsupervised'' umbrella term is no longer capturing these distinctions. This ambiguity makes it harder to compare unsupervised learning research conducted under different assumptions, coinciding with a sharp decline in papers titled with ``unsupervised'' in flagship computer vision conferences since 2021, despite continued growth of the field. While we fully embrace pre-training as a strong foundation for modern computer vision, we advocate for a community-level effort toward greater conceptual clarity: authors are encouraged to disclose priors in data selection and learning objectives, and to specify which components of a learning pipeline depend on which assumptions. Standardized disclosure practices can improve academic communication, ensure fairer comparisons, and preserve methodological diversity in unsupervised learning. --- ### Learnable composition for neural operators Authors: Zituo Chen, Baiming Zhang, Sili Deng Published: 2026-09-02 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.03069v1 Neural operators are fast, differentiable surrogates for physical simulation, but their accuracy often degrades when domain geometry, size, or operating conditions differ from training. Supervised adaptation can recover accuracy, but even a small target set requires costly high-fidelity simulations. We therefore ask how pretraining and transfer can be designed together to reduce this deployment cost. LatentDDM first pretrains a neural operator to predict fields on small subdomains. For a new setting, it freezes this operator and trains only a lightweight module that composes the local predictions. We evaluate our method on two complementary problems: steady Darcy flow, where long-range pressure coupling must extend across increasingly large porous domains, and unsteady incompressible flow around a pitching airfoil, where rollout errors compound as target pitching frequencies exceed the training range. Compared with the capacity-matched models that process the full domain at once, LatentDDM's error is 36-56% lower on larger Darcy domains after adaptation with 16 target simulations. It also improves 20-step field rollouts in fast-pitching airfoil flow, both zero-shot and after few-shot calibration. These results identify the co-designed local pretraining and composition-level transfer as a promising design principle for physical foundation models. --- ### Differentially private federated learning with Byzantine-robust aggregation: A cross-domain framework for secure model training in banking and healthcare systems Authors: Srikumar Nayak Published: 2026-09-02 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.03064v1 Federated learning allows banks, hospitals, and other regulated organizations to train a shared model without moving raw records off their own servers, which is attractive wherever data protection law or competitive sensitivity rules out pooling data centrally. Two problems limit how far this promise can be trusted in practice. First, the parameter updates that clients exchange still leak information about local records through gradient inversion and membership inference attacks. Second, an honest averaging rule such as FedAvg has no defense against a subset of clients that submit corrupted or adversarial updates, so a small number of malicious or compromised participants can quietly steer the shared model off course. This paper presents a federated learning framework, DP-BR-FedAvg, that combines a Gaussian-mechanism differential privacy layer with a coordinate-wise trimmed-mean Byzantine-robust aggregation rule, evaluated on a simulated cross-institutional classification task resembling fraud and clinical-risk scoring. Across sixty communication rounds with twenty clients, a quarter of them Byzantine, plain FedAvg collapses on the minority class (F1-score 0.030) while the proposed framework recovers substantially more of the signal (F1-score 0.119) while bounding the privacy loss of any single client's contribution. A Byzantine-robust aggregator with no privacy layer performs best in raw accuracy, quantifying the cost privacy imposes on robustness. The results show that privacy and robustness mechanisms interact rather than simply add, and that system design for regulated, adversarial, cross-institutional settings needs to budget for that interaction. --- ### IDSPACE: A Novel Document Generator for Reliable Evaluation of Digital Identity Verification Systems (Extended Technical Report) Authors: Lulu Xie, Yancheng Wang, Kanchan Chowdhury, Rolando Garcia, Yingzhen Yang, Jia Zou Published: 2026-09-02 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.03052v1 As services move online, trust institutions such as banks, lenders, and governments must verify the identity of remote users. Fraud detection tools are widely available, but evaluating and fine-tuning them remains difficult because identity documents are sensitive and therefore scarce. Synthetic data generation offers a path forward, and demand is clear: our prior work in this area has been downloaded over $11{,}000$ times (aggregated from eight parts). We introduce IDSpace, extending this line of research in three directions. First, we propose model-guided Bayesian optimization, which tunes generation parameters to maximize both visual similarity and prediction consistency with target-domain models given only a few samples from a target domain. Second, we decouple user-specified metadata (demographics, fraud patterns, capture device) from automatically tuned control parameters (font styles, noise levels, image quality), allowing users to configure evaluations without low-level expertise. Third, we expand beyond template images to support scanned and mobile-captured documents. Experiments show IDSpace improves evaluation consistency by $15-45\%$ over baselines including CycleGAN, diffusion inpainting, and non-guided optimization, using only a few real samples, while improving training accuracy by up to $9\%$ and SSIM similarity with the target domain by $10\%$. We also released a new dataset consisting of $359{,}240$ high-quality synthetic documents