# Research Papers — 2026-09 · bundle page 163 of 266 > Full-text markdown of papers #8101–8150 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/163 [← Previous bundle](https://aipod.dev/papers/archive/2026-09/page/162.md) [Next bundle →](https://aipod.dev/papers/archive/2026-09/page/164.md) ### A multimodal large language model for evidence-based autism spectrum disorder screening Authors: Jun Chen, Qi Zhao, Yunliang Jiang, Shuqin Cao, Yunqiang Lin, Chenglong Jia, Qiang Guo, Guang Dai, Xiongtao Zhang, Mengmeng Wang, Xiaoyue Ma Published: 2026-09-15 Categories: cs.CV, cs.HC, cs.LG arXiv: https://arxiv.org/abs/2609.16464v1 The clinical management of autism spectrum disorder (ASD) faces a bottleneck in early screening, mainly because trained specialists are scarce and conventional assessment tools are subjective. Here, we introduce ASDchat, a multimodal large language model designed for evidence-based ASD screening, which takes video, audio, and dialogue as input. ASDchat adopts a dual-branch architecture, where the decision branch generates screening probabilities and the evidence branch generates traceable, timestamped behavioral evidence aligned with standardized clinical criteria (ADOS-2). The model was trained and evaluated on a dataset of 1,035 participants from 27 sites in China, which covered typically developing (TD) children, children with ASD, and children with other disorders. For ASD versus TD, ASDchat reached an area under the receiver operating characteristic curve (AUC) of 0.953 $\pm$ 0.021. On 9 held-out sites that were not used for training, the mean AUC was 0.932. Furthermore, unsupervised clustering of the behavioral dimensions split the ASD cases into six subtypes with different phenotypic profiles, and ASDchat suggests an intervention for each subtype. ASDchat provides a feasible path for large-scale, evidence-based early ASD screening in clinical practice. --- ### Not All Relations Are Equal: Relation-Balanced and Calibrated Graph Learning for Provenance-Based Intrusion Detection Authors: Lijie Zheng, Ji He, Zhiwei Zhang, Alessandro Brighente, Yulong Shen, Mauro Conti Published: 2026-09-15 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.16462v3 Provenance-Based Intrusion Detection Systems (PIDSs) detect Advanced Persistent Threats (APTs) by analyzing system interactions. However, existing methods largely treat relations uniformly, overlooking statistical heterogeneity; in CADETS, relation frequencies differ by approximately $140{,}000\times$. This may cause PIDSs to focus more on frequent relations and overlook differences in normal error levels across relations, increasing the risk of false alarms and missed detections. We present RECAL, an unsupervised framework using relation-balanced masked graph learning to better capture rare interaction patterns. It further calibrates reconstruction errors against each relation's benign error distribution to produce comparable anomaly evidence, helping distinguish attacks from benign behavior and reduce false alarms. On three DARPA E3 datasets, RECAL achieves F1 scores of 99.99\%, 99.93\%, and 99.99\%, outperforming the best baseline on each dataset by 0.88, 0.82, and 0.42 percentage points, respectively. Compared with the baseline reporting the lowest FPR, RECAL reduces mean FPR by approximately $105\times$, $4\times$, and $41\times$. --- ### Protocol-Preserving Context Trimming for Agentic Workflows: Benefits, Failure Regimes, and Budget Guardrails Authors: Harish Gaggar Published: 2026-09-15 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2609.16461v1 Agentic large language model (LLM) systems rely on long interaction histories to preserve instructions, tool states, intermediate decisions, and unresolved dependencies, but unrestricted context growth increases computational cost and can reduce efficiency. This study evaluates protocol-preserving context trimming as a reliability-constrained approach for multi-step agentic workflows. Five trimming strategies - recency-based, relevance-based, summarization, protocol-aware trimming, and adaptive budget guardrails - were compared across retained-context levels and workflow-complexity classes using task success, protocol adherence, valid tool calls, token savings, latency reduction, cascading failures, and critical context thresholds. Conventional strategies achieved about 60% mean token savings but lower task success (66.6-77.3%) and protocol adherence (85.5-88.6%). Protocol-aware trimming improved task success to 92.2%, while adaptive guardrails achieved 96.0% task success, 96.3% protocol adherence, and 1.0% cascading failure with 56.0% mean token savings. Retained-context budgets of 25% or less increased failure odds 10.92-fold relative to budgets of 50% or more (p < 0.001). Protocol-aware trimming produced 5.24-fold greater odds of successful completion than conventional methods under aggressive budgets, while adaptive guardrails further increased success odds 2.11-fold versus fixed protocol-aware trimming (p < 0.001). Critical context thresholds also increased with workflow complexity. These findings indicate that reliable context reduction depends more on preserving protocol-critical state than on maximizing token removal, and that adaptive guardrails can improve efficiency, scalability, and reliability in long-horizon agentic systems. --- ### OPD-Aha: From Linguistic Momentum to Visual Reflection in Multimodal On-Policy Distillation Authors: Chenhao Qiu, Dawei Li, Yechao Zhang, Lei Gong, Zhen Tan Published: 2026-09-15 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2609.16459v1 Privileged on-policy distillation improves multimodal reasoning by allowing a teacher to evaluate student trajectories using rich, training-only visual evidence. Both models score these trajectories while conditioning on the same student-generated prefix. When a student misinterprets an image early in a response, this accumulating erroneous rationale eventually pulls the teacher away from its visual evidence. The teacher and student converge on the same hallucination, causing standard cross-model supervision to collapse precisely where correction is most needed. We find that the teacher's visual corrective preference is not lost under this misleading agreement. Comparing the predictions of the identical teacher given the real image and a visual null reveals that the privileged evidence still pushes the model toward the correct interpretation. We introduce OPD-Aha, which reconstructs the distillation target directly from this isolated visual preference rather than relying on the fragile teacher-student discrepancy. This reconstructed target aggressively suppresses continuations that contradict the image. Trained with this objective, students learn to naturally interrupt their own flawed reasoning with reflection tokens such as wait and actually. After reflection, subsequent generation relies less on the accumulated erroneous text and more on the visual evidence. Correcting these trajectories mid-generation fundamentally alters the reasoning process, yielding broad and consistent improvements across diverse fine-grained perception and complex multimodal reasoning benchmarks. Our code and models are available at https://github.com/Echochef/OPD-Aha. --- ### Language Orthogonalization for Zero-Shot Cross-Lingual Audio Deepfake Detection Authors: Minu Kim, Ji Sub Um, Hoirin Kim Published: 2026-09-15 Categories: eess.AS, cs.CL arXiv: https://arxiv.org/abs/2609.16458v1 Audio deepfake detectors need to transfer to languages absent from training, as multilingual speech synthesis outpaces labeled anti-spoofing resources. While detectors increasingly rely on self-supervised speech models (S3Ms), these backbones encode language-dependent structure that confounds spoof cues. We address this confound through language orthogonalization, a target-free ridge map that removes S3M variation projected onto continuous language-identification (LID) embeddings. Across six languages, six S3M backbones, and all Leave-N-Out settings, it consistently reduces EER across unseen languages. Cross-lingual EER correlates with LID-space distance, where orthogonalization yields larger gains for more distant transfers. --- ### Fine-Tuning Fixes Mode Collapse and Over-Dispersion in LLMs Authors: Kirill Skobelev, Eric Fithian, X. Y. Han Published: 2026-09-15 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.16454v1 Recent work by Doshi and Hauser (2024), Bisbee et al. (2024), and Xie et al. (2026) raises concerns that outputs from large language models (LLMs) tend to be under-diverse: they repeat or resemble one another more often than responses from the population they are meant to represent, a phenomenon known as mode collapse. In this work, we show that whether mode-collapse, or its opposite, occurs depends on the specific model and dataset used. Further, with sufficient supervised fine-tuning (SFT) data, LLM output diversity converges toward that of the target distribution from which fine-tuning data are sampled. To quantify this comparison, we measure the probability that two responses sampled independently from the same fixed prompt coincide (collide), or their expected similarity under a kernel. We derive a bias-variance decomposition of the expected gap between the model's and target's collision probabilities, showing that SFT is not inherently biased toward mode collapse or its opposite: finite-sample SFT can leave a model either under- or over-dispersed, depending on the model and dataset. Finally, we show that the absolute gap is bounded by the square root of the Kullback-Leibler (KL) divergence from the target distribution to the model. Consequently, a model sufficiently close to optimal under population cross-entropy cannot exhibit arbitrarily miscalibrated diversity. We test the decomposition and the bound in three experiments: small transformers on synthetic languages, four LLMs fine-tuned on human surveys, and these LLMs fine-tuned on CodeNet, a dataset of human code solutions. More target data moves model diversity toward the human (or synthetic target) level in all experiments, consistent with our theoretical predictions. These results show that diversity miscalibration can arise from finite-sample error and shrink as SFT better approximates the target distribution. --- ### Early-Bird Decoding: Accelerating Diffusion LLMs with Learnable Block Sizes and Parallel Sampling Authors: Lixuan Wei, Wei Zhou, Jianwen Wu, Yipeng Shen, Meiling Wang, Haoran You Published: 2026-09-15 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.16450v1 Diffusion large language models (dLLMs) offer a promising parallel decoding paradigm as an alternative to autoregressive generation through iterative unmasking. However, dLLMs typically require many steps before token confidence reaches the decoding threshold, resulting in inefficient