# Research Papers — 2026-05 · bundle page 151 of 326 > Full-text markdown of papers #7501–7550 of 16266 published in 2026-05, newest first. AI-readable mirrors of the full 16266-paper 2026-05 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-05/page/151 [← Previous bundle](https://aipod.dev/papers/archive/2026-05/page/150.md) [Next bundle →](https://aipod.dev/papers/archive/2026-05/page/152.md) ### Guard: Scalable Straggler Detection and Node Health Management for Large-Scale Training Authors: Guanliang Liu, Abhinandan Patni, Congzhu Lin, Zoe Zeng, Jack Wittmayer, Josh Wu, Ashvin Nihalani, Binxuan Huang, Yinghong Liu, Rory Na, Anthony Ko, Alexander Zhipa et al. Published: 2026-05-18 Categories: cs.DC, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2605.17879v1 Training frontier-scale foundation models involves coordinating tens of thousands of GPUs over multi-month runs, where even minor performance degradations can accumulate into substantial efficiency losses. Existing health-check mechanisms, such as NCCL tests or GPU burn-in, primarily focus on functional correctness and often fail to detect fail-slow behaviors that silently degrade system performance. In this paper, we present Guard, a scalable system for detecting stragglers and ensuring node health in large-scale training clusters. Guard combines lightweight online performance monitoring during training with an offline node-sweep mechanism that systematically evaluates and qualifies nodes before they participate in production workloads. This design enables Guard to detect both acute failures and long-running fail-slow behaviors that traditional diagnostics cannot capture. Deployed on large-scale foundation model pretraining workloads, Guard improves mean FLOPs utilization by up to 1.7x, reduces run-to-run training step variance from 20% to 1%, increases mean time to failure (MTTF), and significantly reduces operational and debugging overhead. These results demonstrate that proactive straggler detection and systematic node qualification are critical for maintaining stable and efficient large-scale training. --- ### PAIR: Prefix-Aware Internal Reward Model for Multi-Turn Agent Optimization Authors: Wonjoong Kim, Yeonjun In, Sangwu Park, Dongha Lee, Chanyoung Park Published: 2026-05-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.17877v2 A significant hurdle for current LLMs is the execution of complex, multi-stage tasks. Group Relative Policy Optimization (GRPO) has been emerging as a leading choice, but its reliance on sparse outcome rewards severely limits credit assignment across intermediate steps. Existing remedies such as running full rollouts to assign step-level advantages, calling external LLM judges at each step, or computing intrinsic rewards that require ground-truth answers at every evaluation introduce significant costs or practical constraints. We hypothesize that internal correctness probing over LLM hidden states can be repurposed as a step-level reward signal, potentially addressing all of these limitations at once. However, existing probing research assumes clean inputs, and we first show that this assumption breaks down in multi-step settings: hidden-state probes degrade severely under prefix contamination tracking coherence with the (possibly corrupted) prefix rather than grounded correctness, while attention-based features remain robust to contamination but underperform on clean prefixes. Building on this complementary relationship, we propose the Prefix-Aware Internal Reward (PAIR), a two-stage model with a frozen hidden-state probe estimating belief-consistency and a lightweight attention-based head correcting it toward grounded correctness. Experimental results show that PAIR achieves the highest AUROC on contaminated trajectories while operating at negligible inference cost, enabling dense step-level reward signals for GRPO training without external model calls, ground-truth dependencies, or full-trajectory rollouts. --- ### HexagonalWarriorMamba: Superior Threshold-Dependent Multi-label Classification of 12-Lead ECG Cardiac Abnormalities Authors: Huawei Jiang, Husna Mutahira, Shibo Wei, Jiahang Li, Vladimir Shin, Juneho Yi, Dongryeol Ryu, Wonyoung Park, Mannan Saeed Muhammad Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17875v1 The accurate automated diagnosis of cardiac abnormalities from 12-lead electrocardiograms (ECGs) is critical for managing cardiovascular disease. However, detecting concurrent conditions remains a challenge for traditional deep learning models, which often have limited ability to model the long-range dependencies inherent in ECG signals. This manuscript proposes HexagonalWarriorMamba (HWMamba), a framework built on the Mamba architecture that processes 12-lead ECGs as single-channel 2D images rather than conventional 1D time series. By integrating a hierarchical architecture with a 2D Selective Scan mechanism, HWMamba is designed to model global context and complex spatial relationships within the data. The model is evaluated on the