# Research Papers — 2026-06 · bundle page 24 of 260 > Full-text markdown of papers #1151–1200 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/24 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/23.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/25.md) ### SCARCE: Scalable Cascade Analysis for Rare-event Characterisation via Embeddings Authors: Yingjie Wang, Yi Dong, Edmund Lau, Jie Meng, Taylor T Johnson, Xiaowei Huang Published: 2026-06-28 Categories: cs.AI, cs.LG arXiv: https://arxiv.org/abs/2606.29623v1 Rare events govern the safety profile of modern AI systems, yet their probabilities are extremely difficult to estimate: direct Monte Carlo requires prohibitive sample budgets. Subset Simulation (SS) addresses this by decomposing a rare-event probability into moderate conditional probabilities over nested intermediate events. However, classical SS requires a handcrafted scalar performance function whose sublevel sets define those events, demanding detailed knowledge of the failure geometry and limiting transfer to new domains. We propose SCARCE (Scalable Cascade Analysis for Rare-event Characterisation via Embeddings), which replaces the performance function with learned latent representations and geometric rulers that score proximity to failure regions. Adaptive thresholding constructs nested intermediate events directly from data. We formalise SCARCE through a non-negative supermartingale, yielding a high-probability upper envelope that remains valid under early stopping. On MNIST misclassification, where dense Monte Carlo provides ground truth, SCARCE achieves approximately 400--500 times lower mean absolute error than grid-searched traditional SS while eliminating systematic over-counting. We then study PAIR-style LLM jailbreaks under a fleet-level threat model with adversarial fraction $η$. On Llama-Guard-3-8B hidden states, a PCA-based ruler attains 2.6% mean relative error for $η\geq 10^{-3}$ against finite-sample references whose average bootstrap relative half-width is 27.9%, and transfers to a GCG-style corpus with 2.93% relative error after recalibration. A directional criterion $\mathrm{KL}(p_{\mathrm{good}}\,\|\,p_{\mathrm{bad}})$ ranks rulers consistently with estimation error (Spearman $ρ=0.83$). --- ### Bidirectional Autoregressive Latent Diffusion for Forward and Inverse Magnetohydrodynamics Authors: Alexander Scheinker Published: 2026-06-28 Categories: stat.ML, cs.LG, nlin.AO, physics.plasm-ph arXiv: https://arxiv.org/abs/2606.29620v1 This work presents a new bidirectional autoregressive latent diffusion approach for predicting the evolution of multiple fields (mass density, pressure, velocity, and magnetic field components) for magnetohydrodynamics. We show that this bidirectional flow can be used as a self-supervised consistency metric for uncertainty and error estimation, which enables the model to estimate test-time uncertainty and error without access to ground truth, by comparing how closely flowing forwards and backwards in time returns to the same predicted fields. We also demonstrate this methods's potential to serve as a non-invasive plasma diagnostic, and show how adaptive feedback can be used to make the model more robust based on sparse diagnostics or limited views/measurements. --- ### Do We Still Need Fine Tuning? Turkish Sentiment Analysis in the Era of Large Language Model Authors: Sercan Karakaş, Yusuf Şimşek Published: 2026-06-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.29614v1 This study examines whether supervised fine-tuning remains necessary for Turkish sentiment analysis in the era of large language models. We compare classical machine learning methods, fine-tuned pretrained language models, and prompted large language models on a Turkish e-commerce review dataset with negative, neutral, and positive labels. Fine-tuned BERTurk models perform best overall and outperform all prompted large language models in the full three-class task. The neutral class emerges as the main difficulty: while several large language models are much more competitive in binary positive--negative classification, they degrade substantially in the three-class setting by collapsing neutral reviews into polarized categories. The findings suggest that, in realistic Turkish sentiment classification, prompted large language models do not yet match supervised fine-tuning in the zero-shot setting, and that including the neutral class is crucial for robust evaluation. --- ### Does Role Specialization Matter for Explanation Faithfulness in Mixture-of-Experts? Authors: Yeji Kim, Housam Babiker, Mi-Young Kim, Randy Goebel Published: 2026-06-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.29613v1 Mixture-of-Experts (MoE) architectures have recently been extended with role-based mechanisms for interpretability. This is typically done by assigning semantic roles to individual expert components, for example roles like synergy, redundancy, and uniqueness in multimodal settings. However, whether such structural role decomposition preserves explanation faithfulness of the overall architecture remains largely underexplored. We hypothesize that inter-expert representation overlap weakens effective role separation and degrades attribution-based faithfulness, even when semantic roles are explicitly defined. To address this limitation, we introduce representation-level decorrelation regularization to explicitly reduce inter-expert similarity in latent space. Using representation decorrelation objectives, we encourage clearer specialization among experts by minimizing representation overlap. Our experiments show that across multiple multimodal benchmarks, this separation consistently improves explanation faithfulness, as measured by comprehensiveness, sufficiency, and their Area Over the Perturbation Curve (AOPC) summaries, while preserving task performance. We further show that these improvements are not limited to role-based architectures such as Interpretable Multimodal Interaction-aware MoE (I2MoE). Similar trends are observed in a standard sparse MoE baseline, suggesting that representation-level separation may provide a more general mechanism for enhancing explanation faithfulness in MoE systems. Overall, our findings suggest that structural role decomposition alone may be insufficient to guarantee faithful explanations and that representation-level separation helps improve explanation faithfulness. To support reproducibility, the source code and supplementary material are publicly available at https://github.com/dut0817/FL-I2MoE_Decor. --- ### How much of an LLM-generated clinical corpus is actually new? A production-scale measurement of content redundancy for provenance classification Authors: Ali H. Lazem, William J. Teahan Published: 2026-06-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.29605v1 Clinical machine learning increasingly relies on training corpora generated by large language models (LLMs) rather than annotated by clinicians, and such corpora are described and reused largely on the basis of their reported scale. We test whether volume reflects information content. Analysing the complete output of a multi-agent clinical extraction pipeline applied to 167,034 patient narratives, 2.51 billion generated tokens across the ten text-bearing channels of an eleven-channel pipeline, we introduce Provenance-based Redundancy Decomposition, a token-level classification of the entire output by source. Only 10.9% of the output is trainable-unique content while 79.4% is redundant; raw token count overstates information content by roughly ninefold. The redundancy arises through two distinct mechanisms, verbatim copying of source context into per-item fields, and duplication of generated text across records, of which only the former is losslessly removable. An independent, model-free analysis based on lossless compression confirms the redundancy, recovering the two mechanisms without reference to the provenance labels. One pipeline channel carries almost no redundancy, showing that the level of redundancy depends on how each channel is structured rather than being a fixed property of LLM extraction. Because uncorrected redundancy up-weights the longer, more complex presentations that generate the most items, it skews the token-level training distribution of the corpus, a property we measure directly. In a controlled downstream test, de-duplicating the corpus before adaptation improved a clinical encoder on external disease-recognition benchmarks at equal token budget, robustly across adaptation depths and replicated on a second benchmark, confirming that the redundancy carries a measurable cost beyond storage. The classification tool is released openly. --- ### Mechanistically Eliciting Latent Behaviors in Language Models Authors: Andrew Mack, Nina Panickssery, Alexander Matt Turner Published: 2026-06-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.29604v1 We aim to discover diverse, generalizable perturbations of LLM internals that can surface hidden behavioral modes. Such perturbations could help reshape model behavior and systematically evaluate potential risks. We introduce Causal Perturbative Elicitation (CPE), an unsupervised method for discovering interpretable low-rank adapters (LoRAs) that can elicit these latent behaviors. CPE decomposes the computations of a deep transformer slice using a heuristic tensor-decomposition-based algorithm. CPE exhibits remarkable data efficiency, learning a large number of interpretable LoRAs from a single example. Even though CPE is unsupervised, we find that in some cases it can be competitive with supervised elicitation methods via brute-force enumerative search over weight space. For instance, CPE performs similarly to matched-wall-clock-time GRPO on the Countdown task for Qwen3-8B (85% vs 87%), demonstrating that CPE can efficiently elicit complex multi-token behaviors. Since CPE is unsupervised, it can also surface hidden failure modes, such as sandbagging, restoring 85% of locked BigCodeBench performance on a password-locked version of Llama3-70B introduced by Taylor et al. (2025). Additionally, since CPE explores behaviors in weight-space rather than token-space it can potentially ameliorate exploration hacking, a misalignment failure which may arise in sufficiently self-aware AI models (Ngo, 2022). In fact, we find that CPE virtually eliminates alignment-faking (Greenblatt et al., 