inference even with block-wise KV caching. To accelerate dLLM inference, we for the first time propose an "early-bird (EB)" decoding framework, motivated by the observation that tokens with similarly low entropy tend to cluster and can be jointly decoded earlier, before reaching the confidence threshold. In particular, our EB-Decode framework integrates two key enablers: (1) a learnable network that adaptively groups tokens with similar uncertainty into variable-length blocks, rather than relying on fixed block sizes; (2) a position-aware sampler that learns to unmask tokens in parallel using fewer decoding steps within predicted variable-length blocks. Both components are developed without modifying pretrained dLLM weights and can therefore be directly deployed as plug-ins during serving, with negligible training and inference overhead. Extensive experiments across three models and four benchmarks consistently validate our observation and the effectiveness of EB-Decode, achieving 3.53-18.76$\times$ higher throughput than the vanilla decoding method and up to 1.58$\times$ higher throughput over the strongest baseline, Fast-dLLM, with comparable accuracy. --- ### Decentralized Gossip Learning and Federated Averaging for Histopathology Image Classification Authors: Yusuf Ozturk, Bengisu Atli, Enes Goktekin, Akin Ozturk, Ulas Bagci Published: 2026-09-15 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.16448v2 Breast histopathology analysis increasingly relies on distributed learning because direct data pooling across institutions is often restricted by privacy, governance, and communication constraints. This study compares server-based Federated Averaging (FedAvg), fully decentralized gossip learning, and Hybrid Gossip-FedAvg for invasive ductal carcinoma (IDC) patch classification. Experiments used 277,524 color image patches with patient-disjoint training, validation, and test partitions and a workload-balanced, Dirichlet-guided allocation across six nodes. Ring, random degree-3, and fully connected gossip topologies were evaluated together with sensitivity analyses for statistical heterogeneity, mixing coefficient, learning rate, model drift, prediction disagreement, calibration, clinically motivated operating points, communication payload, and patient-level IDC burden, together with auxiliary backbone robustness analyses. In the principal alpha=0.3 experiment, Hybrid Gossip-FedAvg achieved a test area under the receiver operating characteristic curve (ROC-AUC) of 0.8811, closely followed by FedAvg at 0.8801 and fully connected gossip at 0.8751. Across three independent patient-level repetitions, FedAvg and Hybrid Gossip-FedAvg obtained the same mean ROC-AUC of 0.9082, with standard deviations of 0.0037 and 0.0043, respectively. Hybrid achieved the highest mean area under the precision-recall curve of 0.8240, whereas FedAvg produced the lowest mean Brier score of 0.1335. Denser gossip graphs improved discrimination but increased theoretical model payload, while ring gossip remained sensitive to learning rate and mixing strength. Overall, FedAvg provided the most consistently reliable server-based baseline, topology-aware gossip offered a viable decentralized alternative, and Hybrid Gossip-FedAvg provided a balanced compromise between peer-to-peer diffusion and periodic global coordination. --- ### Adaptive Bayesian Partner Selection for Federated Clinical Centers Authors: Navid Seidi, Satyaki Roy, Sajal K. Das Published: 2026-09-15 Categories: cs.LG, cs.DC arXiv: https://arxiv.org/abs/2609.16446v1 Federated learning (FL) in healthcare faces pronounced heterogeneity and temporal concept drift across clinical centers, where evolving patient populations and care practices shift data distributions. Existing approaches rely on persistent global communication, incurring substantial bandwidth overhead while risking negative transfer from poorly aligned peers. We propose Adaptive Bayesian Partner Selection (ABPS), a peer-to-peer framework that governs who collaborates, when, and at what cost. Each center maintains a Beta-Bernoulli posterior over prospective peers' Shapley marginal utility, ranks candidates with an Upper Confidence Bound (UCB) criterion, and forms collaborations through a lightweight propose-reject mechanism, with the option to abstain from communication when no mutually beneficial partner exists. The framework admits a stochastic decision interpretation, yielding finite-sample concentration guarantees and O(kappa log T) regret in partner selection, along with conditions under which intentional isolation is optimal under negative transfer. Lightweight extensions (head personalization, bfloat16 quantized communication, and a tunable active-set size) further improve efficiency, and a goal-aware metadata filter enables institution-specific collaboration strategies. On binary in-hospital mortality prediction over the first 24 hours of an ICU stay, with 230 non-IID clinical centers drawn from MIMIC-IV, the full ABPS-X variant matches the strongest federated baseline (FedDyn, AUROC 0.758) at 0.09x the communication cost of FedAvg, with reduced variability. A diversity-driven configuration activates intentional isolation for a substantial fraction of centers. These results show that adaptive, utility-aware collaboration reduces communication without sacrificing accuracy when centers are numerous and small, offering a scalable paradigm for healthcare FL. --- ### Beyond the Graph: An Adaptive Meta-Learner Fuses Explainability, Weather, and Dynamics for Robust Bus ETA Prediction Authors: Pratham Payra, Jagadish Published: 2026-09-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.31667v1 Accurate bus Estimated Time of Arrival (ETA) prediction is vital for urban mobility, passenger satisfaction, and transit efficiency, yet existing models falter against nonlinear spatiotemporal dynamics, data sparsity, and factors such as weather. This paper proposes HYB(nm), an adaptive hybrid ensemble framework that dynamically fuses five complementary models - a historical baseline (MST-AV), periodical temporal pattern analysis (GDRN-DFT), Koopman Neural Operators for nonlinear dynamics (KOOP-NET), weather-integrated feature-engineered neural networks (FENN), and real-time graph convolutional networks (MGCN) - via a meta-learner attuned to real-time context. Evaluated on GPS and weather data from three Kolkata bus routes comprising more than 4,000 trips, the framework leverages the individual strengths of its components (for example, the low-latency explainability of MST-AV, the weather resilience of FENN, and the network-dynamics capture of MGCN) to deliver the superior robustness of HYB(2), state-of-the-art accuracy rivalling leading graph neural networks, and balanced trade-offs in stability and efficiency across prediction horizons and operating conditions. The extensible HYB(k) architecture equips transit agencies with flexible tools, ranging from economical single models to tailored high-fidelity hybrids, advancing predictive, equitable urban transport. --- ### Temporal-Attention Head Specialization During Video Diffusion Training Authors: Taewoo Ha, Shafayat Mowla Anik, Dae Yeol Lee, Byeong Kil Lee, Jeeho Ryoo Published: 2026-09-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.31654v2 Video diffusion transformers depend on temporal attention to coordinate information across frames, yet nearly everything known about this mechanism comes from analyzing trained models, so when and where temporal-attention structure forms during training remains poorly characterized. Population averages can also hide it, since a few specializing heads and a diffusing majority cancel in the mean. We therefore conduct a checkpoint-resolved census of every temporal-attention head across nine Open-Sora STDiT training runs spanning three model scales (306M to 1.03B parameters), scoring each head with an entropy-normalized measure of cross-frame attention concentration (CFAC) under a preregistered change-point and effect-size selection rule. The census reveals the sparse picture that averages obscure. Aggregate CFAC is flat or decreasing in every run, while a small minority of heads, roughly 4--13% in full-grid runs, develops pronounced concentration. Across seeds, the reproducible signal is positional but block-level. Selected heads repeatedly arise in the first temporal block, whereas individual head coordinates do not reproduce once block membership is accounted for. Among the analyzed 760M selected heads, attention maps converge to a small repertoire of local frame-routing motifs, self-frame diagonals and adjacent-frame bands, even when the responsible coordinates differ across runs. Correlation and ablation analyses do not establish a causal link to generated video quality, and we bound our claims accordingly. Beyond this STDiT family, the study contributes a transferable methodology. Checkpoint-resolved, per-head analysis under fixed selection rules can expose sparse temporal organization in other factorized video diffusion transformers and, with adapted routing metrics, in joint spatio-temporal architectures. --- ### Distributional sentiment modeling and anomaly detection for consumer complaint assessment Authors: Peiheng Gao, Chen Yang, Shimin Zhang Published: 2026-09-14 Categories: cs.CL, cs.LG, stat.AP arXiv: https://arxiv.org/abs/2609.31653v1 Sentiment analysis is a common tool for converting unstructured text into quantitative signals in finance and risk management. Yet most applications reduce the output to a discrete polarity label or a single predictive feature, overlooking the distributional structure of sentiment intensity in consumer complaint narratives. In this paper we treat negative sentiment in consumer complaints as a bounded continuous variable and study its full distribution rather than a single label. We score each narrative with a transformer classifier, model the scores with Beta distributions, and compare the fitted distributions of meritorious and non-meritorious complaints through the Kullback Leibler divergence and the squared Hellinger distance. The fitted distributions are then linked with dollar amounts and company response outcomes to construct anomaly diagnostics that flag complaints whose textual severity is inconsistent with the recorded relief. We find that the two groups have strongly overlapping distributions, so negative sentiment intensity is not a sharp classifier of outcomes on its own; combined with monetary and categorical attributes, it isolates unusually severe complaints for operational risk monitoring. Treating sentiment analysis as continuous distributional measurement, this study links sentiment extraction, bounded response modeling, and anomaly detection in a unified framework for consumer complaint assessment. --- ### Open-Qwen-Music: An Auditable Framework for