PhysioNet/Computing in Cardiology Challenge 2021 dataset, which includes 26 diagnostic labels and comprises recordings collected from seven institutions across four countries and three continents. Results demonstrate that HWMamba outperforms current state-of-the-art (SOTA) methods across five key threshold-dependent metrics, including Challenge Score and Subset Accuracy. These improvements provide a balance between strong discriminative capability and effective threshold selection derived from the training data, while maintaining near-SOTA performance in Macro AUROC. This Hexagonal Warrior performance, reflecting consistent performance across multiple evaluation dimensions, positions HWMamba as a robust and versatile approach for multi-label ECG classification. --- ### HINT-SD: Targeted Hindsight Self-Distillation for Long-Horizon Agents Authors: Woongyeong Yeo, Yumin Choi, Taekyung Ki, Sung Ju Hwang Published: 2026-05-18 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2605.17873v2 Training long-horizon LLM agents with reinforcement learning is challenging because sparse outcome rewards reveal whether a task succeeds, but not which intermediate actions caused the outcome or how they should be corrected. Recent methods alleviate this issue by generating rewards or textual hints from turn-level action-output signals, or by using feedback-conditioned self-distillation. However, generating feedback at every turn is inefficient when many intermediate turns are already successful or neutral, and applying feedback at a fixed or misaligned turn often fails to supervise the actions that contributed to the failure. To bridge this gap, we propose HINT-SD, a targeted self-distillation framework that uses full-trajectory hindsight to select failure-relevant actions and applies feedback-conditioned distillation only to targeted action spans. Experiments on BFCL v3 and AppWorld show that our method outperforms the dense per-turn feedback baseline by up to 13.60 percentage points on average while achieving a 2.26$\times$ reduction in time per training step, suggesting that selecting where to distill is key to effective and efficient long-horizon agent training. --- ### HINT-SD: Targeted Hindsight Self-Distillation for Long-Horizon Agents Authors: Woongyeng Yeo, Yumin Choi, Taekyung Ki, Sung Ju Hwang Published: 2026-05-18 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2605.17873v1 Training long-horizon LLM agents with reinforcement learning is challenging because sparse outcome rewards reveal whether a task succeeds, but not which intermediate actions caused the outcome or how they should be corrected. Recent methods alleviate this issue by generating rewards or textual hints from turn-level action-output signals, or by using feedback-conditioned self-distillation. However, generating feedback at every turn is inefficient when many intermediate turns are already successful or neutral, and applying feedback at a fixed or misaligned turn often fails to supervise the actions that contributed to the failure. To bridge this gap, we propose HINT-SD, a targeted self-distillation framework that uses full-trajectory hindsight to select failure-relevant actions and applies feedback-conditioned distillation only on targeted action spans. Experiments on BFCL v3 and AppWorld show that our method improves over the dense per-turn feedback baseline by up to 18.80 percent while achieving 2.26$\times$ lower time per training step, suggesting that selecting where to distill is a key factor for both effective and efficient long-horizon agent training. --- ### PySIFT: GPU-Resident Deterministic SIFT for Deep Learning Vision Pipelines Authors: Sivakumar K. S., Mohammad Daniyalur Rahman, Gopi Raju Matta Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17869v1 A widespread assumption in local feature research holds that classical handcrafted descriptors are accuracy-limited relics best replaced by learned alternatives. We show this is wrong. Through an 8-configuration ablation spanning four benchmarks (HPatches, ROxford5K, IMC Phototourism, MegaDepth), we demonstrate that classical SIFT with DSP multi-scale pooling outperforms neural descriptor and orientation replacements (HardNet, OriNet) on every accuracy metric--while running 2--18$\times$ faster--and that learned matchers (LightGlue) complement rather than supersede classical features. The conclusion reframes a decade of work: not "replace SIFT" but "compose with SIFT," classical extraction paired with learned matching only where geometric context demands it. This finding was invisible because no prior GPU SIFT kept the complete pipeline in VRAM or offered modularity for controlled classical-vs-learned ablations. We present PySIFT, the first fully GPU-resident SIFT, implemented in CuPy/Numba CUDA kernels with DLPack zero-copy handoff to downstream DL frameworks--submillisecond O(1) metadata swap regardless of keypoint count. On a laptop-grade NVIDIA RTX 3050 (4 GB VRAM), PySIFT achieves: (i) higher Mean Matching Accuracy (MMA) than OpenCV SIFT