2024) behavior in a Llama3-70B-based model organism developed by Hughes et al. (2025). Finally, we find that CPE can be used to initialize GPT-OSS-20B in an aligned basin when running GRPO on an environment prone to reward-hacking. By providing a data-efficient method to systematically explore the space of latent model behaviors, CPE yields a powerful tool for aligning AI systems and evaluating their safety. --- ### Langshaw: Declarative Interaction Protocols Based on Sayso and Conflict Authors: Munindar P. Singh, Samuel H. Christie, Amit K. Chopra Published: 2026-06-28 Categories: cs.PL, cs.AI, cs.MA arXiv: https://arxiv.org/abs/2606.29601v1 Current languages for specifying multiagent protocols either over-constrain protocol enactments or complicate capturing their meanings. We propose Langshaw, a declarative protocol language based on (1) sayso, a new construct that captures who has priority over setting each attribute, and (2) nono and nogo, two constructs to capture conflicts between actions. Langshaw combines flexibility with an information model to express meaning. We give a formal semantics for Langshaw, procedures for determining the safety and liveness of a protocol, and a method to generate a message-oriented protocol (embedding needed coordination) suitable for flexible asynchronous enactment. --- ### One Scene, Two Depths: Probing Geometric Ambiguity in Monocular Foundation Models Authors: Xiaohao Xu, Feng Xue, Xiang Li, Haowei Li, Shusheng Yang, Tianyi Zhang, Matthew Johnson-Roberson, Xiaonan Huang Published: 2026-06-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.29600v1 A faithful 3D world representation should account for layered geometry, where a single camera ray may contain multiple visible and geometrically valid surfaces. Monocular depth estimation, however, reduces this structure to one scalar depth per pixel. Transparent scenes make this ambiguity measurable: the same ray can pass through foreground glass and observe the background, turning the supervised target into a convention of annotation, data, and training rather than a scene-intrinsic truth. A learned predictor exposes this convention as its depth-layer preference. We introduce MultiDepth-3k (MD-3k), a sparse two-layer ordinal benchmark for measuring depth-layer preference and multi-layer spatial relationship accuracy (ML-SRA). On MD-3k, leading depth foundation models exhibit diverse layer preferences under standard RGB input, showing that the same layered geometry can be resolved differently across models. We further find that Laplacian Visual Prompting (LVP), a training-free spectral input transformation, can substantially change the reported layer for certain frozen models. The strongest RGB/LVP pair, DAv2-L, reaches 75.5% ML-SRA. These results suggest that depth foundation models may express complementary geometric hypotheses that standard RGB inference leaves unexpressed. We invite the community to rethink depth supervision and evaluation through an ambiguity-aware lens, where multiple valid 3D interpretations are treated as geometric structure to be measured, preserved, and expressed. --- ### Boundary Degree as a Node-level Feature for Epidemic Scenario Identification in Agent-based Cascade Simulations Authors: Amro Alabsi Aljundi, Galen Harrison, Jiangzhuo Chen, Abhijin Adiga, Anil Kumar Vullikanti, Madhav V. Marathe Published: 2026-06-28 Categories: cs.SI, cs.LG arXiv: https://arxiv.org/abs/2606.29596v2 Characterizing the scenario underlying an epidemic from its disease cascade is an important task in simulation analytics. We propose boundary degree, the count of an infected node's contacts in the underlying contact network that were not infected, as a per-node cascade feature for this task. Through systematic ablation on realistic social contact networks of Tennessee and Virginia, we show that boundary degree alone improves scenario identification accuracy by 19%. Edge features, whose importance was observed empirically by prior work, consistently improve accuracy across all settings; we provide theoretical grounding for this observation. These effects are complementary. We prove that certain epidemic scenarios are indistinguishable without boundary or edge information. Prior feature engineering approaches included aggregate boundary statistics, but these were not among the top-ranked feature groups; the per-node representation we propose reveals their importance clearly. Our results suggest that contact tracing applications should track contacts with non-infected individuals, not only transmissions. --- ### How AI settled the complexity of the oldest SGD algorithm Authors: Michał Dereziński, Xiaoyu Dong Published: 2026-06-28 Categories: cs.LG, cs.AI, math.NA, math.OC, stat.ML arXiv: https://arxiv.org/abs/2606.29593v1 In 1937, Stefan Kaczmarz proposed a simple algorithm for solving systems of linear equations. This algorithm turned out to be the earliest known example of stochastic gradient descent, a ubiquitous computing paradigm that drives the training of modern AI models such as ChatGPT and Gemini. Now, those AI models have joined forces to discover the worst-case complexity of the Kaczmarz algorithm. This paper tells the story of how it happened. --- ### STEMGym: Benchmarking Sequential Decision-Making under Dose Budgets in Autonomous Electron Microscopy Authors: Can Polat, Erchin Serpedin, Mustafa Kurban, Hasan Kurban Published: 2026-06-28 Categories: cs.LG, cond-mat.mtrl-sci, physics.atom-ph, physics.optics, quant-ph arXiv: https://arxiv.org/abs/2606.29592v1 A central premise of autonomous scientific imaging is that smarter navigation, whether Bayesian, RL-based, or otherwise adaptive, is the principal lever for sample-efficient acquisition. We present evidence to the contrary in scanning transmission electron microscopy (STEM), an atomic-resolution imaging modality whose every measurement deposits damaging electron dose. We introduce STEMGym, an open-source Gymnasium benchmark of 15 physics-simulated STEM worlds spanning five materials, three difficulty levels, and four characterisation tasks, scored by the Dose-Efficiency Curve area (DEC-AUC), a single scalar capturing the information-vs-dose Pareto frontier. Across 33 agent configurations under realistic dose budgets, the dominant determinant of dose efficiency is the analyst (perception) pipeline, not the navigator: pairing a trained CNN analyst with naïve raster scanning raises DEC-AUC by 5.5x over a CNN-free raster baseline (0.287 vs.\ 0.052), while substituting Bayesian or adaptive finite-state-machine navigation for raster yields no statistically significant further gain. Production-tier vision-language models further underperform task-specific CNNs by {\sim}13x on crystallographic defect analysis. By decoupling perception, navigation, and planning under a unified dose budget, STEMGym reframes where ML effort should be invested in autonomous electron microscopy and provides the measurement infrastructure to test it. --- ### SonoCLIP: Mask-Guided Region-Aware Vision-Language Pretraining for Fetal Ultrasound Analysis Authors: Hang Su, Chao Sun, Zhaofan Li, Wei Hu, Juhua Liu, Bo Du Published: 2026-06-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.29586v1 Vision-language foundation models have shown strong potential in medical image analysis. Although foundation models for ultrasound imaging have recently emerged, the domain remains particularly challenging due to severe speckle noise, acquisition variability, and subtle anatomical boundaries, leading to high inter-observer variability. Existing CLIP-based models rely primarily on global image-text alignment, limiting their sensitivity to clinically decisive local structures. We propose SonoCLIP, the first million-scale region-controllable fetal ultrasound vision-language foundation model that integrates segmentation masks as mask-channel visual prompts within the vision encoder, enabling joint global-local contrastive representation learning. To support scalable region-text alignment, we introduce a sigmoid-based pairwise contrastive loss that improves stability under large-scale supervision. We further curate a 1.44M-image multimodal fetal ultrasound dataset spanning 24 standard planes for large-scale pretraining. Extensive cross-center evaluations demonstrate that SonoCLIP achieves superior zero-shot transfer performance under both global and mask-guided inference, establishing a controllable and clinically oriented foundation model for fetal ultrasound analysis. Our code and data are available at https://github.com/Harrison-one/SonoCLIP. --- ### Geometric Algebra Meets Cartesian Tensors: Higher-Order Equivariance for Interatomic Potentials Authors: Can Polat, Erchin Serpedin, Mustafa Kurban, Hasan Kurban Published: 2026-06-28 Categories: physics.chem-ph, cond-mat.mtrl-sci, cs.LG arXiv: https://arxiv.org/abs/2606.29584v1 $\mathrm{Cl}(3,0)$ interatomic potentials, despite their algebraic elegance, predict force magnitudes accurately but force directions poorly. Across ten rMD17 molecules, every $L \leq 1$ baseline in our twelve-model study attains aggregate force-cosine similarity below $0.25$. The cause is structural. The geometric product of two vectors in $\mathbb{R}^3$ realises only the $L=0$ and $L=1$ components of its irreducible representation content, leaving the symmetric-traceless rank-2 component absent from the per-edge bilinear that drives each message-passing layer. We address this with CliffordSTF, which couples the Clifford multivector to closed-form symmetric-traceless tensor tracks at ranks two and three through bilinear cross-track contractions, using a single learned bilinear and no Clebsch--Gordan tables, Wigner-$D$ matrices, or e3nn calls. On rMD17, CliffordSTF raises aggregate force-cosine similarity from $0.055$ (base Clifford) to $0.551$, an order-of-magnitude relative directional gain, alongside improved magnitude accuracy (force MAE $15.8\%$ lower; energy MAE $10.9\%$ lower). It outperforms all CG-free or body-ordered baselines in our study (all $\leq 0.17$). On catalysis benchmarks, CliffordSTF achieves the best out-of-distribution S2EF energy MAE on OC22 in our experiments, and the best in-distribution energy MAE among $L \geq 2$ methods on OC22 IS2RE. An eleven-variant ablation shows the two tracks are complementary: neither alone matches the combined model. --- ### Bilevel Optimization for Neural Architecture Search Authors: Abhishek Shukla, Ankur Sinha, Faiz Hamid Published: 2026-06-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.29582v1 Bilevel optimization has become an influential and widely adopted framework for addressing hierarchical optimization problems in machine learning, providing an effective approach to modeling the interaction between two levels of optimization, with applications such as hyperparameter tuning, meta-learning, adversarial training, and data poisoning. Neural Architecture Search (NAS), a subfield of hyperparameter optimization, is a prime example of a bilevel optimization problem, with architecture parameters optimized at the outer-level and network weights optimized at the inner level. This paper presents a structured overview of NAS through the lens of bilevel optimization. We categorize existing NAS approaches into two main classes: sampling-based methods, which search optimal architectures using different architecture samplers, and bilevel theory-based methods, which solve the architecture search problem using bilevel optimization principles. We further highlight our current research direction, wherein the bilevel NAS formulation is addressed through an auxiliary mathematical programming framework. This framework enables the systematic integration of second-order information from the model's training loss function and ensures the optimality of the model parameters while modifying architecture parameters. By simultaneously updating the architecture and model parameters along their respective optimal descent directions derived from the auxiliary mathematical program, these methods achieve more principled and theoretically consistent results. The same auxiliary program can also be used for simultaneous hyperparameter and model fine-tuning. A comparative analysis shows that bilevel theory-based approaches generally outperform sampling-based methods, both in accuracy and efficiency. --- ### The Joint Effect of Quantization and Sampling Temperature on LLM Safety Alignment: A Factorial Analysis Authors: Hari Prasad, Ritam Pal Published: 2026-06-28 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.29581v2 Modern LLM deployments often combine quantization with higher sampling temperatures to reduce cost, latency, or repetition, yet safety evaluations usually treat these as fixed implementation details. We test whether models that are safe at FP16 with greedy decoding remain safe after quantization and stochastic sampling, or whether the two factors amplify each other. We evaluate 8 instruction-tuned models from five families across 3 precisions and 6 temperatures, covering 144 configurations on 7 harmfulness benchmarks and generating about 2.0 million responses, which are scored by a six-judge safety ensemble. Contrary to concerns that low-bit deployment erodes alignment, we find that standard quantization is approximately safety-neutral: for 7 of 8 models, AWQ INT4 keeps attack success within about 1.6 percentage points of FP16 or lowers it, with clear degradation only for SmolLM3-3B (34.5% to 44.1%). However, the larger risk comes from sampling: higher temperatures sharply increase decision instability, with DFR reaching 41.9% at T = 1.0, even when average ASR changes only modestly. The two factors do not compound: our Compound Degradation Index remains sub-additive (-0.071 to +0.018), indicating that quantization partially offsets rather than amplifies temperature-induced degradation. Finally, a per-benchmark breakdown shows that single-benchmark evaluation badly understates risk: several models scoring 0% on AdvBench exceed 80% on ManyHarm. Standard INT4/INT8 quantization can therefore be reasonable for well-aligned models, but safety claims should report multi-sample stability across multiple benchmarks rather than rely on a single benchmark at greedy decoding. --- ### MAM-AI: An On-Device Medical Retrieval-Augmented Generation System for Nurses and Midwives in Zanzibar Authors: Yi Ren Published: 2026-06-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.29580v3 Maternal and newborn mortality remain among the highest in sub-Saharan Africa, where midwifery care is often delivered by nurses who lack midwifery training to international standards, and consulting authoritative guidance at the point of care is hard: the guidelines are long and connectivity is intermittent. We present MAM-AI, a medical question-answering assistant for nurse-midwives in Zanzibar that runs entirely on a commodity Android device: a question is embedded (EmbeddingGemma, 300M) and matched against a curated corpus of 87 guideline documents (63,650 passages), then answered with citations by a 4B int4 generator (Gemma 4 E4B), fully offline, with no query leaving the device. We evaluate the exact deployed configuration with a layered methodology -- retriever, generator under oracle context, end-to-end, and latency -- scored by LLM judges validated against physician rubrics. The evaluation relocates the hard problem. On-device retrieval is essentially solved: the 300M embedder ranks third of seven retrievers and rivals cloud systems, so the passages the system needs are usually found. The small generator is what remains in doubt: adding retrieved context does not improve its answers, and at 4B it cannot be both helpful and safe at once -- of two same-size candidates, the more helpful one commits genuine dangerous errors, so we deploy the other, which is about twice as faithful to its sources (as faithful as a frontier model), and recover its helpfulness with a redesigned prompt that cuts deflection from 33% to 3%. Corpus quality is decisive for the same reason: where the corpus holds the right passage the answer is specific and actionable, and where it does not it goes vague. MAM-AI is a thoroughly evaluated, open-source research prototype, not a fielded product; the system, knowledge base, benchmarks, and evaluation harness are released. --- ### ScAle: Attention Head Scaling as a Minimal Adapter for Spatial Reasoning in Vision Language Models Authors: Rahul Chowdhury, Timothy A Rupprecht, Xuan Shen, Pu Zhao, Yanzhi Wang Published: 2026-06-28 Categories: cs.CV, cs.AI, cs.LG, cs.MM arXiv: https://arxiv.org/abs/2606.29579v1 Spatial reasoning remains a persistent challenge for many vision language models (VLMs), and improving it typically requires fine-tuning with substantial additional parameters. Our preliminary analysis reveals that rescaling activations in selected transformer layers-without modifying pretrained weights-can significantly influence downstream performance. Motivated by this observation, we propose ScAle, an ultra-lightweight adaptation method that learns a small set of scalar coefficients to modulate last-token attention and MLP activations in a fully frozen backbone. We evaluate our method on the synthetic spatial reasoning benchmark SpatialEval and on real-world VQA datasets (COCOQA and VGQA) across multiple model families. Our method, ScAle, achieves up to 134.1% relative accuracy gains using only 1K trainable parameters without requiring millions of trainable parameters as in standard PEFT methods such as LoRA. Despite its extreme compactness, our approach recovers a substantial fraction of standard PEFT performance while preserving strong non-spatial VQA accuracy. These results demonstrate that bounded activation reweighting provides a simple, architecture-agnostic, and highly parameter-efficient alternative for adapting pretrained VLMs. --- ### ReMAP-PET: Beyond Visual Understanding -- Learning Region-Guided Metabolic Alignment Semantics from Brain PET Authors: Dasen Dai, Yanteng Zhang, Shuoqi Li, Yuxiang Wei, Hongjie Yu, Qingxin Zhang, Qizhen Lan, Jagath C. Rajapakse, Vince D. Calhoun Published: 2026-06-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.29577v1 Positron Emission Tomography (PET) reveals brain metabolism and is clinically central to neurodegenerative disease assessment, yet existing 3D brain foundation models treat PET as generic volumetric data, missing the structured regional metabolic information that distinguishes it from structural neuroimaging. To address these limitations, we propose ReMAP-PET, a framework that moves beyond visual encoding by supervising a partially-tuned MedicalNet 3D ResNet-50 with brain regional standardized uptake value ratio (SUVR) profiles through joint regression and contrastive objectives, enabling the encoder to learn the metabolic semantics underlying PET modality. On 1015 paired PET--SUVR samples, ReMAP-PET achieves 0.070 SUVR MAE and 77.8% PET SUVR Recall@1, substantially outperforming five frozen pretrained baselines. We further connect the metabolic embedding to clinical language via contrastive alignment with frozen BioClinicalBERT and demonstrate end-to-end PET-to-report generation through SUVR-constrained verbalization. Linear probing on diagnostic classification and cognitive regression tasks confirms that the embeddings retain clinically relevant information without task-specific fine-tuning. Our results show that grounding PET encoders in regional metabolic semantics -- rather than treating PET as generic volumetric data -- yields representations that are structured, interpretable, and language-compatible, pointing to a new direction for metabolic-aware PET understanding. --- ### TF-MoE: Time-Frequency Mixture-of-Experts for Efficient Speech Separation Authors: Qinzhe Hu, Chenda Li, Wangyou Zhang, Shujie Liu, Yan Lu, Yanmin Qian Published: 2026-06-28 Categories: cs.SD, cs.AI arXiv: https://arxiv.org/abs/2606.29575v3 Recent advances in speech separation (SS) have led to compact front-end models with small parameter sizes, yet their high computational cost remains a major barrier for deployment on edge devices. To address this, we propose TF-MoE, a sparse Mixture-of-Experts (MoE) framework that enhances model capacity with almost no increase in inference cost. Our method introduces dynamic expert specialization in time and frequency dimensions through alternating time-wise and frequency-wise MoE modules, each dynamically selecting experts per frame or mel band. Built upon a mel-band-splitting Conformer backbone, TF-MoE achieves strong performance on SS tasks under low-compute settings. Experimental results demonstrate that TF-MoE consistently improves separation performance under computation cost constraints, outperforming BSRNN by +3.8 dB SDR on Libri2Mix with comparable 4.1 GMACs/s inference