LLM-Based Music Composition and Diffusion Rendering Authors: Yangbin Yu, Mingyu Yang Published: 2026-09-14 Categories: cs.SD, cs.AI, eess.AS arXiv: https://arxiv.org/abs/2609.31652v1 We present Open-Qwen-Music, an open reconstruction of Qwen-Music and a fully specified research system for text-to-music generation that couples LLM-based semantic composition with diffusion-based acoustic rendering. The system comprises a 25 Hz single-codebook music tokenizer, a 3B-parameter autoregressive Music LLM, and a diffusion renderer producing 48 kHz stereo audio, following the cross-module interfaces reported by Qwen-Music. The strongest systems of this design remain closed, and prominent open music-generation projects release weights and inference code without their training corpora or end-to-end training implementations. This limits independent and controlled study of how information loss and prediction errors propagate from semantic representation through autoregressive planning to acoustic rendering. To our knowledge, Open-Qwen-Music is the first fully open release of an LLM-composition-plus-diffusion-rendering text-to-music system. Beyond model weights and inference code, the release includes the training datasets and provenance manifests, complete data-processing, annotation, training, inference, and evaluation pipelines, configurations, and pretrained weights for every learned module. Artifact manifests bind the identities of these artifacts across the complete workflow. Together, these artifacts establish a reproducible implementation of the modular architecture and provide an empirical basis for component-level analysis and future evaluation. We present the system as a transparent, executable research baseline and a starting point for the community, not as evidence of quality parity with Qwen-Music. Open-Qwen-Music is an ongoing effort, and we will continue to improve its generation quality, controllability, and robustness. All release artifacts are available at https://github.com/biang15343100-source/Open-Qwen-Music. --- ### PalmLeaf-VQA: A Multi-Script Visual Question Answering Benchmark for Historical Palm-Leaf Manuscript Understanding Across Diverse Regions Authors: Nimol Thuon, Jun Du, Panhapin Theang Published: 2026-09-14 Categories: cs.CV, cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2609.31651v1 Historical manuscripts remain largely absent from modern vision-language benchmarks, leaving open how well multimodal large language models (MLLMs) handle culturally diverse, degraded, and non-Latin document images. We introduce \textbf{PalmLeaf-VQA}, a multi-script visual question answering benchmark for historical palm-leaf manuscript understanding across South and Southeast Asian traditions. PalmLeaf-VQA contains \textbf{923 curated manuscript images} and \textbf{7,384 question--answer pairs} from eight collection groups: Balinese, Grantha, Jathakam, Kambaramayanam, Kannada, Khmer, Sundanese, and Tamil. Unlike recognition-oriented resources, the benchmark targets manuscript-aware visual reasoning over preservation-relevant cues, including physical condition, line structure, material and coating, binding holes, margins, symbols, drawings, and localized visual artifacts. We evaluate recent proprietary and open-weight MLLMs under open-answer and constrained-answer prompting and provide fine-grained analysis across collections, question categories, and task types. The strongest evaluated model reaches only \textbf{58.00\% exact-match accuracy} on the held-out test split, revealing substantial limitations in current MLLMs for rare-script, degraded-layout, and preservation-oriented document understanding. PalmLeaf-VQA provides a standardized benchmark for advancing culturally grounded and layout-aware multimodal document analysis. --- ### A literature-guided descriptor-based framework for filtering composition search spaces Authors: Lei Zhang, Markus Stricker Published: 2026-09-14 Categories: cs.CL, cond-mat.mtrl-sci arXiv: https://arxiv.org/abs/2609.31650v1 Scientific literature contains latent knowledge about materials behavior, but much of this knowledge is expressed through words, contexts, and recurring associations rather than explicit design principles. This raises a central question: how can large-scale scientific corpora be used for practical problems in materials discovery? Here, we present a literature-guided descriptor-based filtering framework for reducing composition search spaces. For a given performance metric, the framework selects two descriptors from a filtered vocabulary in a literature-trained word embedding model and uses the selected descriptors to construct a Pareto-based filter for candidate compositions. Across the evaluated performance metrics and composition search spaces, the framework filters out an average of 74.27\% of the candidate compositions, with an average best-value error of 1.93\% relative to experimental measurements. Compared with expert-chosen and random descriptors, our performance metric-dependent descriptors provide a more controlled balance between retained fraction and best-value error. These results show that literature-derived embeddings can support intuitive and reproducible filters for narrowing candidate composition spaces while preserving high-performing compositions. --- ### NeuralCert: certified computational discovery of extremal mathematical constructions Authors: Mark Patrick Roeling Published: 2026-09-14 Categories: cs.LG, stat.CO, stat.ME arXiv: https://arxiv.org/abs/2609.30296v1 Neural networks are becoming popular in solving mathematical problems, but stochastic models do not provide mathematical exactness by themselves. This study introduces a discovery-to-certification framework in which high-dimensional variational trial functions are learned in a compact separable representation, spectrally diagnosed and pruned, and then certified exactly through multimodular evaluation. Exact certification makes the numerical proofs fully explicit and independently verifiable. This framework can be run on a standard personal computer. Across three extremal problems, we show that neural optimization can contribute to rigorous mathematics in three distinct ways: by discovering improved constructions, by exposing empirical invariants that lead to proofs, and by revealing optimization barriers whose geometry motivates new analytic or numerical representations. More broadly, these results suggest a path toward AI-assisted mathematics in which flexible computational discovery and exact certification become complementary components of a single rigorous workflow. --- ### Long-Tail Adaptive Flow Matching with Explicit Conditional Consistency Guidance for Precise Multimodal Face Synthesis Authors: Yushe Cao, Xuechao Zou, Xing Xi, Dianxi Shi, Chun Yu, Junliang Xing Published: 2026-09-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.29581v1 Although diffusion-based methods have substantially improved the controllability of multimodal face synthesis, their semantic alignment remains suboptimal because most existing approaches rely on implicit latent-space objectives to model the relationship between denoising variables and multimodal conditions. Such implicit modeling is often insufficient to enforce precise correspondence between synthesized faces and conditional inputs, especially under long-tailed semantic mask distributions where rare attributes receive weak optimization signals. To address these limitations, we propose EC\textsuperscript{2}Face, a multimodal face synthesis framework that improves semantic alignment through explicit semantic supervision and distribution-aware optimization. First, we introduce Explicit Conditional Consistency Guidance (ECCG), which imposes direct consistency supervision in pixel space by decoding an approximate reverse estimate of the clean latent and explicitly aligning the synthesized image with textual descriptions and semantic masks. A temporal dynamic modulation function is further designed to adapt the supervision strength according to the timestep-dependent reliability of reverse estimation. Second, we propose Long-Tail Adaptive Flow Matching (LAFM), which reweights spatial optimization signals based on semantic attribute frequency, with normalized weights to maintain numerical stability during training. Importantly, all additional modules are used only during training and introduce no extra inference overhead. Extensive experiments show that EC\textsuperscript{2}Face consistently outperforms competitive baselines in both generation quality and semantic alignment, achieving a 29.38\% improvement in mask accuracy on rare attributes. --- ### Gödel's and Scott's Variants of the Ontological Argument in Lean 4 and TPTP THF Authors: Christoph Benzmüller Published: 2026-09-14 Categories: cs.LO, cs.AI arXiv: https://arxiv.org/abs/2609.26806v2 This paper presents a complete, structure-preserving port to Lean 4 of the Isabelle/HOL dataset accompanying Benzmüller and Scott's study of Gödel's ontological argument and Scott's variant: 30 modules, one per theory, retaining section structure, declaration order and names up to documented renamings; a comparison tool certifies the 548 statements identical as parsed. Every named result the original proves is proved again, from the inconsistency of Gödel's 1970 axioms to modal collapse, monotheism and the ultrafilter property of positive properties. Five statements the original reports proved but does not replay are proved here. The 45 statements it refutes with Nitpick (35) or leaves open (10) are anonymous sorrys nothing depends on. Lean 4 has neither a sledgehammer nor a model finder, so automated proofs become explicit proof terms and the 65 Nitpick invocations are documentation. #print axioms then lists, as Isabelle/HOL's thm_deps would, the postulates each proof consumes, hence an upper bound on the modal logic it needs: the proofs of Scott's necessary-existence theorem and of modal collapse consume only symmetry of the accessibility relation, so KB suffices; those of the essence and monotheism lemmas, of the possible existence of a God-like being (with one recorded exception) and of the 1970 inconsistency consume none. The port also renders the dataset in TPTP THF, the format in which Gödel's argument was first mechanised, and in SMT-LIB: a metaprogram prints the 294 theorems as problems. Six provers (E, Vampire, Zipperposition, cvc5, Leo-II, Leo-III) prove 227 of them within ten seconds on one core, 231 within sixty, and none of the 45 left unproved; Leo-II, repaired here and released as 2.2, is level with E at ten seconds. The development needs no library beyond Lean 4's core; sources, tools, cross-checks and both renderings are ancillary files. --- ### Authority-Preserving Evaluation of Medical Vision-Language Assistants Authors: Flint Xiaofeng Fan, Cheston Tan, Yew-Soon