on HPatches, (ii) 383 ms faster per pair on high-resolution MegaDepth, (iii) higher geometric accuracy on cross-dataset benchmarks (+5.6 pp AUC@10${}^\circ$ on MegaDepth, more inliers on IMC Phototourism), and (iv) bitwise deterministic output--identical keypoints and descriptors across runs, with detection reproducing identically even across GPU architectures: a guarantee that learned extractors cannot match without significant performance sacrifice, and cannot achieve at all across GPU architectures due to cuDNN's architecture-dependent algorithm selection. PySIFT is open-source, requiring no C++ compilation. --- ### DAD4TS: Data-Augmentation-Oriented Diffusion Model for Time-Series Forecasting with Small-Scale Data Authors: Masahiro Suzuki, Bohui Xia, Hiroto Yamamoto, Masanori Miyahara Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17866v2 Small-scale data is a critical problem in time-series forecasting tasks. Data augmentation is an effective strategy for this task, but it has a limitation in generating meaningful data. To address this limitation, we propose DAD4TS, a diffusion-model-based data augmentation method with reinforcement learning, designed for time-series forecasting with small-scale data. In DAD4TS, a data generator is simultaneously trained with a time-series model and controlled by a reinforcement learning model to efficiently generate samples that improve the forecast accuracy of the time-series model. To support small-scale data, we use mathematical methods instead of conventional VAE methods to train the diffusion model by projecting the time-series data into the geometric space. We validated the effectiveness of DAD4TS with seven comparative methods through qualitative and quantitative experiments on six real-world datasets and eight time-series models. As a result, DAD4TS was validated on five datasets. --- ### Imaging Hidden Objects with Consumer LiDAR via Motion Induced Sampling Authors: Siddharth Somasundaram, Aaron Young, Akshat Dave, Adithya Pediredla, Ramesh Raskar Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17865v1 LiDARs are being increasingly deployed for consumer imaging in handheld, wearable, and robotic applications. These sensors can capture the time-of-flight of light at picosecond resolution, which in principle, enables them to capture information about objects hidden from their field of view. While such non-line-of-sight (NLOS) imaging capabilities have been shown on research-grade LiDARs, they are challenging to achieve on consumer devices due to poor signal quality resulting from low laser power, low spatial resolution, and object and camera motion. Inspired by burst photography and synthetic aperture radar, we propose a multi-frame fusion strategy to overcome these challenges and demonstrate NLOS imaging on consumer LiDAR. We first introduce the motion-induced aperture sampling model to unify the effects of object shape, object motion, and camera motion under a single measurement model. Using this model, we demonstrate several NLOS capabilities on a smartphone-grade LiDAR: (1) 3D reconstruction, (2) single and multi-object tracking, and (3) camera localization using hidden objects. Previously, NLOS imaging capabilities were largely restricted to bulky and expensive research-grade hardware that requires extensive setup and calibration. Our results represent a shift towards plug-and-play NLOS imaging, where anyone can image hidden objects with off-the-shelf hardware ($<100) and no additional setup. We believe that democratization of such capabilities will advance consumer applications of NLOS imaging. --- ### $\boldsymbol{f}$-OPD: Stabilizing Long-Horizon On-Policy Distillation with Freshness-Aware Control Authors: Xianwei Chen, Shimin Zhang, Jibin Wu Published: 2026-05-18 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.17862v1 Scaling on-policy distillation (OPD) for large language models (LLMs) confronts a fundamental tension: asynchronous execution is necessary for system efficiency, but structurally deviates from the ideal on-policy objective. To address this challenge, we theoretically decompose the objective discrepancy into rollout drift and supervision drift, capturing staleness in student rollout and teacher context, respectively. Building on this, we introduce a sample-level freshness score that quantifies the reliability of a buffered sample with respect to the on-policy objective. Guided by this signal, we further propose f-OPD, a novel framework that adaptively regulates stale-sample influence and constrains policy drift accumulated under asynchronous training. Across reasoning, tool-use, and coding-agent tasks of increasing interaction horizon, f-OPD consistently achieves task performance comparable to synchronous optimization while largely retaining the throughput advantages of asynchronous execution. Our results establish the first recipe for achieving a performance-efficiency trade-off in OPD, paving the way for long-horizon agentic post-training at scale. --- ### PAREDA: A Multi-Accent