cost. This positions TF-MoE as a promising candidate for edge-device deployment. --- ### TF-MoE: Time-Frequency Mixture-of-Experts for Efficient Speech Separation Authors: Qinzhe Hu, Chenda Li, Wangyou Zhang, Shujie Liu, Yan Lu, Yanmin Qian Published: 2026-06-28 Categories: cs.SD, cs.AI arXiv: https://arxiv.org/abs/2606.29575v2 Recent advances in speech separation (SS) have led to compact front-end models with small parameter sizes, yet their high computational cost remains a major barrier for deployment on edge devices. To address this, we propose TF-MoE, a sparse Mixture-of-Experts (MoE) framework that enhances model capacity with almost no increase in inference cost. Our method introduces dynamic expert specialization in time and frequency dimensions through alternating time-wise and frequency-wise MoE modules, each dynamically selecting experts per frame or mel band. Built upon a mel-band-splitting Conformer backbone, TF-MoE achieves strong performance on SS tasks under low-compute settings. Experimental results demonstrate that TF-MoE consistently improves separation performance under computation cost constraints, outperforming BSRNN by +3.8 dB SDR on Libri2Mix with comparable 4.1 GMACs/s inference cost. This positions TF-MoE as a promising candidate for edge-device deployment. --- ### Reliability-Prioritized Fine-Grained Generation in Multimodal Large Authors: Xiaomeng Fan, Wei Wu, Yuwei Wu, Zhi Gao, Shiyu Luo, Mingyang Gao, Haoyu Zhao, Zhenxin Diao, Yuxuan Ba, Lijia Feng, Yunde Jia, Mehrtash Harandi Published: 2026-06-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.29573v3 Multimodal large language models (MLLMs) are increasingly expected to generate fine-grained descriptions of visual content. However, we observe and theoretically show that generating fine-grained responses poses a reliability challenge, \textit{i.e.}, fine-grained generation is more error-prone than coarse-grained generation. This phenomenon suggests that models should generate the finest description that remains reliable rather than simply produce more specific outputs. To investigate this problem, we develop \textsc{GranFact}, a granularity-aware benchmark consisting of expert-verified multi-object images with coarse-to-fine category annotations. Then, we design a hierarchy-aware evaluation algorithm, which assesses both whether model predictions are visually correct and how specific the correct predictions are. We also propose a reliability-prioritized preference optimization method based on Direct Preference Optimization, which penalizes unreliable fine-grained claims while rewarding reliable specificity. Experiments on \textsc{GranFact} show that our method improves fine-grained generation while preserving reliability. Code and data are available \href{https://github.com/WeiWu2025/GranFact}{here}. --- ### Anisotropy Decides Cosine vs. Rank Metrics for Text Embeddings Authors: V. S. Raghu Parupudi Published: 2026-06-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.29571v1 The standard way to compare two text embeddings is cosine similarity. Scattered studies report that a different metric does better, but never pin down the geometric condition that decides when, or why. We settle both with a comprehensive empirical study: nineteen parameter-free similarity metrics on nineteen encoders, from compact sentence transformers up to seven-billion-parameter large language models, across seven datasets. The answer is geometric. When an encoder spreads its variance evenly across directions, cosine is the best parameter-free choice and no other metric helps by a usable margin. When the variance concentrates into a few dominant directions, a property known as anisotropy, rank-based and L1-type metrics beat cosine by a clear margin. The absolute gain is modest, but because cosine starts low on these encoders it is a sizable relative improvement, around twenty percent on average and largest where cosine is weakest. What decides this is the geometry of the embedding space, not how the model was trained: where the two disagree, the metric follows the geometry. One number, the fraction of variance held by the single most dominant dimension, predicts how much the alternatives help across all nineteen encoders, with a rank correlation of 0.86 and a linear correlation of 0.95. To test this as the cause rather than a correlate, we project out the dominant directions: cosine recovers and the advantage of the other metrics nearly vanishes, but only on the encoders that were anisotropic to begin with. The effect is directional, not magnitude based, since it survives normalizing every vector to unit length. Among parameter-free metrics, then, cosine is the right tool wherever an encoder is well spread, which includes the fine-tuned embedders commonly deployed for retrieval, and we give a one-number diagnostic for when it is not. --- ### SurrogateShield: Beyond Redaction for High-Utility, Privacy-Preserving LLM Interactions Authors: Sherwin Vishesh Jathanna Published: 2026-06-28 Categories: cs.CR, cs.CL arXiv: https://arxiv.org/abs/2606.29567v1 LLM-based assistants transmit user queries verbatim to third-party API endpoints that lie outside the user's audit or control. When those queries contain personally identifiable information (PII), the data persists on remote infrastructure subject to breach, subpoena, or policy change. Placeholder redaction (the prevailing mitigation) suppresses PII at the cost of semantic coherence, producing structurally degraded queries and correspondingly degraded responses. We present SurrogateShield, a client-side proxy that substitutes detected PII with locally generated, type-consistent surrogate values prior to transmission and restores originals in the response. No real PII crosses the network boundary. Detection runs through a three-stage cascade (PatternScan, EntityTrace, and ContextGuard) covering 22 PII types and quasi-identifier combinations grounded in Sweeney's k-anonymity framework. Surrogate-to-original mappings are sealed in an AES-256-GCM encrypted per-conversation ShadowMap that never leaves the device. Evaluations on a 1,124-query corpus demonstrate that the cascade reliably detects PII, achieving an overall F1 score of 98.87%. Surrogate substitution substantially outperforms placeholder redaction in semantic utility, yielding a 13.26 pp improvement in BERTScore (roberta-large), from 81.59% to 94.85%. Within this corpus, the local pipeline restricted real PII transmission across all tested query types; in a 100-query adversarial trial, a prompted LLM adversary recovered no original values from surrogate-substituted messages. --- ### Speculative Pre-Positioning: Decoding Stateful Sessions to the Next Decision Point Off the Critical Path Authors: Victor Norgren Published: 2026-06-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.29565v1 A stateless inference server (vLLM, SGLang, TensorRT-LLM) idles between requests while the accelerator waits; a stateful session reclaims that idle time. Speculative pre-positioning decodes the session forward to its next decision point with the target model's own forward pass and no draft model, moving the cross-request prefill and entry-decode off the critical path: the next request resumes from a pre-paid entry on its delta, or, when a confidence gate fires, is answered from a cached distribution in one near-constant vocabulary scan with no decode, at a cost only of energy and a rare, bounded false accept. The payoff is conditional on capability: a capable model fires the gate at near-full coverage and about 87% precision (a smaller one never clears it), returning the first token in about 1.0 ms versus the 39 ms decode a prefix cache still pays. --- ### Coverage-Driven KV Cache Eviction for Efficient and Improved Inference of LLM Authors: Shuvendu Roy, Mengyao Zhai, Hossein Hajimirsadeghi, Golnoosh Samei Published: 2026-06-28 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.29563v1 Large language models (LLMs) excel at complex tasks like question answering and summarization, thanks to their ability to handle long-context inputs. However, deploying LLMs is costly, not only due to the high computational demands of quadratic complexity of self-attention and auto-regressive generation, but also because of the significant memory overhead required for storing the key-value (KV) cache during inference. To reduce the memory cost, existing KV-cache eviction strategies leverage the sparsity in attention to selectively store a subset of tokens. While reducing the memory footprint, such approaches show a considerable drop in performance, especially in tasks that require long-context reasoning. We identify that the drop in performance is linked to a reduction in the coverage of unique tokens. Additionally, we theoretically show that reduced coverage limits the mutual information between inputs and outputs, thereby impairing predictive accuracy. To this end, we introduce K-VEC, a novel coverage-aware KV-cache eviction strategy that prioritizes token coverage while evicting tokens in the cache. K-VEC introduces a cross-head and a cross-layer coverage module to enhance token retention across attention heads and model layers, mitigating performance degradation caused by low coverage. Evaluated on 16 LongBench subsets, K-VEC exhibit up to 10.35 points improvement over the existing methods under the same eviction rate and memory constraint. Comprehensive evaluations validate the effectiveness of our approach and demonstrate its potential for efficient LLM deployment in resource-constrained settings. --- ### Persona-Trained Monte Carlo: Estimating Market-Outcome Distributions via Swarms of Persona-Conditioned Neural Policy Bots in a Limit Order Book Authors: Salavat Ishbulatov Published: 2026-06-28 Categories: cs.LG, cs.MA arXiv: https://arxiv.org/abs/2606.29556v1 We propose Persona-Trained Monte Carlo (PTMC), a method for estimating distributions of market-outcome statistics by repeatedly simulating limit-order-book interaction among swarms of persona-conditioned neural-policy trading bots. Each run instantiates many bots sharing one trained policy network but conditioned on heterogeneous, individually sampled persona parameters drawn from a learned trader-heterogeneity distribution; the bots interact in a continuous double auction, and the resulting price path is one Monte Carlo sample. Repeating this over independent persona-population