Ong, Roger Wattenhofer Published: 2026-09-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.22302v1 Medical vision-language models can propose how urgently a skin lesion should be reviewed, but the local service retains authority to accept or replace that proposal under referral policy, capacity, and locally held patient context. Proposal quality and selected-action quality are therefore distinct evaluation targets, and benchmark evidence transfers between them only when local review preserves the expected action score. We introduce AuthEval, a logging and evaluation framework that records both actions, scores the selected action under declared local criteria, and, where feasible, scores the declined proposal under the same rule. It reports the resulting authority gap only when the record supports it. Because the gap is the product of the proposal-change rate and the mean score change on changed cases, that rate alone determines neither its magnitude nor its sign. On ISIC 2019, with MedGemma and simulated local review, two constraint regimes with similar change rates produced an optimistic image-equal gap under capacity ($+0.744$ simulator units) but no detectable gap under safety. The declared evaluation unit also mattered: under mixed constraints the gap reversed from $+0.374$ to $-0.206$ when weighting shifted from image to lesion-aware cluster. AuthEval thus clarifies whether a study's records support claims about the model, the workflow, or both. --- ### When Does Test-Time Physical Diagnosis Pay? A Frozen Policy Buys Evidence It Never Reads Authors: Zhengshu Zhang Published: 2026-09-14 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2609.22299v1 When a robot faces unfamiliar physical conditions, a common approach is to collect evidence about what changed and adapt. For such diagnosis to improve behavior, six ordered empirical conditions must hold: a meaningful reference, identifiability of the physical condition, use of the acquired evidence, decision value, selection value over a fixed alternative, and safe realization. We test this chain in controlled and public environments. It holds end to end in our controlled environments. After transfer to unseen mechanisms, however, it breaks at evidence use. On decisions requiring the full trace, the frozen decoder does not change its choice. A linear model using only trace increments recovers the correct choice on mechanisms excluded from fitting, showing that the trace is informative but unused. The failure is concentrated at the richest evidence level: those decisions fall to chance, while decisions settled with lower-cost evidence remain correct, a split hidden by aggregate accuracy. The same chain can fail at other links in public environments. Successful physical identification therefore guarantees neither evidence use nor useful adaptation; evaluation should identify where the chain breaks rather than rely on recovery accuracy or aggregate performance alone. --- ### On The Robustness-Resolution Tradeoff In Temporal Quantization Of Event Streams Authors: Sayeed Shafayet Chowdhury, Ruhi Sharmin, Syed Ishtiaque Ahmed Published: 2026-09-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.22295v2 Event pipelines often discretize asynchronous timestamps before learning. This step looks harmless, but its stability depends directly on temporal resolution. We study this dependence at the representation level. We first show that hard temporal binning is discontinuous: an arbitrarily small timestamp shift near a boundary can move unit event mass between bins. We then define a class of nonnegative, mass-preserving, resolution-faithful continuous encoders and prove that every encoder in this class has global L1 sensitivity at least 2/Delta, where Delta denotes bin width. Linear two-bin interpolation attains this limit. Local support and first-moment preservation also make it unique. Experiments on SHD, N-MNIST, and DVS128 Gesture support the analysis. Across uniform timestamp budgets, linear interpolation lowers mean representation drift by 47-72% while keeping clean accuracy nearly unchanged. On DVS Gesture, it produces zero prediction flips across all tested budgets and three seeds. On SHD, measured drift follows 1/Delta with R^2 = 0.992. --- ### Machine Learning for Underwater Optical Wireless Communication Systems: A Comprehensive Survey Authors: Shaymaa Mahmoud, Ardimas Purwita, Mohamed-Slim Alouini Published: 2026-09-14 Categories: eess.SP, cs.LG arXiv: https://arxiv.org/abs/2609.22294v1 Underwater Optical Wireless Communication (UOWC) has emerged as a promising technology for high-speed underwater data transmission, offering significantly higher bandwidth and lower latency compared to acoustic and Radio Frequency (RF) technologies. However, the underwater medium introduces severe impairments that degrade link performance and limit communication range. The growing complexity of these challenges has increased research interest in Machine Learning (ML) and Deep Learning (DL) approaches, which offer powerful tools for channel modeling, signal processing, and system adaptation in ways that conventional analytical methods struggle to achieve. This survey provides a comprehensive review of ML methods applied across the full UOWC system pipeline, covering channel modeling, transmitter design, receiver detection and equalization, link alignment, and emerging applications, including wireless power transfer, semantic communication, object detection, optical sensing, and localization. Finally, we discuss open challenges and future research directions to motivate further work in this rapidly growing field. --- ### Validating, Not Sampling: Region-Level Robustness of Vision-Language and Vision-Language-Action Models Authors: Bogdan Aron, Christopher Brix, Benedikt Brückner, Yanghao Zhang, Panagiotis Kouvaros, Alessio Lomuscio Published: 2026-09-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2609.22293v1 Vision-language models (VLMs) and vision-language-action models (VLAs) are increasingly deployed in real-world applications. There, a small perturbation to the recorded camera image may change a decision significantly. However, existing benchmarks for these models only sample perturbations, which does not guarantee the absence of a failure in the untested region. We present the first robustness validation of six VLMs (drawn from the Gemma, InternVL, LLaVA, and Qwen families) and five VLAs (drawn from the GR00T, OpenVLA, and $π$ families) over entire continuous regions of photometric and geometric image perturbation: brightness shifts, camera rotations, and their composition. To this end, we build on the validation framework H$^2$V and introduce H$^2$V-M, a margin-aware convergence rule that makes validation affordable at the 32B parameter scale. We demonstrate that H$^2$V-M outperforms H$^2$V by an order of magnitude in model queries and that it finds counterexamples faster than random sampling while providing soundness guarantees. Our VLM and VLA robustness validation shows that robustness is mostly dependent on the perturbation type, rather than the model, and that VLMs are more robust to large camera rotations than VLAs. For VLAs, even perturbations as small as $\pm1^\circ$ can change the commanded action in many cases. We also show that robustness depends more on model family than on model size. --- ### Beyond the Survey: A Systematic Empirical Study of Detection and Association in Visual MOT Authors: Linh Van Ma, Juhua Hu, Wei Cheng, Unse Fatima, Moongu Jeon Published: 2026-09-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.22291v1 This paper presents a comprehensive experimental evaluation and detailed analysis of state-of-the-art multi-object tracking algorithms, with an emphasis on quantifying the individual contributions of detection and association components to overall tracking performance. Unlike existing surveys that primarily offer theoretical categorizations or taxonomies of tracking methods, our work adopts a rigorous experimental perspective grounded in publicly available implementations, providing practical guidance for researchers and practitioners in method selection and system design. We introduce a unified pipeline diagram that consolidates the core components across the two main branches of visual multi-object tracking: tracking-by-detection and end-to-end deep learning paradigms, and systematically analyze the object detection, feature extraction, and data association modules. Through extensive empirical studies on standard benchmarks, including MOT16, MOT17, MOT20, SportsMOT, DanceTrack, and CrowdTrack datasets, we reveal critical insights: (1) detection quality dominates association strategy performance, with detector improvements yielding more than 10% gains compared to less than 5% from refined association strategies; (2) modern deep learning detectors paired with specialized re-identification models significantly outperform joint detection and embedding approaches; and (3) transformer-based end-to-end methods exhibit greater robustness to detection quality variations but at a substantial computational cost. Our findings from extensive experiments provide key insights into component-level effects in MOT, particularly the dominant influence of detection quality relative to association, while offering practical insights for designing and optimizing MOT systems under varying performance and robustness requirements. Code and experimental setups are available at github.com/linh-gist/VisualMOT. --- ### AgentPProf: Semantic Profiler for Long Horizon AI Agents Authors: Yusheng Zheng, Chaokun Chang, Yu Mao, Tianyuan Wu, Yuxi Huang, Tao Ma, Wenan Mao, Shuyi Cheng, Andi Quinn, Wei Wang Published: 2026-09-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.20301v1 AI agents increasingly orchestrate long-running activities with users, tools, and system resources for days and weeks. To improve agent quality, safety, and cost efficiency, developers need to determine where failures happen, what triggers unsafe effects, and which tasks consume the most budget, then optimize those tasks. In systems software, profiling answers similar questions by aggregating resource consumption and attributing it to responsible code paths to identify hotspots. Yet existing agent observability tools focus on per-execution debugging and tracing rather than cross-run, long term profiling, making these questions difficult to answer at scale. Agent observability needs profiling, not only debugging, but profiling agents is challenging: the responsible entities are task intent like diagnose authentication, compare branches rather than code paths, and lack stable identifiers for aggregation. We propose a semantic operation stack model that adapts profiling to agent trajectories. Uniform operations represent all activities, and operation stacks replace