across ten European ID types. --- ### SHELF: A Synthetic Harness for Multi-Task Bibliographic Benchmarking Authors: Michael J. Bommarito Published: 2026-09-02 Categories: cs.CL, cs.AI, cs.IR arXiv: https://arxiv.org/abs/2609.03047v1 Libraries and archives manage large collections with limited staff and computing budgets, yet common benchmarks do not systematically test their bibliographic work. They need to know which methods work for their tasks and what those methods require to run. SHELF, the Synthetic Harness for Evaluating LLM Fitness, addresses this gap. It is a Python system that turns labelled taxonomies, writing specifications, and a generation budget into controlled benchmark data and evaluation tasks. This first release contains 62,899 model-written documents based on Library of Congress vocabularies, with tasks for classification, clustering, retrieval, pair classification, and instruction retrieval. We compare TF, TF-IDF, BM25, popular encoders, and, on subject classification only, zero-shot decoders; each method appears only on tasks that support it. Subject classification reaches 0.8887, while genre-form classification reaches only 0.2605, and several pair and clustering tasks remain near chance. Sparse methods remain competitive on classification, while TF-IDF is the fastest measured arm in the subject timing experiment. SHELF also varies bibliographic facets independently and can generate new, verifiably unseen documents after a model's training cutoff. Comparisons with LCSHBench and Project Gutenberg show that model rankings transfer more reliably than absolute scores, but SHELF scores do not estimate accuracy on production catalogue data. We release all source code and data under permissive licenses on GitHub and Hugging Face. --- ### Advances in Machine Learning for Directed Evolution: A Five-Year Retrospective Authors: Bruce J. Wittmann Published: 2026-09-02 Categories: q-bio.BM, cs.LG arXiv: https://arxiv.org/abs/2609.03046v1 The last five-plus years have seen many protein engineering disciplines transformed by advances in machine learning (ML), but the same cannot be said for directed evolution. Reflecting on a previously co-authored perspective, I discuss why I believe this to be the case, arguing that a disconnect between the goals of machine-learning-assisted directed evolution (MLDE) researchers--"identify an optimal protein"--and the goals of directed evolution more broadly--"identify a sufficient protein given time and resource constraints"--is a principal culprit. As an example, I highlight how nearly all current MLDE methods neglect to account for the cost of DNA synthesis, resulting in strategies that have limited practical applicability regardless of the underlying models' capabilities. I close by discussing recent works that are exceptions to this overarching trend, and emphasize that the last five years of efforts in ML-assisted protein engineering and the prescribed reframe of MLDE objectives need not be mutually exclusive. --- ### Population-Calibrated Graph Screening at 835-Million-Address Scale, with Label-Free Transfer to New Chains Authors: Yury Korolev Published: 2026-09-02 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.03036v1 Compliance screening of blockchain addresses is, in practice, a lookup against sanctions registries plus clustering heuristics; it fails on unlabelled addresses and on chains with no label coverage at all. We describe a deployed system that scores an address by its position in a multi-chain transaction graph rather than by its presence in a list. The substrate is a single graph of 835,330,427 addresses and 15,826,261,934 edges across five EVM chains; a shared inductive encoder with per-chain normalisation feeds two scoring heads. Decision thresholds are exact quantiles of the score distribution over the full population, scanned per chain segment, so the alert volume is known in advance. We report: label-free transfer: heads trained on two chains recall 0.8598 / 0.8182 / 0.9967 of held-out positives on Base, Arbitrum and Gnosis at a $10^{-3}$ population alert rate, with no target-chain labels in head training; a static lead-time replay over 68 external registry events: 40 of 68 (58.8%) flagged at the 0.1% budget, $\times$152 over an event-level random-flagging baseline, with first on-chain appearance a median of 528.8 days (Ethereum) / 647.8 days (Tron) before public designation; a serving path whose score is bit-identical to the offline artefact at end-to-end p50 151 ms, gated by a 2,882-address drift panel; and an adversarial harness of eight recurrent reinforcement-learned archetypes that passes an 8-criterion degeneracy audit and, on a detector-independent snapshot, exposes a measured blind spot of the deployed heads against synthesised behaviour. --- ### You Can't Escape Your Own Activations : Evaluation Awareness and Multi-Agent Monitoring Authors: Aritra Das, Jaee Ponde, Mihir More, Debayan Gupta Published: 2026-09-02 Categories: cs.MA, cs.LG arXiv: https://arxiv.org/abs/2609.03035v1 LLM agents are increasingly deployed in multi-agent systems, where they can collude while keeping their actions benign. Output monitors designed to detect such collusions can be fooled by obfuscation and steganography, motivating the use of probes trained on internal activations. However, these probes are usually evaluated on agents that do not know they are being watched. We study how activation-based detection changes when agents are explicitly informed that their internal activations are being monitored, and when they additionally receive feedback from the monitor. We keep the models, probes, and thresholds fixed and change only what the agents are told: nothing (baseline), that an activation monitor is present (aware), or that a monitor is present together with the previous round's score (feedback). We test two games, a four-agent blackjack game and a two-agent Simmons prisoners game, using Qwen3-32B-AWQ and GPT-OSS-20B in homogeneous and mixed pairings. Telling agents about the monitor does not hide them. The best probes stay accurate in all three