draws yields an ensemble from which a target market statistic is estimated. Randomness enters through persona draws, within-run action sampling, and optional exogenous shocks, not solely through price as in classical Monte Carlo. We distinguish PTMC from adjacent paradigms, including classical Monte Carlo, hand-coded agent-based models, single-agent reinforcement learning, and large-language-model-based generative agents. To justify the design, we survey cross-disciplinary foundations -- agent-based computational economics, market microstructure, behavioral finance, deep reinforcement learning, generative/LLM-based agents, news-driven trading, systemic risk, econophysics, and game theory -- connecting each literature to a specific design choice in the policy network, training data, or validation protocol. We formalize the PTMC estimator and its convergence properties, specify a candidate bot architecture and training objective, and propose a four-level validation methodology: stylized-fact matching, microstructure- and agent-level checks, and historical stress-test comparison against a zero-intelligence baseline. The framework is proposed but not implemented: we contribute a formal estimator, a cross-disciplinary design justification, and a validation roadmap, and conclude with open research questions. --- ### Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise Authors: John Sweeney Published: 2026-06-28 Categories: cs.LG, math.NA, math.OC, stat.ML arXiv: https://arxiv.org/abs/2606.29554v1 Shuffle order can be a larger source of fine-tuning noise than a memoryless analysis predicts: fixed-clock optimizer memory makes local equal-multiset contrasts first order in the learning rate rather than second order, and the resulting order channel can be large enough for a single seed to flip a close A/B comparison. We isolate this mechanism and derive a fit-free way to size the noise it produces. For a memoryless optimizer, reordering an equal multiset has no first-order endpoint term; the leading local contrast is the $O(η^2)$ gradient bracket. Fixed-clock optimizers such as AdamW are different. Their moment buffers, preconditioner state, and de-biasing counters advance with the step index rather than with the learning-rate-scaled time $τ=ηk$, so the same gradient can receive a position-dependent endpoint weight. For any fixed finite measurement window, a lifted-state expansion gives an $O(η)$ equal-multiset contrast whenever the first-order replay coefficient is nonzero, while regular and clock-matched controls remain $O(η^2)$; a bare fixed-$β$ momentum buffer is already enough. A bitwise-deterministic replay from one warmed optimizer state isolates the mechanism, giving order-variance slopes 1.83 for AdamW, 2.00 for fixed-$β$ momentum, and 4.00 for SGD; matching the memory clock to $τ$ restores the regular exponent. For AdamW with a frozen preconditioner, the same impulse-weight kernel gives a closed-form asymptotic order-variance floor after the local potentials are measured, with no fitted coefficients. The result is local to the measurement window (independent reshuffling can average the channel across windows), but it yields order-noise error bars, positional attribution weights, and a seed-budget criterion for fine-tuning comparisons. --- ### VISTA-DZ: Visual Semantic Trajectory Adaptation for Personalized Dilemma Zone Prediction Authors: Chuheng Wei, Ziye Qin, Ziran Wang, Guoyuan Wu Published: 2026-06-28 Categories: cs.LG, cs.AI, cs.DB, cs.HC, cs.RO arXiv: https://arxiv.org/abs/2606.29548v1 Driver decision making in the dilemma zone at signalized intersections is safety critical, as vehicles approaching a yellow signal must decide whether to stop or proceed within limited time and distance margins. Accurate prediction of both stop-go decisions and decision timing is important for adaptive signal control, advanced driver assistance systems, and human-centered intelligent transportation applications. However, dilemma zone behavior is strongly driver dependent. Similar approach trajectories may lead to different decisions across drivers because of differences in risk preference, braking habit, and decision threshold. Existing personalized models often rely on handcrafted scalar descriptors, which provide useful but limited summaries of individual behavior. This paper proposes VISTA-DZ, a semantic-profile-conditioned framework for personalized stop-go and decision-time prediction. Historical trajectories are converted into visual representations, interpreted by a vision-language model to generate behavioral profiles, and encoded as semantic embeddings to condition a dual-output prediction network. The final model combines a bidirectional GRU encoder, driver-conditioned multi-head cross-attention, and Feature-wise Linear Modulation for temporal evidence selection and feature adaptation. Experiments on the SDZ dataset and a newly collected FDZ dataset show that VISTA-DZ outperforms trajectory-only and handcrafted personalization baselines, achieving 93.26% in-domain simulation accuracy and 90.22% mean accuracy across 20 held-out simulation drivers. Cross-domain results further show feasible zero-shot simulation-to-real transfer and better real-world generalization when simulation data are combined with limited field data. --- ### AURORA: Asymmetry and Update-Induced Rotation for Robust Hallucination Detection in Large Language Models Authors: Zishuai Zhang, Hainan Zhang, Zhiming Zheng Published: 2026-06-28 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.29545v1 Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks. However, their tendency to generate hallucinations, namely factually incorrect or unfaithful outputs, poses a critical obstacle to their deployment in high-stakes applications. Although recent hallucination detection methods have made encouraging progress, they typically rely on costly output-level consistency checks or static hidden-state probes that capture shallow dataset-specific patterns, leading to substantial degradation under cross-dataset evaluation. In this work, we propose AURORA, a novel hallucination detection framework that shifts the focus from static representations to the weight-gradient dynamics of LLMs. Our key insight is that hallucinated and faithful answers induce qualitatively different gradient update patterns on the model's parameters. Specifically, hallucinated samples trigger asymmetric and structurally misaligned gradients, which can be captured through two complementary features: (1) the skewness of the cosine similarity distribution between weight matrices and their gradient update directions, and (2) the rotation ratio, which quantifies how much the gradient update reorients the singular-vector basis of weight matrices via SVD. AURORA achieves strong hallucination detection performance across four model families and four benchmark datasets. Further analyses demonstrate that our method scales effectively across model sizes and transfers to out-of-domain tasks, including mathematical reasoning and vision-language scenarios. --- ### Proteus: Automated Adversarial Robustness Testing for Audio Deepfake Detectors Authors: Nicolas M. Müller, Aditya Tirumala Bukkapatnam, Zohaib Ahmed Published: 2026-06-28 Categories: cs.SD, cs.AI arXiv: https://arxiv.org/abs/2606.29544v1 We present Proteus, a framework developed at Resemble AI for automated robustness testing of our audio deepfake detection system. Given a detector, Proteus systematically searches over sequences of everyday audio transformations (codec transcoding, additive noise, reverberation, dynamic-range compression, and VoIP simulation) to find combinations that fool the detector while preserving speech quality. We propose two complementary search strategies: (1) a breadth-first search that exhaustively maps augmentation effectiveness across the parameter space, and (2) a Q-learning agent designed to efficiently discover deeper attack chains by exploiting structural patterns in the BFS data. We report findings from continuous deployment of Proteus against our production detector, showing that specific augmentation chains can reliably flip detection verdicts while preserving speech intelligibility and speaker identity. We discuss how these findings are used to harden the detector through targeted retraining. --- ### Learned Coordination Conventions in Cooperative MARL: Measuring the Translation Gap Between Theory-Informed Roles and Learned Routing Authors: Yoosung Hong Published: 2026-06-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.29541v1 Role-semantic assignments provide priors over how heterogeneous agents may coordinate, but cooperative MARL systems instead settle on conventions through decentralized, non-stationary learning, with no guarantee that the resulting structure matches those priors. We study this translation gap between theory-informed role expectations and learned coordination structure through a diagnostic combining a role-routing matrix, formation sensitivity ($Δ_{\max}$), and gradient/occlusion attribution across three-role MiniGrid and SMACv2 (Terran) environments. We show that label-conditioned attention produces substantially more concentrated and role-specific routing than flat MLP baselines, remains stable under 3v3--9v9 scaling, transfers zero-shot across team sizes, and is invariant to ally-slot padding. A 5-seed re-evaluation shows partial alignment between learned conventions and designer-specified priors while revealing where small-n noise can manufacture apparent strategic divergence. We present these results as an empirical framework for measuring coordination structure in cooperative MARL rather than as a new equilibrium concept or causal explanation. --- ### Em-ergence of the em-dash: a population-level rise in em-dash frequency in medRxiv preprints at the dawn of the large-language-model era Authors: Przemysław Czuma Published: 2026-06-28 Categories: cs.DL, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2606.29540v1 Large language models (LLMs) can leave subtle stylistic traces in assisted text; one of the most cited is the em-dash (Unicode U+2014). Yet no one has measured whether em-dash use has changed in the scientific literature. This study, pre-registered on the Open Science Framework (HFT8C), used the full set of medRxiv full-text XML preprints from the official Text-and-Data-Mining resource. The primary cohort was first, original