Speech Dataset of Natural Language Processing Research Discussions Authors: Sicheng Jin, Dipankar Srirag, Aditya Joshi Published: 2026-05-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.17860v1 While modern Automatic Speech Recognition (ASR) systems achieve high accuracy on benchmark corpora, their performance often degrades when there is real-world variability. This work focuses on variability arising due to accented, spontaneous, and domain-specific speech. In particular, we introduce PAper REading DAtaset (PAREDA), a first-of-its-kind multi-accent speech dataset consisting of discussions on academic Natural Language Processing (NLP) papers between speakers with Australian, Indian-English, and Chinese English accents. Each session elicits a spontaneous monologue (a summary of a paper's abstract) and a non-monologue (a question-and-answer session between participants), resulting in a corpus rich with technical jargon and conversational phenomena. We evaluate the performance of SOTA ASR models on PAREDA, analysing the impact of accent mixing and increased speech rate. Our results show that, in the zero-shot setting, models perform worse, confirming the dataset's challenging nature. However, fine-tuning on PAREDA significantly reduces the Word Error Rate (WER), demonstrating that our dataset captures linguistic characteristics often missing from existing corpora. PAREDA serves as a valuable new resource for building and evaluating more robust and inclusive ASR systems for specialised, real-world applications. --- ### Multi-site PPG: An In-the-Wild Physiological Dataset from Emerging Multi-site Wearables Authors: Jiayi Shao, Jiaying Ye, Shengyao Liu, Zachary Englhardt, Girish Narayanswamy, Vikram Iyer, Qiuyue Shirley Xue Published: 2026-05-18 Categories: cs.HC, cs.LG arXiv: https://arxiv.org/abs/2605.17859v2 Wearables are widely used for mobile health monitoring, and photoplethysmography (PPG) is a key sensing modality for heart rate and related physiological measurements. However, public in-the-wild PPG datasets remain largely wrist-centric or limited to short, controlled studies, constraining research on emerging wearable form factors. We present Multi-site PPG, an in-the-wild physiological dataset collected from four custom-developed unobtrusive wearables: a smart earring, ring, watch, and necklace. Each device records green and infrared reflective PPG, 3-axis acceleration, and temperature with timestamps for cross-device alignment, while a Polar H10 chest strap provides reference electrocardiogram (ECG). Participants wore the devices for multiple days during daytime activities while continuing their normal routines. The dataset contains over 350 hours of raw data and 230-290 hours of modeling-ready 8-second windows per wearable. We benchmark heuristic, supervised, and self-supervised heart-rate estimation methods, showing substantial body-site differences: the best methods achieve mean absolute errors (MAEs) of 2.30 bpm on the earring, 5.13 bpm on the ring, 8.37 bpm on the watch, and 8.68 bpm on the necklace. We further analyze motion effects and evaluate multi-site and PPG-accelerometer fusion, demonstrating the dataset's value for robust physiological sensing across emerging wearable form factors. --- ### KISS - Knowledge Infrastructure for Scientific Simulation: A Scaffolding for Agentic Earth Science Authors: Ziwei Li, Liujun Zhu, Yuchen Liu, Yichen Zhao, Birk Li, Ruiqi Wu, Junliang Jin, Jianyun Zhang Published: 2026-05-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.17856v1 Process-based simulation models encode decades of scientific understanding across the Earth sciences, yet the communities most exposed to climate risk and resource scarcity are the least able to use them. Here, we introduce knowledge infrastructure (KI), an agent-actionable scaffold that externalizes expertise into validated modelling operators, staged domain protocols, and diagnostic recovery mechanisms. Across a 3,000-trial coupled-hydrology benchmark, agents equipped with KI produced physically plausible, verifiable end-to-end simulations in up to 84% of trials, while agents without KI plateaued below 40%. KI generalizes across disciplines. We packaged its construction into a Knowledge Dissection Toolkit (KDT) that autonomously produced KI enabling end-to-end agent execution of 117 additional process-based models across 14 Earth-science domains. Across all 119 KIs, modelling decisions and failure remedies converged despite different underlying physics, showing that operational expertise is structured and extractable rather than ad hoc. Demonstrations show KI-equipped agents lowering both the access barrier between non-specialist users and process-based simulation, and the integration barrier between modelling communities. Through this scaffold, process-based science can then evolve as a living scientific commons, answerable to whoever needs to know and extendable by whoever can contribute. --- ### Learning over Positive and Negative Edges with Contrastive Message Passing Authors: Peter Pao-Huang, Charilaos I. Kanatsoulis, Michael Bereket, Jure Leskovec Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17854v1 Conventional approaches to learning on graphs involve message passing along existing (i.e., positive) edges to update node features. However, these approaches often disregard the potentially valuable information contained in the absence (i.e., negative) of edges. Here, we theoretically analyze the value of negative edges in graph representations and prove that in settings of low label rates, high homophily, and high edge density, access to negative edges provides significant information gain over using only positive edges. Motivated by this insight, we introduce Contrastive Message Passing (CMP), a general message passing architecture that enable graph neural network layers to reason over positive and negative edges. By imposing soft positive semidefinite constraints on the learnable weights, our approach differentially applies similarity-preserving transformations to positively connected nodes and dissimilarity-inducing transformations to negatively connected nodes. Over simulated and real datasets in varying data regimes, CMP consistently outperforms baselines in low-label settings when negative edges are informative. --- ### Simple Approximation and Derivative Free Inference-Time Scaling for Diffusion Models via Sequential Monte Carlo on Path Measures Authors: Chenyang Wang, Weizhong Wang, Yinuo Ren, Jose Blanchet, Yiping Lu Published: 2026-05-18 Categories: stat.ML, cs.CV, cs.LG, math.NA, math.PR arXiv: https://arxiv.org/abs/2605.17850v1 iffusion-based generative models increasingly rely on inference-time guidance, adding a drift term or reweighting mixture of experts, to improve sample quality on task-specific objectives. However, most existing techniques require repeated score or gradient evaluations, introducing bias, high computational overhead, or both. We introduce \texttt{URGE}, Unbiased Resampling via Girsanov Estimation, a derivative-free inference-time scaling algorithm that performs path-wise importance reweighting via a Girsanov change of measure. Instead of computing gradient-based particle weights in previous work, \texttt{URGE} attaches a simple multiplicative weight to each simulated trajectory and periodically resamples. No score, no Hessian, and no PDE evaluation is required. We establish an equivalence between path-wise and particle-wise SMC: the Girsanov path weight admits a backward conditional expectation that recovers the previous particle-level weights, guaranteeing that both schemes produce the same unbiased terminal law. Empirically, \texttt{URGE} outperforms existing inference-time guidance baselines on synthetic tests and diffusion-model benchmarks, achieving better generation quality, while being significantly simpler to implement and fully gradient-free. --- ### Generating Pretraining Tokens from Organic Data for Data-Bound Scaling Authors: Zichun Yu, Chenyan Xiong Published: 2026-05-18 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2605.17849v2 LLM pretraining is shifting from a compute-bound to a data-bound regime, where available human (organic) text falls far short of scaling demands. However, reaching the data-bound regime does not mean the model has fully utilized its organic corpus. In this paper, we introduce SynPro, a synthetic data generation framework that helps LLMs more thoroughly learn from limited organic data. SynPro applies two operations, rephrasing and reformatting, that present the same organic source in diverse forms to facilitate deeper learning without introducing external information. Both generators are optimized via reinforcement learning with quality, faithfulness, and data influence rewards, and are continuously updated as pretraining plateaus to target content the model has yet to absorb. We pretrain 400M, 1.1B, and 2B models with 10% of their Chinchilla-optimal tokens (0.8B, 2.2B, and 4B) from DCLM-Baseline, reflecting a realistic data-bound regime in frontier pretraining. Our results reveal that organic data is significantly underutilized by standard repetition: SynPro unlocks 3.4--5.2x the effective tokens of repetition, even surpassing the non-data-bound oracle that trains on equivalent unique data at the 1.1B and 2B scales. Analyses confirm that faithful, model-aware synthesis sustains data-bound scaling without causing distribution collapse. We open-source our code at https://github.com/cxcscmu/SynPro. --- ### Generating Pretraining Tokens from Organic Data for Data-Bound Scaling Authors: Zichun Yu, Chenyan Xiong Published: 2026-05-18 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2605.17849v1 LLM pretraining is shifting from a compute-bound to a data-bound regime, where available human (organic) text falls far short of scaling demands. However, reaching the data-bound regime does not mean the model has fully utilized its organic corpus. In this paper, we introduce SynPro, a synthetic data generation framework that helps LLMs more thoroughly learn from