the runtime call stack, enabling hierarchical attribution at different granularities. We observe that an agent's task occupies a contiguous span and decomposes into subtasks, so we introduce recursive operation segmentation, which recursively splits trajectories at task boundaries. AgentPProf is a profiler that aggregates agent trajectories into pprof-compatible profiles, enabling flame graph visualization and analysis. AgentPProf reaches 0.764 $B^3$ F1 against human annotations on CodeTraceBench. On three problem-localization benchmarks, the profile raises MAP by up to 56%, demonstrating that it effectively attributes resources, locates problems, and helps optimize token cost at practical profiling cost. AgentPProf is available at https://github.com/eunomia-bpf/agentsight. --- ### Regularized Emphatic Temporal-Difference Learning: Stability under Constant Stepsizes Authors: Xingguo Chen, Zhaohui Wu, Jinguo Ye, Chao Li, Shangdong Yang, Guang Yang, Skylar Liang, Wenhao Wang Published: 2026-09-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.19170v1 Emphatic temporal-difference learning (ETD) stabilizes the expected off-policy TD update and changes its projection geometry, but neither property determines constant-stepsize sampled dynamics. We construct an ergodic two-state counterexample in which the ETD mean map contracts while the sampled product has a positive top Lyapunov exponent. Regenerative-cycle analysis separates this sign from the infinite variance of the follow-on trace. We introduce regularized emphatic TD (RETD), a normalized first-order post-shock repair that leaves the trace and importance ratios unchanged, stores the emphatic TD signal in a leaky scalar state, and releases a delayed correction. RETD's raw equilibrium is an affine shift of the ETD equilibrium; single- and two-regularization readouts recover the ETD fixed point exactly. We prove almost-sure convergence for harmonic diminishing stepsizes and a conditional constant-stepsize moment-contraction result from a Markovian random-product bound. RETD has certified negative exponents on the two-state construction and one Baird point, whereas the positive Baird ETD sign remains numerical. Paired 10,000-run experiments validate both separations, fixed-point recovery, a nonmonotone stability region, and task dependence. RETD changes post-shock dynamics; it does not reduce the shared follow-on-trace variance. --- ### Democratizing Clinical Tumor Whole Genome Sequencing: 18-hour End-to-end Analysis via Trillion-parameter Large Language Models Locally Deployed on Consumer-grade Hardware Authors: Rui Xiao, Yili Xu Published: 2026-09-14 Categories: q-bio.GN, cs.LG arXiv: https://arxiv.org/abs/2609.17620v1 Whole genome sequencing (WGS) is essential for precision oncology, yet its clinical adoption remains limited by prohibitive computational costs and multi-day turnaround times. This work presents a fully localized low-resource framework enabling stable deployment of a trillion-parameter biomedical LLM on a single consumer-grade RTX 4060 laptop with 32GB system memory and 8GB VRAM, as well as on routine clinical workstations in general hospitals, completing the entire tumor-paired WGS workflow from raw FASTQ input to clinical-grade full-variation-spectrum report output. Under standard 30X depth configurations, our implementation finishes a single tumor-paired WGS analysis within 18 hours, achieving 99.62% F1 score for somatic variant detection with over 99.9% concordance to the industrial-standard A100 cluster pipeline, fully meeting clinical oncology accuracy requirements. Quantitative profiling shows adaptive heterogeneous memory scheduling accounts for 71% of total execution time, while model optimization introduces less than 9% of total detection error. This work is the first engineering implementation of trillion-parameter biomedical LLM-driven clinical-grade genomic analysis on consumer-grade hardware, breaking the industry paradigm that trillion-scale genomic LLMs require hundred-thousand-dollar GPU clusters and multi-day turnaround, establishing a low-resource pathway for global primary medical institutions to adopt whole-genome precision oncology at zero additional cost. --- ### Stability-Constrained Approximation in Spline KANs: Exact Layer Balancing and Budget-Compatible Saturation Authors: Aleksander Tankman Published: 2026-09-14 Categories: stat.ML, cs.LG, math.NA arXiv: https://arxiv.org/abs/2609.17619v1 Deep spline superposition networks face a tension between approximation order and stability across depth. We study approximation under a hard layerwise Lipschitz budget, and organise it around two quantities: the factorisation stability complexity of a given deep factorisation, and the budget-compatible approximation complexity of a discretisation operator. First, we solve exactly the finite-depth diagonal balancing problem for a fixed chain of nonnegative envelope matrices: the optimal uniform layer budget equals $\|M_{L-1}\cdots M_0\|_{\infty\to\infty}^{1/L}$, attained by an explicit one-pass minimiser, for rectangular layers, with a complete treatment of degeneracies and non-attainment. The optimum can be arbitrarily larger than the Lipschitz constant of the network itself, because passing to envelopes destroys sign cancellation. Second, we give a constructive spline discretisation theorem preserving the budget up to a controlled slack, with an explicit grid threshold. Conversely, for linear spline-valued operators that preserve the budget exactly, we prove budget-compatible minimax lower bounds on classes constrained simultaneously in the first and third derivative norms -- a constraint pair that is forced by the problem and that rules out the usual scaling escapes. Finally, we show that the corresponding layer errors need not cancel under composition: for every operator of the class there is a stable depth-$L$ tower realising a constant fraction of the accumulated error, so the linear-in-depth accumulation of the upper bound is not a proof artefact. --- ### DualCount: Structurally Consistent Density and Point Modeling for Zero-Shot Object Counting Authors: Xuan Cuong Ngo Published: 2026-09-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.17613v1 Zero-shot object counting aims to estimate the number of objects specified by a text query without category-specific training. Recent approaches primarily rely on density regression or detection-style instance prediction. While effective, density-based models often suffer from spatial ambiguity and background leakage due to weakly regulated mass allocation, leading to fragmented or part-biased representations that increase counting error in complex scenes. In this work, we propose an instance-aware dual-decoder framework that structurally couples density and point representations for zero-shot object counting. Instead of treating density estimation as independent pixel-wise regression, we interpret it as a structured mass allocation problem over a latent set of object instances. Predicted instance centers induce a soft instance-wise decomposition of the density map, upon which we enforce two geometric constraints: (1) per-instance mass conservation, ensuring each object contributes approximately one unit of density mass, and (2) center-of-mass alignment, encouraging each density component to concentrate around its corresponding predicted center. These constraints introduce instance-level geometric consistency and lead to more accurate mass allocation, thereby reducing counting error. Extensive experiments on FSC-147, PUCPR+, and CARPK show that our approach consistently reduces counting error and establishes new state-of-the-art performance in zero-shot object counting. --- ### Timbre Analysis of the Hulusi, a Southwestern Chinese Free-Reed Instrument, using Machine Learning Authors: Yang Xia, Rolf Bader Published: 2026-09-14 Categories: cs.SD, cs.LG, nlin.AO, physics.data-an arXiv: https://arxiv.org/abs/2609.17612v1 The hulusi is a wind instrument that was invented in Yunnan Province, China, and has become tremendously popular in recent years. It consists of a mouthpiece, a gourd, and three bamboo tubes, all with free reeds made of copper. The main bamboo tube in the middle has seven finger holes. In this instrument, the pipe length, not the free reed's eigenfrequency, determines the instrument's pitch, unlike, for example, with the Western accordion or the blues harp. In this study, a machine learning model implemented in the COMSAR framework (https://github.com/ifsm) was used to investigate the timbre characteristics of the \emph{hulusi} to cluster different instruments and pitches. The measured \emph{hulusi} pitches C, B, A, G, and F were analyzed according to seven psychoacoustic features, among which only the spectral centroid, sharpness, and fractal correlation dimension are shown to form pitch clusters. These timbre features were used to train Kohonen self-organizing maps (SOMs) for clustering. Brightness and sharpness analysis revealed that the highest pitches were less bright and less sharp than mid- and low-range pitches were. Furthermore, the fractal correlation dimension, which mainly determines the chaoticity of the initial transients, was the best-clustering timbre feature for the hulusi, with the highest pitches showing the least chaoticity. This result is supported by defining a cluster quality index for the SOMs. --- ### The Cost of Becoming: Developmental Encoding Increases Phenotypic Diversity but Reduces Locality and Recombination Robustness in Evolved Robots Authors: Lyes Saad Saoud Published: 2026-09-14 Categories: cs.NE arXiv: https://arxiv.org/abs/2609.17606v1 Developmental encodings are often motivated by the expectation that a structured genotype-to-phenotype process can improve evolvability, robustness, and adaptation. Yet an encoding that expands phenotypic variation may simultaneously make useful parental structure harder to preserve under mutation and recombination. We test this trade-off in a controlled evolutionary-robotics benchmark comparing three matched representations: direct encoding, a static generative encoding, and a temporal zygotic developmental encoding. All three conditions use 34 genome parameters, 34 adult-controller parameters, identical initial genome matrices, the same optimizer, the same mutation and crossover operators, the same environments, and the same number of fitness evaluations. Across 30 paired evolutionary runs per representation, developmental encoding does not reliably exceed direct encoding in final training or out-of-distribution fitness. It does, however, substantially increase final population phenotypic diversity and mutant-reachable phenotype diversity. The same representation exhibits lower mutation viability, lower genotype--phenotype locality, sharply reduced crossover offspring fitness and viability, and