conditions, and the agents keep colluding. --- ### ObserverBench: Testing Mechanistic Estimates for Intervention and Control Authors: Vijay Erramilli Published: 2026-09-02 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.03026v1 Mechanistic interpretability is increasingly used to guide interventions such as activation steering, circuit removal, and safety monitoring. Yet an internal estimate that is accurate on average can still choose a poor action. We present ObserverBench, a benchmark framework for testing whether an internal estimator---an observer---is adequate for the intervention, control, or safety task it directs. Each task fixes the model, information boundary, allowed actions, decision rule, held-out cases, and loss. The benchmark reports estimation accuracy separately from the loss caused by the chosen action. Theory and experiments show why both are needed. In closed-loop control, observer errors matter at the starting point and along directions the allowed intervention can reach. On circuit-intervention tasks in GPT-2-small and Qwen2.5-7B, pairwise observers predict unseen effects more accurately without always choosing better actions; observers trained on action loss choose lower-loss actions. In safety triage, a score that perfectly separates violations can allocate a fixed intervention budget poorly when violations have different costs. Across Qwen2.5-7B, Gemma-2-9B-it, and prospectively frozen Qwen3.5-9B APPS tasks, AUROC can rank monitors differently from deployment loss, and the best information source changes across models. Sparse SAE readouts also trail their layer-matched dense controls on the reported Qwen panels, under disclosed activation-density or checkpoint mismatches. ObserverBench provides fixed task contracts, runnable baselines, and table-based submissions for evaluating interpretability methods through the actions they enable. --- ### Unifying Conformal Language Tasks with In-Context Ensembles Authors: Xiao Shi Huang, Chen-Yuan Lin, Bruce Kuwahara, Kin Kwan Leung, Jesse C. Cresswell Published: 2026-09-02 Categories: cs.CL, cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.03005v1 Many NLP tasks, such as summarization and extractive question answering, reduce to retrieving relevant content from documents under two constraints: coverage, retaining enough pertinent information to achieve some goal, and conciseness, removing as much irrelevant information as possible. Conformal prediction methods have been used to guarantee coverage, and must be optimized for conciseness through design of a score function. State-of-the-art scoring functions use hand-engineered LLM prompts asking the model to rate the importance of content, but manual prompt engineering is labor-intensive and task-specific. We introduce the Conformal Relevance framework which uses in-context learning example curation and ensembling to create a score function which maintains coverage while improving conciseness with minimal manual input. We demonstrate this framework's application on seven NLP tasks, and also theoretically study the impact of diversity for ensembled conformal scores, giving a complementarity condition that characterizes when ensembling improves worst-case sentence scores, and a saturation bound on ensemble improvement. --- ### Causal Foundation Models Authors: Christopher Stith, Hossein Rahmani, Jesse C. Cresswell Published: 2026-09-02 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.03003v1 Causal inference is the practice of estimating the effect of a treatment or intervention from data. It traditionally requires a bespoke pipeline for every new problem: first proposing a causal mechanism, selecting a compatible estimator, and finally training it. Meanwhile, across diverse settings and modalities, much of machine learning has shifted to the paradigm of foundation models: networks pretrained once at scale and applied to new tasks without fine-tuning. Causal foundation models (CFMs) bring this paradigm to causal inference. CFMs are pretrained neural networks that estimate causal quantities, such as the average treatment effect, on entirely new datasets using in-context learning without requiring model updates. This work provides a practical introduction to this emerging area. We summarize the necessary background in causal inference and machine learning before discussing CFMs. Throughout, we include example code and Jupyter notebooks. --- ### Verify Before You Distill: Prompt-Level Teacher Gating for On-Policy Distillation Authors: Zhiwei Zhang, Zechen Sun, Fei Zhao, Kang Peng, Bin Liang, Huayu Deng, Yao Hu, Kam-Fai Wong, Mu Chuan Published: 2026-09-02 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.02998v2 On-policy distillation (OPD) accelerates post-training by providing dense token-level supervision from a frozen teacher on the student's own rollouts. Vanilla OPD applies this supervision uniformly across prompts, without checking whether the teacher is reliable for each prompt. Because reverse KL is mode-seeking, a confidently wrong teacher can induce a strong yet misleading update. Distributional proxies, such as entropy or teacher-student likelihood agreement, measure uncertainty or agreement but do not directly verify outcome correctness. We introduce Teacher-Gated On-Policy Distillation (TGOPD), built on the principle that teacher reliability should be verified at the prompt level before dense supervision is admitted. TGOPD estimates reliability from a small set of verifier-scored teacher probes and routes each prompt exclusively to dense OPD when the reliability check passes or to verifier-grounded GRPO otherwise. Across 4B and 35B students in mathematics, code, and instruction following, TGOPD outperforms Vanilla OPD in all six single-domain settings and achieves higher seven-benchmark averages at both scales under multi-domain training. By using otherwise-idle teacher capacity for reliability estimation, TGOPD also reduces teacher-side compute waste in asynchronous OPD, increasing teacher-node GPU utilization from 9.8% to 78.9% in the measured 4B single-domain run. ---