limited organic data. SynPro applies two operations, rephrasing and reformat, that present the same organic source in diverse forms to facilitate deeper learning without introducing external information. Both generators are optimized via reinforcement learning with quality, faithfulness, and data influence rewards, and are continuously updated as pretraining plateaus to target content the model has yet to absorb. We pretrain 400M and 1.1B models with 10% of their Chinchilla-optimal tokens (0.8B and 2.2B) from DCLM-Baseline, reflecting a realistic data-bound regime in frontier pretraining. Our results reveal that organic data is significantly underutilized by standard repetition: SynPro unlocks 3.7-5.2x the effective tokens of repetition, even surpassing the non-data-bound oracle that trains on equivalent unique data at the 1.1B scale. Analyses confirm that faithful, model-aware synthesis sustains data-bound scaling without causing distribution collapse. We open-source our code at https://github.com/cxcscmu/SynPro. --- ### SNLP: Layer-Parallel Inference via Structured Newton Corrections Authors: Ligong Han, Kai Xu, Hao Wang, Akash Srivastava Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17842v3 Autoregressive language models execute Transformer layers sequentially, creating a latency bottleneck that is not removed by conventional tensor or pipeline parallelism. We study whether this layerwise dependency can be relaxed by treating the hidden-state trace across layers as the solution of a nonlinear residual equation and solving it with parallel Newton-style updates. While this view is principled, exact Newton corrections require expensive Jacobian-vector products and naive fixed-point iterations are unstable on trained Transformers. We introduce Structured Newton Layer Parallelism (SNLP), a training and inference framework that replaces exact layer Jacobians with cheap architecture-induced surrogate dynamics. In residual Transformers, this yields Identity Newton (IDN), where the correction reduces to a prefix-sum-like update; in mHC-style architectures, HC Newton (HCN) uses the model's residual mixing matrix. We also study SNLP-aware training, including pretraining regularization and direct SNLP-forward SFT. Experiments on Nanochat-scale Transformers show that SNLP exposes a practical speed-quality frontier: on 0.5B models, it reaches up to 2.58x wall-clock speedup, and a less aggressive configuration reaches 1.40x speedup without increasing PPL. The useful tradeoff comes from the biased finite-iteration computation induced by IDN/HCN rather than exact recovery of the sequential trace. We further show that SNLP-forward SFT can preserve downstream task accuracy, and that SNLP can serve as a drafter for self-speculative decoding while a sequential verifier preserves output correctness. --- ### Balancing Knowledge Distillation for Imbalance Learning with Bilevel Optimization Authors: Anh B. H. Nguyen, Ba Tho Phan, Viet Cuong Ta Published: 2026-05-18 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.17839v3 Knowledge distillation transfers knowledge from a high capacity teacher to a compact student using a mixture of hard and soft losses. On imbalanced data, a fixed weighting between hard and soft losses becomes brittle the learning process. Recent studies try to reweight these components in long-tailed settings. However, most of these methods do not adapt weights at the sample-wise level and do not take into account the students behavior during training. To address this, we propose BiKD -- a bilevel framework that dynamically balances hard and soft losses for each sample. We employ a weight generation network that produces adaptive per-sample weights, guided by a small balanced validation set. The student is now trained with an unconstrained combination of weighted hard and soft losses, allowing the student to relax both terms. We further propose a multi-step SGD strategy to optimize the weight model more accurately and efficiently. Experiments on long-tailed CIFAR-10/100 show that our approach surpasses recent balanced distillation methods across imbalance factors. --- ### Temporal Aware Pruning for Efficient Diffusion-based Video Generation Authors: Sheng Li, Yang Sui, Junhao Ran, Bo Yuan, Yue Dai, Xulong Tang Published: 2026-05-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2605.17837v2 Video diffusion models have recently enabled high-quality video generation with ViT-based architectures, but remain computationally intensive because generation requires attention computation over long spatiotemporal sequences. Token pruning has proven effective for ViTs and VLMs. However, most prior pruning methods are attention-based and operate per frame, failing to ensure the vital temporal coherence across frames in video generation tasks. In practice, naively adopting attention-only pruning causes noticeable degradation due to worsened background consistency, flickering, and reduced image quality. To address this, we propose TAPE, a training-free Temporal Aware