increased recombination novelty. Continued developmental dynamics also attenuate damage applied early in development relative to otherwise identical late damage. These results identify a representation-level trade-off: development can enlarge the neighborhood of reachable phenotypes and provide within-development damage attenuation while degrading the local inheritance of already-adapted structure. We argue that developmental encodings should therefore be evaluated not by performance alone, but by a joint geometry of generativity, locality, robustness, and inheritance compatibility. --- ### The Neverwhere Visual Parkour Benchmark Suite Authors: Ziyu Chen, Henghui Bao, Haoran Chang, Alan Yu, Ran Choi, Kai McClennen, Gio Huh, Kevin Yang, Ri-Zhao Qiu, Yajvan Ravan, John J. Leonard, Xiaolong Wang et al. Published: 2026-09-14 Categories: cs.RO, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2609.16443v1 State-of-the-art visual locomotion controllers are increasingly capable at handling complex visual environments, making evaluating their real-world performance before deployment increasingly difficult. This work intends to narrow this train/evaluation gap by developing a collection of hyper-photo-realistic, closed-loop evaluation environments - The Neverwhere Benchmark Suite - comprised of over sixty 3D Gaussian Splatting reconstructions of urban indoor and outdoor scenes. Our goal is to encourage large-scale and reproducible robot evaluation by making it easier to create and integrate Gaussian splats-based reconstructions into simulated continuous testing setups. We also underscore the potential pitfalls of relying exclusively on 3D Gaussian-generated data for training, by providing policy checkpoints trained over multiple Neverwhere scenes and their performance when evaluated in novel scenes. Our analysis illustrates the necessity of sourcing diverse data to ensure performance. Code and data are available on the project page: https://ziyc.github.io/neverwhere-bench/. --- ### Learned Look-Ahead Splitting Rule for CART Authors: Andrew Gao, Tianlin Liu, Ruichen Han, Lu Tian Published: 2026-09-14 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2609.16440v1 Classification and regression trees are typically constructed using a greedy splitting rule that maximizes the immediate reduction in prediction error at each node. Although this strategy is computationally efficient, it can miss splits that yield small short-term gains but create substantial downstream improvements after further partitioning. We propose a look-ahead tree-building method that evaluates each candidate split by the prediction error reduction achieved after growing a conventional CART subtree below that split. Because the full look-ahead procedure can be computationally expensive, we also describe a smart look-ahead algorithm that learns downstream split values using node-level features. The proposed framework preserves the interpretability of recursive partitioning while improving split selection in hierarchical or interaction-driven settings. We conduct a simulation study comparing conventional, full look-ahead, and smart look-ahead methods under several settings and apply the proposed methods to analyze two real data examples demonstrating the merit of the new methods. --- ### Interpreting and Steering LLM Agents for Social Simulations Authors: Jiayue Gaveal Fan, Arul Murugan, Shreyas Krishnan, Abhishek Nagaraj Published: 2026-09-14 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2609.16436v1 Simulations based on large language models (LLMs) have proven to be powerful for understanding human behavior, making them valuable additions to the social scientific toolkit. However, LLMs are ultimately black boxes based on deep neural networks which limits their value for social science. This is because of a lack of (i) interpretability: i.e. the ability to assign clear mechanisms driving observed behavior; and a lack of (ii) steerability: i.e. the ability to mute or amplify specific theoretically meaningful mechanisms of action to drive specific model behavior. Here, we demonstrate how the black box could be opened up to further enrich LLM-based simulations. Specifically, we compare three types of methods: (1) prompt-based manipulation, (2) SAE-derived feature steering, and (3) probe-based direction steering and examine their utility for LLM-based social scientific simulations. We do so by interpreting and steering two foundational components of human behaviors, namely preferences (risk attitudes, altruism) and capabilities (divergent creativity, product innovation), operationalized using four classic economic and creative tasks implemented as natural-language interactions. Overall, our results show that SAE- and probe-based techniques often outperform basic prompt-based methods for steering LLM agents, although this advantage depends on the specific prompting strategy involved. Together, SAEs and probes constitute an effective pipeline for social scientists seeking to interpret and steer agents in social simulations: SAEs decompose agents' internal representations into human-readable features, after which probes can reliably shift agents' behaviors in specified directions. We discuss implications of these methods for future work using LLM agents for social scientific simulations. --- ### Scaled Hippocampus-inspired Neural Networks on Neuromorphic Memristive Hardware Authors: Joseph A. Kilgore, Jeffrey D. Kopsick, Zahin Ahmed, Giorgio A. Ascoli, Gina C. Adam Published: 2026-09-14 Categories: cs.NE, cs.ET arXiv: https://arxiv.org/abs/2609.16429v1 The hippocampus, a key brain region for learning and memory, exhibits rich structural diversity, sparse communication, and robust dynamics with incredible energy efficiency. It offers promising insights for novel computing capabilities, particularly when co-designed with emerging hardware technologies. In this work, we draw inspiration from the rodent CA3 hippocampal subregion to develop the first spiking neural network with neuronal diversity and biologically-realistic resting state dynamics demonstrated on memristor hardware. We propose a network downscaling methodology utilizing a 4-prong objective function and demonstrate a small-scale CA3-inspired network with 179 Izhikevich-modeled neurons, 3 neuronal types and 17,996 synapses with similar resting-state dynamics as the orders-of-magnitude larger full-scale network. The small-scale network is mapped to an FPGA/memristor platform using a greedy algorithm and 18,316 memristors. Benefiting from memristor noise, the hardware implementation shows continuous periodic behavior, outperforming simulated hardware. This work showcases the potential of biologically-realistic algorithms on emerging hardware for neuromorphic computing. --- ### ReMova: Fine-tuning LLMs for English to Belarusian translation Authors: Mikita Pilinka, Aliaksandr Kliujeŭ, David Samuel, Yves Scherrer Published: 2026-09-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.16427v1 This paper presents a Belarusian-specific data-cleaning pipeline and fine-tuning for English-Belarusian machine translation. Our cleaning pipeline distinguishes itself from others by employing a correction tool that addresses the issue of the two orthographies of the Belarusian language, noise in the training data, interference from other languages and other misspelling issues common in Belarusian on the internet. A matched ablation on unfiltered training data shows substantial benefits from filtering for all fine-tuned models, with the LLM-based models gaining roughly twice as much from filtering as the dedicated encoder-decoder MT system, supporting the view that for Belarusian MT one of the primary bottlenecks is data quality. --- ### How Good Are Time-Series Foundation Models for Pedestrian Crowd Count Forecasting? A Cross-Dataset Comparative Study Authors: Theivaprakasham Hari, Ziteng Li, Yanan Xin, Winnie Daamen, Serge Hoogendoorn Published: 2026-09-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2609.16415v1 Pedestrian-count forecasting supports pedestrian-oriented Intelligent Transportation Systems (ITS), including crowd monitoring, pedestrian-traffic staffing and routing, and proactive risk mitigation during surges. Recent time-series foundation models (FMs) report strong zero-shot accuracy on heterogeneous forecasting benchmarks, but it remains unclear whether these gains transfer reliably to pedestrian sensing deployments. We benchmark seven univariate forecasting approaches spanning four paradigms: Seasonal Naive, gradient-boosted trees (LightGBM, CatBoost), deep learning models (N-HiTS, PatchTST), and two pretrained FMs (TimesFM, Chronos-2). Experiments cover two complementary regimes: (i) a five-day special event dataset SAIL2025 at 3-minute resolution with limited in-domain history; and (ii) Melbourne pedestrian sensors as a multi-year hourly dataset (2010--2017) with strong seasonality. We compare the MAE and RMSE results per sensor across datasets and multiple forecast horizons. Results show three consistent findings. First, with limited historical data, Seasonal Naive remains a strong baseline for long-horizon forecasting on high-volume sensors, while trained models can degrade when the next day differs substantially from prior days. Second, boosted trees can be competitive on lower-volume sensors but exhibit higher sensitivity on high-volume sensors under event-driven shift. Third, FMs excel in the seasonal and data-rich regime under long-context configuration. The findings highlight the importance of choosing pedestrian forecasting models based on both the underlying data conditions and the forecasting horizon. --- ### On the Expressive Power of Implicit Line-Graph Higher-Order Weisfeiler--Leman Authors: Fan Yang Published: 2026-09-14 Categories: cs.SI, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.16412v1 Whitney's theorem allows isomorphism testing for connected simple graphs, apart from $K_3$ and $K_{1,3}$, to be formulated as distinguishing their line graphs. However, the relation between fixed-dimensional Weisfeiler--Leman (WL) expressivity on line graphs and on their roots remains unresolved. We study this relation through Implicit Line-Graph WL (ILG-$k$-WL), which is exactly $k$-WL on $L(G)$, executed over the edges of $G$ with line-graph relations derived from endpoint incidence and without explicitly constructing $L(G)$. On the Whitney-general class, the relation between root-domain and line-graph WL depends on $k$. For $k=1,2$, ILG-$k$-WL adds no distinguishing power beyond root-domain $1$-WL and misses some pairs that $1$-WL separates. For $k=3$, we prove the backward containment $L(G)\equiv_{3\text{-WL}}L(H)\Rightarrow G\equiv_{3\text{-WL}}H$. Strongly regular witness pairs, including the Shrikhande/rook pair, show that ILG-$3$-WL is strictly more