versions deposited 2020-2025 with an extractable Discussion section of at least 500 characters (N = 69,632). The primary endpoint was the presence of at least one em-dash in the Discussion; the principal measure was the absolute change in its prevalence between the pre-ChatGPT era (before 30 November 2022) and the post-ChatGPT era, estimated with a logistic model with standard errors clustered by first author. The analysis plan (six supporting analyses, six sensitivity analyses, two falsification tests) was frozen before any confirmatory result was computed. Em-dash prevalence in Discussion sections rose from 4.23% before ChatGPT to 11.58% afterward, an absolute increase of 7.35 percentage points (95% CI 6.94-7.77; odds ratio 2.96, 95% CI 2.77-3.17). The rise was not a sharp jump but a gradual, delayed acceleration: near 4% through 2023, 8.0% in 2024, and 20.3% in 2025. The effect survived every feasible sensitivity analysis (7.35-7.60 pp) and both falsification tests; a placebo split within the pre-LLM era showed no meaningful change (+0.13 pp, 95% CI -0.33 to +0.58), and was essentially absent in boilerplate sections. Independent LLM-associated lexical markers and within-paper section comparisons pointed the same way. The em-dash is a population-level indicator, not a per-paper detector of LLM use, and the design cannot establish causality; it shows that something in how scientific literature is written changed markedly in the early 2020s, and roughly when. --- ### RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources Authors: Yijia Fan, Zonglin Di, Zimo Wen, Yifan Yang, Mingxi Cheng, Qi Dai, Bei Liu, Kai Qiu, Yue Dong, Ji Li, Chong Luo Published: 2026-06-28 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2606.29538v4 Skills are a useful abstraction for software agents, turning human and agent experience into reusable procedural knowledge. Yet existing skill libraries are mostly hand-written, text-centric, or derived from agent traces, leaving tutorial videos and other multimodal human resources largely underused. We present RESOURCE2SKILL, a framework that distills multimodal resources, including tutorial videos, repositories, articles, and reference artifacts, into executable skills for software agents. RESOURCE2SKILL organizes these skills as a hierarchical multimodal Skill Wiki, where each entry combines structured text, code, visual examples, metadata, and provenance. This design preserves complementary signals from different resources: videos capture temporal operations and visual effects, code captures executable tool patterns, and articles or artifacts provide conceptual and stylistic grounding. At inference time, agents retrieve and compose relevant skills from the wiki; when coverage is insufficient, the same construction operator can acquire new skills online. Across seven practical authoring domains, RESOURCE2SKILL improves average overall score by +11.9 percentage points over no-skill agents and outperforms strong harness baselines in 26 of 28 main-aggregate model-domain cells. Ablations confirm the value of multimodal skill format, hierarchical organization, source diversity, selection strategy, and online acquisition. --- ### OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks Authors: Mengqi Yuan, Zilong Zhou, Xinzhuang Xiong, Weiming Wu, Jiayang Sun, Jiamin Song, Kaiqian Cui, Bowen Wang, Haoyuan Wu, Yitong Li, Dunjie Lu, Haikong Lu et al. Published: 2026-06-28 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.29537v2 Existing computer-use benchmarks fail to capture the realism, complexity, and long-horizon demands of real-world computer use, limiting their ability to reveal the limitations of frontier agents. We introduce OSWorld 2.0, a benchmark of 108 long-horizon computer-use workflows across everyday and professional tasks, designed to capture complex and challenging real-world phenomena. Each task represents a realistic end-to-end workflow that takes human users a median of about 1.6 hours to complete and requires an average of 318 tool calls with Claude Opus 4.7 using maximum thinking, compared with about 30 in OSWorld 1.0. OSWorld 2.0 targets challenge phenomena that are common in real workflows yet underrepresented in prior benchmarks, spanning interaction-design challenges such as streaming interaction and dynamic environments, as well as agent-pattern challenges such as cross-source reasoning, implicit-state inference, and visual-spatial precision. Tasks are grounded in authentic input artifacts and cross-referenced against realistic stateful user profile data, and include separate safety reports auditing safety-sensitive execution. Under our primary binary-completion metric at 500 steps, Claude Opus 4.8 with maximum thinking and batched tool calls scores best but still completes only 20.6% of tasks at a 54.8% partial score; GPT-5.5 is far more token-efficient yet plateaus near 13%. These results show that current agents are still far from professional-level computer use: rather than stumbling on basic GUI control or coding, they lose track of constraints, miss information that arrives mid-task, guess rather than ask the user, and skip verification, struggling most when a task hinges on hidden state they must recover. --- ### GarmentZoom: Generating Zoomable Images from Garment Listings Authors: Renjie Zhao, Jingwei Ma, Huy Huynh Cao, Brian Curless, Steven M. Seitz, Ira Kemelmacher-Shlizerman Published: 2026-06-28 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.29535v2 Online product listings for garments often include an overview photo and a close-up to show garment details. However, each photo focuses on either field of view or garment detail, forcing users to alternate between views and breaking browsing continuity. We present GarmentZoom, a system that enhances the full-view photo to match the fidelity of its accompanying close-up, enabling seamless zoom-and-pan exploration. Unlike standard reference-based super-resolution, our setting involves close-up references that are spatially unaligned with the full view, and scale factors that vary substantially across garments 3-20$\times$. Prior work typically relies on alignment to transfer details or requires per-instance fine-tuning to memorize them. Instead, we train a single model that supports a continuous range of scales across diverse garments. Our approach synthesizes details without requiring spatial alignment and matches the quality of per-instance methods with a fraction of the training cost. --- ### Preference-ASR: A Preference-Aware Test Set for Benchmarking ASR in the Era of Speech LLMs Authors: Nithin Rao Koluguri, Sasha Meister, Nikolay Karpov, Piotr Zelasko, Desh Raj, Jagadeesh Balam, Boris Ginsburg Published: 2026-06-28 Categories: cs.CL, eess.AS arXiv: https://arxiv.org/abs/2606.29534v1 Popular ASR test sets adopt inconsistent conventions for numbers, disfluencies, entities, and casing, while standard normalizers erase the format distinctions users care about. Current benchmarks therefore cannot measure whether a model follows user preferences for output style. We introduce PreferenceASR, a test set evaluating ASR systems on their ability to follow natural-language preference instructions across four categories: normalization, entities, disfluencies, and case. Built from seven open-source corpora via a two-stage LLM-assisted pipeline with human verification, it is evaluated with a preference-aware normalizer that selectively skips steps matching the active instruction. Benchmarking four models shows rankings shift across preference types, exposing quality differences traditional evaluation obscures. We publicly release the dataset. --- ### Improved Multi-Dimensional Forecasting for Swap Regret Authors: Joey Rivkin, Ramiro N. Deo-Campo Vuong, Robert Kleinberg, Chido Onyeze, Erald Sinanaj, Eva Tardos Published: 2026-06-28 Categories: cs.GT, cs.LG arXiv: https://arxiv.org/abs/2606.29533v1 We study the problem of forecasting for an arbitrary number of downstream agents with unknown objectives, each of whom best responds to the forecaster's predictions. We seek a single forecaster that guarantees sublinear swap regret for all downstream agents simultaneously. For two-dimensional outcome spaces, we give a polynomial time algorithm that guarantees $\tilde{O}(\sqrt{kT})$ swap regret for any downstream agent with $k$ actions. This improves over the previously known bound of $\tilde{O}(kT^{5/8})$ and avoids the exponential in $T$ runtime of prior algorithms in this setting. Our algorithm extends nicely to other low dimensional environments, retaining $\tilde{O}(\sqrt{T})$ downstream swap regret while the exponent of $k$ in the regret bound and the exponent of $T$ in the running time both grow with dimension. For arbitrary dimension $d$, we give a forecasting algorithm that guarantees $\tilde{O}(d\sqrt{kT})$ swap regret, assuming the forecaster knows an upper bound $k$ on the number of actions available to any downstream agent, albeit with a much longer runtime. This improves upon previous high dimensional guarantees that had $\tilde{O}(T^{2/3})$ dependence and required additional behavioral assumptions. --- ### SemJoin: Semantic Join Optimization Authors: Christopher Gou, Aditya Banerjee, Jiaxuan Wang, Chunwei Liu Published: 2026-06-28 Categories: cs.DB, cs.AI arXiv: https://arxiv.org/abs/2606.29532v1 Integrating unstructured data into relational database systems is increasingly important as demand grows for natural language querying and analysis. A semantic join, joining two tables under a natural-language predicate, can be evaluated with a large language model (LLM), but comparing every pair of tuples requires O(M x N) LLM invocations and is cost-prohibitive at scale. Existing systems reduce this cost but typically commit to a single fixed strategy (e.g., embedding similarity or one batched scheme) regardless of the data or the join predicate. We propose an LLM-agent-based decision pipeline that optimizes semantic joins by matching the execution strategy to the characteristics of the underlying tables. An LLM advisor routes each join to one of two strategies: a Cluster Join, which prunes candidates via unsupervised embedding clustering and sample-based filtering, or a Classifier strategy for predicates that reduce to a shared discrete label set. Across three diverse datasets (IMDb reviews, email contradictions, and Stack Overflow tags), the advisor consistently identifies the optimal execution strategy for each workload. This dynamic routing proves decisive: it outperforms adaptive block join (ABJ) by 20-33 F1 points across all datasets while consuming fewer tokens on two of the three, and achieves higher F1 scores than featurized-decomposition join (FDJ) at one to two orders of magnitude lower token cost. --- ### MotionAtlas: Detailed Region Captioning for Motion-Centric Videos Authors: Weisong Liu, Haochen Wang, Kuan Gao, Yuhao Wang, Yikang Zhou, Zhongwei Ren, Jacky Mai, Anna Wang, Yanwei Li, Jason Li, Zhaoxiang Zhang Published: 2026-06-28 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.29531v1 We propose MotionAtlas, a system for detailed captioning of motion-centric videos, comprising (1) a dedicated human-annotated benchmark, (2) a scalable, high-quality pipeline to construct training samples, and (3) a family of powerful Video-MLLMs. Unlike conventional global motion captioning datasets, we focus on region-aware motion captioning: given a video and a spatiotemporal mask, the model generates precise descriptions of motion within the target region, thereby alleviating visual clutter and motion entanglement and enabling reliable, quantifiable evaluation. Concretely, we first build MotionAtlas-Bench, a comprehensive benchmark comprising 2,073 multiple-choice questions, meticulously annotated for a curated set of high-quality, motion-centric videos, to evaluate fine-grained motion understanding of the objects in question. Second, we design a rigorous and scalable data pipeline that leverages self-bootstrap refinement to suppress fine-grained hallucinations, yielding 159k high-quality motion captioning data. Third, we design a tailored training data composition strategy, which achieves consistent and substantial performance gains across diverse baseline Video-MLLMs, including Molmo2 and Qwen3-VL. For instance, MotionAtlas-4B surpasses Qwen3-VL-4B by an average of 5.2 percentage points across general motion benchmarks. The benchmark, dataset, and code have been released. --- ### Supervised Hebbian learning in Deep Counterstream Associative Networks Authors: Andreas Knoblauch Published: 2026-06-28 Categories: cs.NE arXiv: https://arxiv.org/abs/2606.29528v1 Modern machine learning applications employ deep neural networks training with the error backpropagation algorithm. Although this algorithm is very effective, it lacks biological realism. For example, backpropagation requires symmetric connectivity, and a separate neural processing channel for error signals. Prior works have therefore proposed a number of more realistic alternatives for error backpropagation. However, most of them still suffer from demanding preassumptions that may be not fulfilled in the real brain, for example, they often still require either symmetric connectivity or two separate processing channels, and often require also special mathematical operations like subtractions or function inversions. Here I propose supervised counterstream learning in deep associative networks as a simpler approach that requires only recognition of errors during training, and then backpropagates correcting target activity through the same activity channel as used for forward propagation. For this, two activity waves are initiated at the same time in input and output layers and then traveling in opposite directions to meet in one of the hidden layers. By employing simple local Hebbian-type learning rules, the corresponding activity pattern sequences get linked bidirectionally, thereby decreasing error rates over time. Despite its simplicity and an incomplete hyperparameter optimzation, a high high test accuracy is achieved on the (binarized) MNIST data set that is comparable to more demanding architectures. --- ### The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning Authors: Jing Liang, Hongyao Tang, Yi Ma, Yancheng He, Weixun Wang, Xiaoyang Li, Ju Huang, Wenbo Su, Jinyi Liu, Yan Zheng, Jianye Hao, Bo Zheng Published: 2026-06-28 Categories: cs.LG arXiv: https://arxiv.org/abs/2606.29526v1 Reinforcement learning (RL) has gained growing attention in large language model (LLM) post-training, yet RL training remains fragile and can suffer from instability or collapse. One vital cause is training-inference mismatch: LLM adopts separate inference and training engines for generation efficiency and training precision, which in practice exhibits inconsistent probabilities for the same trajectories on training and inference sides, even with synchronized model parameters. This naturally induces a special type of off-policyness ever existing and poisoning the training. Prior works have made various efforts in addressing the off-policyness to stabilize the training policies under the mismatch. In this paper, we point out the objective misalignment neglected by existing works that an effective update to the policy in the training engine not necessarily ensures the improvement of the inference policy, i.e., the one used in deployment. To this end, we propose a new policy optimization objective for LLM RL, named Monotonic Inference Policy Improvement (MIPI). Following this principle, we introduce Monotonic Inference Policy Update (MIPU), a two-step LLM RL framework that constructs sampler-referenced candidate updates and selectively accepts synchronized candidates using an inference-side gap proxy. Experiments conducted on two model scales under high mismatch show that MIPU improves average reasoning performance and training stability. --- ### When Does Activation Steering Change What a Model Computes From? Authors: Benjamin Shih, John Winnicki, Eric Darve Published: 2026-06-28 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2606.29522v2 Activation steering can reliably change an agent's output by modifying its internal activations. Yet arriving at the same answer need not involve the same computation: behavioral equivalence does not imply mechanistic equivalence. We test whether an activation edit changes the state used in subsequent computation or biases that computation toward the desired output. In a controlled state-tracking task, trace supervision creates an editable register: changing its internal value makes the model apply the next operation to the edited state. We next ask whether widely used mean activation steering admits a similar interpretation: does steering recreate the internal configuration the model naturally uses when performing a task? If so, transplanting that activation should be effective, and steering should remain effective at the scale of the natural source-to-target activation change. Neither prediction holds in the two selected model-task settings. Replacing the activation at one layer produces less than 2% of the target-answer margin gain from patching through all remaining layers, while natural-scale steering is similarly ineffective. The steering vectors have norms 77 and 26 times the median natural change in Qwen and Llama, producing 54% and 82% of the reference effect. Thus, a successful steering intervention need not reproduce the natural target activation at the intervention layer. More generally, an intervention should be interpreted as changing the computational state only when a later computation uses the edited value according to the semantics of that state. --- ### Do Models Read What They Write? Causal Registers in Scratchpad Reasoning Authors: Benjamin Shih, John Winnicki, Eric Darve Published: 2026-06-28 Categories: cs.LG, cs.CL arXiv: https://arxiv.org/abs/2606.29522v1 A central hope behind process supervision is that models can expose intermediate variables that matter for their later behavior. For this to help with alignment, a scratchpad must be tied to the computation: when the model writes a state, later steps should compute from that state. To test this requirement, we use a controlled state-tracking task with a known update rule, comparing models trained to report only the final state with models trained to write intermediate states before giving the final answer. At evaluation, we edit the internal representation of one written state while leaving the visible scratchpad text fixed. Because the transition rule is known, the edit has a single correct downstream consequence. In Qwen2.5-Coder-7B, the state-writing model predicts the next phase bit implied by the edited state on 80% and 91% of held-out examples across the two task variants, while pretrained and final-answer-only controls remain near baseline. Additional controls rule out generic next-token steering and copying another continuation: the prediction depends on both the edited state and the current move. The same causal-use pattern replicates across model families. Together, these results suggest a sharper goal for scratchpad oversight: not just to make intermediate reasoning legible, but to train written states that the model uses as part of its computation. --- ### Not All Objectives Are Born Equal: Priority-Constrained Descent for Hierarchical Multi-Objective Optimization Authors: Dara Varam, Mohamed I. Alhajri Published: 2026-06-28 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2606.29521v2 Deep learning problems rarely involve objectives that are equal in importance. A primary objective defines the goal, whilst secondary objectives, such as sparsity, compression, or robustness constrain the solution. While existing multi-objective methods have proven effective in practice, they have a clear symmetry problem and neglect the inherent objective hierarchy built into these objective spaces. We introduce Priority-Constrained Descent (PCD), a gradient-based optimization framework designed to explicitly exploit hierarchical objective structures. PCD preserves the direction of primary descent whilst allowing for the minimal distortion necessary to guarantee progress on secondary objectives, controlled by a single $τ\in (0, 1)$ that dictates the strength of the distortion. The resulting formulation is invariant to objective scaling and admits exact closed-form solutions for problems with two and three objectives. We evaluate PCD within structured network compression settings, unstructured sparsity and low-rankness, and across a variety of synthetic experiments, showing Pareto dominance and better per-objective performance with secondary progress guarantees over existing methods, further exhibiting the interpretable trade-off that $τ$ provides. --- ### Not