expressive than $3$-WL. The backward containment also extends to disconnected graphs with no isolated vertices when every connected component is Whitney-general. Deterministic ILG-$3$-WL separates all three substructure-counting witness pairs, all $105$ pairs in SR25, and $359$ of $400$ BREC pairs. An untrained dense ILG-$3$-GNN gives the same pairwise verdicts on these evaluations. --- ### Reasoning with Image Generation Authors: Nishad Singhi, Hector Garcia Rodriguez, Aditya Arora, Marcus Rohrbach, Anna Rohrbach Published: 2026-09-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.16409v1 Chain-of-thought reasoning has revolutionized natural language processing by enabling large language models (LLMs) to decompose problems into intermediate steps before answering. Yet confining reasoning to the textual domain presents limitations for tasks requiring direct manipulation of visual representations. Recent efforts augment multimodal LLMs with external visual expert tools such as depth estimation or object detection modules, but these remain fundamentally limited by their reliance on narrow, rigid operations that cannot flexibly generate or transform visual content. We propose ReImaGin, which leverages image generation models as a flexible visual reasoning mechanism for multimodal LLMs: unlike fixed-function tools, they accept natural language commands and can perform open-ended visual operations, like removing an occlusion or generating a floorplan from multiple disjoint views of a room. Across six diverse visual reasoning tasks including multi-view spatial reasoning and collision prediction, ReImaGin consistently outperforms both text-only reasoning and specialist vision-tool baselines, with gains of up to 25\%, demonstrating the advantage of flexible, generative visual reasoning. --- ### Balancing Trial and Reorder: A Hybrid Sequential Transformer-GBDT Ranker for On-Demand Delivery Authors: Marcel Kurovski, Attila Nagy, Steffen Klempau, Aleksandr Fedintsev Published: 2026-09-14 Categories: cs.IR, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2609.16407v1 On a delivery platform, personalized store ranking greatly influences what users find and order. Unlike digital-only domains, candidate stores are local and bound by real-time availability and delivery operations. One central modeling tension is between surfacing new stores for trial and preserving ranking quality for sessions with reorder intent. We present Universal Venue Ranker (UVR), a production system deployed at Wolt that pairs a bidirectional transformer encoder for sequential user modeling with a GBDT ranker integrating contextual, user, and store features. Trained across all stores and domains of a country while enforcing local delivery constraints at inference, UVR replaces four previously separate ranking models (three for restaurants, one for retail) with a single unified system. Label smoothing and trial-biased sample weighting steer the model toward new stores, lifting offline trial MRR by +12% to +30% over production while regressing reorder MRR in five of six countries. These regressions leave Global CVR, our core online metric, which blends trial and reorder sessions, statistically unchanged. We validate UVR in three consecutive A/B tests, the first two across Wolt's largest operating markets and the third spanning all operating countries and both domains. UVR V1 delivers +5.5% Merchant Trial Rate and +0.16% Global CVR over the previous production ranker; V2 adds a further +0.45% Merchant Trial Rate on top; and V3, our cross-domain unification of the restaurant and retail rankers, adds a further +1.31% Retail Merchant Trial Rate, together accounting for substantial incremental gross order value and a materially simplified serving stack. --- ### Physics Informed Random Feature Neural Networks for Solving PDEs Authors: Chi-An Chen, Chunyang Liao, Ming Zhong Published: 2026-09-14 Categories: math.NA, cs.LG arXiv: https://arxiv.org/abs/2609.16406v1 Machine learning-based partial differential equations (PDEs) solvers have attracted significant attention in recent years. Most progress in this area has been driven by deep neural networks such as physics-informed neural networks (PINNs) and kernel method (such as physics-informed Gaussian Processes). We introduce a physics-informed random feature method for countering part of the spectral bias which PINN-based solvers are facing for a certain class of PDEs. Random feature method was originally proposed to approximate large-scale kernel machines and can be viewed as a specialized randomized neural network. Compared to other state-of-the-art PINN-based solvers which require a large number of collocation points, our proposed method reduces the computational complexity. In this paper, we develop a rigorous approximation error analysis and derive high-probability error bounds on the $H^1$ norm. We provide extensive numerical tests for verifying our theoretical guarantees on error decay rates, as well as several comparison tests to showcase our claimed capability for combating spectral bias in these deep learning based methods. --- ### Implementing a White-Box Undetectable Backdoor for Random Fourier Features Authors: Michael Collins, Jada Cumberland, Brianne Dunn, Ross Gore, Samuel Jackson, Sachin Shetty Published: 2026-09-14 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2609.16403v1 Goldwasser et al. showed that undetectable backdoors can be planted in machine learning models trained with the Random Fourier Features (RFF) algorithm, under a hardness assumption tied to the Continuous Learning With Errors (CLWE) problem. Under standard cryptographic assumptions, even a full white-box audit of a model's weights cannot detect this class of backdoor. The construction is stated in terms of cryptographic reductions and probabilistic lemmas, without a reference implementation, and relies on secondary machinery such as the Sparse Gaussian Pancakes distribution and a homogeneous CLWE conditional density. Its realizability in ordinary numerical code is not obvious from the paper alone. This paper implements the white-box CLWE-RFF backdoor construction end to end using only numpy and scipy, to test whether this threat is realizable with commodity scientific-computing tools or requires specialized cryptographic infrastructure. We give two samplers for the core $GP_d(b_k)$ distribution. The first is a rejection-sampling proxy. The second is an exact closed-form sampler derived from the homogeneous CLWE density and verified against its own analytic form. Using this implementation, we run statistical indistinguishability tests, covering both weight-space and functional black-box comparisons. We find no evidence of detectable difference between backdoored and clean models across a range of sparsity ratios $ρ= d_{\text{sparse}}/D$. We report which parts of the construction were straightforward to realize, which required derivation not spelled out in the paper. We also highlight which parts we did not attempt to reproduce, including the underlying lattice hardness reduction. We see this work as a contribution to understanding the practical realizability of the Goldwasser white-box CLWE core, not as a new theoretical result. --- ### Negation Beyond the Verbal Channel: Temporal Multimodal Correlates in Dialogue Authors: Leon Hammerla, Patrick Schrottenbacher, Alexander Mehler Published: 2026-09-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.16396v1 Negation is typically modeled through its linguistic realization, although spoken interaction is accompanied by tightly coordinated nonverbal behavior. We ask whether contexts centered on spoken negation cues contain measurable multimodal behavioral information: whether they can be distinguished from matched control contexts without lexical or acoustic input, where this information occurs in time, which modalities carry it, and whether it extends to the dialogue partner. We study 27 human-human interviews conducted in virtual reality, comprising temporally aligned gaze, facial, head, body, hand, and finger behavior and 964 annotated negation cues. Treating classification as a predictive probe, we compare 20 time-series models while excluding lexical and acoustic information, and then systematically vary temporal context, interactional source, modality availability, and event timing. Across grouped 10-fold cross-validation, the strongest probes reach up to .75 mean held-out AUROC from speaker-side behavior. Temporal analyses show that predictive information is concentrated around cue onset but remains detectable over a broader surrounding interval, while dialogue-partner behavior carries weaker predictive information with a comparatively diffuse temporal profile. Ablation and timing perturbations further show that facial features produce the largest modality-ablation effect and that the trained probe is sensitive to the temporal organization of the observed events. --- ### ParsHate: A Benchmark Dataset for Hate and Target Detection in Persian Authors: Zahra Bokaei, Walid Magdy, Bonnie Webber Published: 2026-09-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.16393v1 We introduce ParsHate, a manually annotated dataset of 10,000 Persian tweets spanning 2013-2022, representing the first decade-long benchmark for hate speech detection in Persian. The dataset contains 31% hateful content and supports both hate detection and multi-label fine-grained target identification across seven structured target categories. ParsHate also distinguishes explicit and implicit hate, marks explicit and implicit targets, and provides span-level rationales. Data collection combines random and score-stratified temporal sampling to reduce keyword-driven bias while preserving natural label distributions. Applying SOTA models for Persian hate-speech detection on ParsHate shows moderate performance (79% F1), especially with samples from earlier years, and low performance with target identification (25.5% macro-F1). This emphasizes the diverse sampling of hate speech in ParsHate and its challenging nature that requires more advanced methods for better performance. Dataset is made publicly available. --- ### Where Post-Training Quantization Breaks Text Embedders: A Measured Map Across Four Embedder Families Authors: Hyojung Han Published: 2026-09-14 Categories: cs.IR, cs.CL arXiv: https://arxiv.org/abs/2609.16391v1 Weight-only post-training quantization is the cheapest way to shrink a retrieval embedder, and the received advice for applying it -- protect the embedding table, allocate bits by module sensitivity, prefer a ranking-aware objective over weight reconstruction -- was carried into LLM quantization largely intact. We test that advice on retrieval embedders directly, quantizing five checkpoints from four architecture families across a grid of bit widths and group