All Objectives Are Born Equal: Priority-Constrained Descent for Hierarchical Multi-Objective Optimization Authors: Dara Varam, Mohamed I. Alhajri Published: 2026-06-28 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2606.29521v1 Deep learning problems rarely involve objectives that are equal in importance. A primary objective defines the goal, whilst secondary objectives, such as sparsity, compression, or robustness constrain the solution. While existing multi-objective methods have proven effective in practice, they have a clear symmetry problem and neglect the inherent objective hierarchy built into these objective spaces. We introduce Priority-Constrained Descent (PCD), a gradient-based optimization framework designed to explicitly exploit hierarchical objective structures. PCD preserves the direction of primary descent whilst allowing for the minimal distortion necessary to guarantee progress on secondary objectives, controlled by a single $τ\in (0, 1)$ that dictates the strength of the distortion. The resulting formulation is invariant to objective scaling and admits exact closed-form solutions for problems with two and three objectives. We evaluate PCD within structured network compression settings, unstructured sparsity and low-rankness, and across a variety of synthetic experiments, showing Pareto dominance and better per-objective performance with secondary progress guarantees over existing methods, further exhibiting the interpretable trade-off that $τ$ provides. --- ### SAKE: Software Architectural Knowledge Evaluation Benchmark for Large Language Models Authors: Tiziano Santilli, Francesco Daghero, Mayhar Tourchi Moghaddam Published: 2026-06-28 Categories: cs.SE, cs.AI, cs.DB arXiv: https://arxiv.org/abs/2606.29520v1 Large Language Models (LLMs) are increasingly used as assistants across the software development lifecycle, yet their ability to reason about software architecture remains largely unmeasured. Architectural decision-making depends on quality attribute trade-offs, design patterns, and system-level constraints, none of which are exercised by benchmarks that target syntactic or algorithmic tasks. We introduce SAKE (Software Architectural Knowledge Evaluation), a standardized and reproducible benchmark for assessing software architectural knowledge in LLMs. SAKE comprises 2154 expert-curated multiple-choice questions, each with four options, stratified across eight architectural categories and four context-length levels. We evaluate 11 proprietary and open-weight models in zero-shot and five-shot settings. Overall accuracy is high, but performance varies markedly across categories, revealing competency gaps in areas central to professional practice. SAKE, its evaluation scripts, and all results are released as open source to give the community a baseline for tracking architectural reasoning in LLMs. --- ### Anti-Collapse Dynamics and the Emergence of Multi-Time-Scale Learning in Recurrent Neural Networks Authors: Lorenzo Livi Published: 2026-06-28 Categories: cs.LG, physics.data-an arXiv: https://arxiv.org/abs/2606.29519v3 Long-range learning is hard for recurrent networks trained with stochastic gradient descent, because the influence of a past input fades with the lag $\ell$, and if it fades too fast the dependence cannot be learned from finite data. This fade is captured by an envelope $f(\ell)$. An exponential fade makes the data needed to learn a lag-$\ell$ dependence grow exponentially, putting long horizons out of reach; a power-law fade keeps the cost polynomial. We show that the asymptotic decay behavior of $f(\ell)$ is not fixed by the architecture. Instead, it emerges from the coupling between the state dynamics and parameter dynamics, settling into either a collapsed regime (fast, exponential forgetting) or an extended, anti-collapsed regime (slow, power-law forgetting). The intuition is a competition within these coupled dynamics. Training drives the network's effective time scales toward short ones, while rare, heavy-tailed fluctuations of the learning dynamics push a few of them to very long values. Along the route studied here, the extended regime survives only when these heavy-tailed pushes are strong enough to balance the pull. We make this mathematically precise with a coarse-grained stochastic process and derive an explicit threshold at which this route to the extended regime becomes available. A single exponent, the spectral exponent~$β$, then governs both the spread of time scales and how slowly the network forgets. Realizing the regime in practice needs one more ingredient: the joint action of the architecture and the optimizer must be able to hold such a broad spread. A network whose capacity to generate broad time-scale spectra is severely constrained still collapses, even when supplied with strong heavy-tailed forcing. Heavy-tailed fluctuations thus act not as noise to be suppressed, but as the mechanism that sustains long-range learning. --- ### Anti-Collapse Dynamics and the Emergence of Multi-Time-Scale Learning in Recurrent Neural Networks Authors: Lorenzo Livi Published: 2026-06-28 Categories: cs.LG, physics.data-an arXiv: https://arxiv.org/abs/2606.29519v2 Long-range learning is hard for recurrent networks trained with stochastic gradient descent, because the influence of a past input fades with the lag $\ell$, and if it fades too fast the dependence cannot be learned from finite data. This fade is captured by an envelope $f(\ell)$. An exponential fade makes the data needed to learn a lag-$\ell$ dependence grow exponentially, putting long horizons out of reach; a power-law fade keeps the cost polynomial. We show that the asymptotic decay class of $f(\ell)$ is not fixed by the architecture. Instead, it emerges from the coupling between the state dynamics and parameter dynamics, settling into either a collapsed regime (fast, exponential forgetting) or an extended, anti-collapsed regime (slow, power-law forgetting). The intuition is a competition within these coupled dynamics. Training drives the network's effective time scales toward short ones, while rare, heavy-tailed fluctuations of the learning dynamics push a few of them to very long values. The extended regime survives only when these heavy-tailed pushes are strong enough to balance the pull. We make this mathematically precise with a coarse-grained stochastic process and prove exactly when the extended regime exists. A single exponent, the spectral exponent~$β$, then governs both the spread of time scales and how slowly the network forgets. Realizing the regime in practice needs one more ingredient: the joint action of the architecture and the optimizer must be able to hold such a broad spread. A network whose capacity to generate broad time-scale spectra is severely constrained still collapses, even when supplied with strong heavy-tailed forcing. Heavy-tailed fluctuations thus act not as noise to be suppressed, but as the mechanism that sustains long-range learning. --- ### Harvesting AI Computation at the Edge via Generic Approximation Authors: Yihan Wang, Huiru Yan, Luxin Zhang, Long Cheng, Weiwei Chen, Ying Wang, Lei Zhang, Cheng Liu, Huawei Li Published: 2026-06-28 Categories: cs.AR, cs.LG arXiv: https://arxiv.org/abs/2606.29518v1 With the widespread adoption of AI in various IoT scenarios such as smart sensing and processing, AI chips have become a common component at the edge. These chips are typically specialized for structured neural network (NN) processing and are designed to meet peak workload demands. However, they are often underutilized and suffer from considerable computational waste due to temporal or spatial redundancy in processing. Conversely, general-purpose processing engines at the edge may struggle with compute-intensive tasks such as signal processing and complex numerical operations because of stringent resource constraints. To address this imbalance, we propose a framework that harvests unused AI computation resources using general-purpose approximation techniques. The core idea is to automatically convert traditional computing tasks into neural network models via a representative neural architecture search (NAS) method. These approximate versions of general-purpose tasks are then deployed on AI engines during their idle periods. Specifically, we introduce a runtime scheduler that offloads these tasks to AI chips without compromising the performance of primary AI workloads, thereby alleviating the burden on general-purpose processors. Experiments on a representative AIoT processor show that our proposed AI computation harvesting strategy delivers substantial performance improvements across a set of edge processing tasks. --- ### A Mathematical Optimization Approach for Expert-Informed Bayesian Best Subset Selection Authors: Nolan Alexander, Henning Mortveit Published: 2026-06-28 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2606.29516v1 A central challenge in statistical modeling is identifying the subset of features that belong in the true regression model. The classical best subset selection problem, recently made tractable via mixed-integer optimization (MIO), finds the globally optimal sparse solution. It does not, however, make use of any information beyond the observed data. In many applied settings, domain experts can meaningfully rank or score the relevance of candidate predictors, yet no existing framework integrates such probabilistic expert assessments directly into the best-subsets objective. This paper presents Expert-Implied Bayesian Best Subsets (EBBS), a method that incorporates domain-expert probability estimates of feature relevance into the MIO best-subsets problem through a maximum a posteriori (MAP) framework. Expert views from multiple respondents are aggregated into a single prior probability per feature using the Poisson binomial distribution for marginal probability estimates, the pairwise win rate for pairwise comparisons, or the normalized mean rank for ordinal rankings. This probability enters the objective function as a log-odds penalty term that smoothly encourages or discourages the selection of each feature consistent with the expert consensus. This paper provides analytic derivations of the MAP formulation and characterizes its theoretical properties. The proposed model reduces to Best Subsets when experts all have no views. Empirical results on synthetic and real datasets are forthcoming. ---