sizes, and isolating the embedding, attention and feed-forward blocks at each width. Every heuristic fails to transfer as stated. The embedding table never emerges as the dominant isolated protection priority in any family, despite being the largest tensor in several of them. Module sensitivity does not survive as a transferable ordering: at INT4/g16 the spread between modules is too small to allocate against, at INT3 the ordering becomes family-dependent and joint damage stops being the sum of its parts, and at INT2 comparable reconstruction error accompanies retention ranging from 1.3 to 65.9 percent of full precision. A cheap reconstruction proxy is useful for screening uniform bit widths but substantially less reliable for choosing which tensors to protect; its apparent strength across the whole grid is a range-extension artifact. A distilled 109M student at INT3 holds 78.04 NDCG@10 in 68.4 MB and dominates the extreme-PTQ arm of its own 0.6B teacher, 297.9 MB at 64.46, on both size and quality -- but only inside the task it was distilled for. Sizes are byte counts of files that exist rather than arithmetic estimates, and the measurement repository carries the byte provenance for every one of them. --- ### Attention Mean Fields Predict Average Representation Dynamics and Reveal Context-Specific Computation Authors: Micah Adler, John W. Byers, Mark Crovella Published: 2026-09-14 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2609.16382v1 A language model's representation geometry is not predetermined; it evolves as the model runs. A faithful account of that geometry must capture that dynamic process, and so cannot be based solely on model-independent statistics such as co-occurrence. Here we introduce a mean-field analysis of attention. The average attention from one token to another defines a kernel that carries representations layer to layer and can be iterated through the network to model how the geometry is transformed. We condition this average two ways. Conditioned on a whole corpus, the kernel predicts the average-case evolution of representation geometry. Conditioned instead on a single context, it predicts the expected geometry for that context. A head's departure from that prediction, its \emph{mean-field deviation}, isolates the context-specific computation that the mean field misses. Under the corpus-conditional reading, the kernel yields an open-loop model: from the input embeddings and the frozen weights alone, we can iterate the kernel and the model's own MLPs over token representations, never consulting a measured deviation at any layer. The resulting prediction is highly accurate. In early training the model and its corpus mean field are indistinguishable. Replace every attention head with its mean field, and the substitution leaves the loss on real text unchanged. Around the onset of induction, the two diverge, and the gap widens as representations become contextualized. Under the context-conditional reading, deviation from the mean field is a task-agnostic measure of context-specific computation. The residual decomposes additively into unusual attention routing and contextualization of the transported values. Across controlled induction and few-shot settings, greater deviation tracks greater reliance on in-context information. --- ### Bounded Adjustment with Reliability-Guided Embedding for Imbalanced Learning with Noisy Labels Authors: Mushir Akhtar, Akarsh J., M. Tanveer, Mohd. Arshad Published: 2026-09-14 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2609.16380v1 Class-balanced learning and label noise create a coupled failure mode: frequency correction prevents majority classes from dominating the decision rule, but can amplify incorrectly labeled minority examples. We introduce BARGE (Bounded Adjustment with Reliability-Guided Embeddings), a single-stage objective combining a bounded, prior-adjusted density-power score with reliability-guided angular geometry. Its classification score is strictly proper in the adjusted probability space and recovers balanced Bayes ordering under clean supervision and the true class prior. Under label contamination, its finite range bounds classification-risk perturbation at a fixed predictor, while its logit gradient redescends when the model confidently contradicts the supplied label. The adjusted target probability also weights class-equal feature compactness, and a one-sided separation term discourages aligned class directions. BARGE requires neither a noise rate nor a transition matrix, uses one network, and leaves inference unchanged. We evaluate it on CIFAR-10, CIFAR-100, and Tiny ImageNet under long-tail and step imbalance, clean labels, and 20% and 40% random incorrect-label replacement. Across 12 clean settings, BARGE ranks second overall and attains the lowest error in four. Under corruption, it achieves the lowest mean balanced error in all six dataset-corruption settings, reducing the six-setting average from 72.32% for the strongest competitor to 70.00%. It also obtains the highest macro-F1 and macro-AUPRC in every corrupted-label setting. Ablations show that class-equal angular compactness improves on the bounded score alone. These results support bounded predictive influence and reliability-guided geometry as complementary mechanisms for imbalanced learning with uncertain labels. --- ### Geometry vs Structure: Graph-Based Diagnostics for LiDAR Point-Cloud Simulation Fidelity Authors: Ghazal Farhani, Taufiq Rahman Published: 2026-09-14 Categories: cs.RO, cs.CV arXiv: https://arxiv.org/abs/2609.16378v1 Digital twins provide a scalable and cost-effective complement to real-world testing for validating autonomous-driving and advanced driver-assistance system (ADAS) sensor pipelines. However, quantifying their fidelity remains challenging, particularly for 3D LiDAR point clouds, where conventional geometric metrics may overlook important structural discrepancies. We present a graph-based framework for evaluating the structural fidelity of simulated LiDAR point clouds against real-world scans. While scan-level metrics such as Chamfer distance capture point-wise geometric similarity, they do not explicitly represent connectivity, topology, or object-level organization. Our framework constructs graphs from real and simulated point clouds, applies Louvain community detection to identify spatially coherent subgraphs, and matches corresponding communities using centroid proximity. For each matched pair, we compute $r_λ$, a bounded graph-spectral metric motivated by Weyl's inequality, and compare it with density-aware Chamfer distance (CDC) as a geometric baseline. Controlled perturbation experiments demonstrate that $r_λ$ is invariant to rigid transformations and robust to sensor noise while remaining sensitive to structural deformation. We evaluate the framework on 50 paired real and simulated LiDAR scans acquired using a Velodyne VLP-32C sensor and CARLA, respectively. The dataset contains more than 1,000 matched communities across four representative classes: vehicles, vegetation, trees, and building walls. The results show that geometric and structural measures capture complementary aspects of simulation fidelity, supporting graph-spectral analysis as an additional diagnostic layer for validating digital twins in ADAS and autonomous-driving applications. --- ### Certified Uncertainty Propagation in One-Shot Federated Bayesian Models via Posterior Event Transport Authors: Mahyar Mohammadi, Mohammad Hossein Badiei, Abolfazl Yaghmaei, Hamed Kebriaei Published: 2026-09-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2609.16373v1 Probabilistic certification of Bayesian neural networks lower-bounds the posterior probability that a model satisfies a verifier-defined safety property. In one-shot federated Bayesian learning, however, the deployed model is obtained by aggregating parameters drawn from client-specific posterior distributions, so local certificates do not directly guarantee safety of the aggregated model. This paper develops a deployment-consistent certification framework by propagating local posterior events through the deployment aggregation rule, with an exact geometric characterization for Federated Averaging (FedAvg). Each client constructs disjoint hyper-rectangular regions in parameter space and computes their probability masses. The server forms Cartesian products of these regions, maps them through the deployment rule, and retains a product event only when its aggregation image is verified to satisfy the safety property. Under independent client posteriors, each product-event probability factorizes into local masses, and summing verified disjoint events yields a lower bound on safety probability of the deployed model. For FedAvg with nonnegative aggregation coefficients, the image of a Cartesian product of axis-aligned hyper-rectangles is exactly a weighted hyper-rectangle, introducing no set over-approximation. We distinguish the proposed transported-event certificate from direct certification under posterior distributions induced by FedAvg and Product-of-Gaussians aggregation. Experiments on MNIST and Fashion-MNIST under label-Dirichlet heterogeneity show that the transported FedAvg certificate ranges from 22.51% to 46.89%, while direct global certificates range from 72.05% to 91.39%. Results show that predictive accuracy and certifiable safety do not necessarily follow the same trend, and that global posterior constructions can exhibit distinct certification behavior across architectures. --- ### Register Tokens for Bounded-State Reasoning in Diffusion Language Models Authors: Albert Ge, Chandan Singh, Yufan Zhuang, Xiaodong Liu, Jianfeng Gao, Frederic Sala Published: 2026-09-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.16372v1 Masked diffusion language models (dLLMs) generate text by iteratively denoising masked tokens with bidirectional attention. Extending reasoning across generation chunks normally requires keeping earlier generated text in context. We ask whether a dLLM can instead continue reasoning after that text is cleared, using only a fixed-size carried state. We implement this state as a small number of register tokens: dedicated fixed-position tokens whose continuous hidden states are trained to carry reasoning progress across generation chunks. We post-train dLLMs to decode a chunk of text, clear it while preserving the register values, and continue decoding from the prompt and carried state. In our main comparisons on LLaDA and Dream, registers outperform discrete-text carry on every benchmark, with gains of up to 8.5 points on math and 19.5 points on code. Registers are especially effective for bounded code generation, where correct programs usually span several chunks. Finally, registers can be further refined with reinforcement learning on long-horizon reasoning tasks. ---