Pruning for Efficient diffusion-based video generation. TAPE (i) applies temporal smoothing to align token-importance across adjacent frames and suppress selection jitter; and (ii) performs token reselection in selected layers to align token pruning with layers' diverse semantic focus and avoid error accumulation in specific areas; it also (iii) adopt a timestep-level budget scheduling that prunes aggressively at early noisy steps and relaxes pruning during fidelity-critical refinement. The experimental results show that TAPE delivers significant speedups while preserving high visual fidelity, outperforming prior token reduction approaches. --- ### Stabilizing, Scaling & Enhancing MeanFlow for Large-scale Diffusion Distillation Authors: Xiao He, Yang Li, Peizhen Zhang, Songtao Liu, Zhao Zhong, Nannan Wang Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17834v1 Diffusion models exhibit remarkable generative capability, but their high latency limits practical deployment. Many studies have attempted to reduce sampling steps to accelerate inference. Among them, MeanFlow has attracted considerable attention due to its concise formulation and remarkable performance. Nevertheless, the instability of its optimization objective and the ''mean-seeking bias'' have limited its applicability to distill large-scale industrial models. To stabilize MeanFlow for distilling large-scale models, we first introduce a warm-up technique, in which the original differential solution of MeanFlow is replaced by a discrete solution. This design avoids training collapse caused by the MeanFlow target containing a stop-gradient term from an undertrained model. Once the model acquires a preliminary ability to fit the average velocity field, we switch the optimization objective back to the differential solution, enabling further refinement. Meanwhile, to alleviate the ''mean-seeking bias'' of MeanFlow under extremely few-step inference with complex target distributions, we incorporate trajectory distribution alignment as an auxiliary objective, encouraging the student model's trajectory distribution to align more closely with that of the teacher model. Our proposed distillation framework achieves superior performance compared to existing distillation approaches when applied to the text-to-image (T2I) model FLUX.1-dev (up to 12B parameters). Furthermore, when extended to the 80B-parameter state-of-the-art (SOTA) T2I model HunyuanImage 3.0, our method continues to demonstrate robust generalization and strong performance. --- ### Efficient Bilevel Optimization for Meta Label Correction in Noisy Label Learning Authors: Ba Hoang Anh Nguyen, Viet Cuong Ta Published: 2026-05-18 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.17833v1 Training a deep neural network with noisy labels could reduce data annotation cost but may introduce noise into the learned model. In meta label correction approaches, an additional meta model besides the main model is trained with a small, clean dataset to correct the large, noisy dataset. However, the update of the meta model requires the computation of hypergradients at the inner step of the main model which signif- icantly increases the computational cost. To improve the training efficiency, we first introduce the dynamic barrier gradient descent into standard meta label correction. While this naive extenstion is able to speed up the training process to approximately first- order complexity, it lacks mechanisms to prevent the leakage of noisy signals to the main model and to stabilize the learning of the meta model. Based on this observation, we propose the EBOMLC method, which is designed with three key improvements including one-step inner loop update, mixture upper loss and alignment- aware dynamic barrier. Empirical results on CIFAR-10 and CIFAR-100 demonstrate that EBOMLC consistently outperforms other baselines, especially under high noise rate settings, while reducing training time of the meta label correction approach. --- ### Agentic Cost-Aware Query Planning with Knowledge Distillation for Big Data Analytics Authors: Mahdi Naser-Moghadasi Published: 2026-05-18 Categories: cs.LG, cs.DB arXiv: https://arxiv.org/abs/2605.17831v1 Query optimization in big data analytics remains computationally expensive, particularly for resource-constrained environments where traditional optimizers fail to satisfy memory and latency constraints. We present an agentic query planning system that combines a rule-based teacher planner, UCB1 bandit exploration, cost-aware prediction, and knowledge distillation to a lightweight student planner. Our teacher planner generates SQL plans using six key optimization strategies, while UCB1 bandit search efficiently explores the plan space under explicit resource constraints. A Random Forest cost model predicts query latency from plan features, enabling cost-aware decisions. A distilled student planner (Logistic Regression or Gradient Boosting) learns to mimic teacher-bandit decisions for fast inference. Evaluation on NYC Taxi and IMDB datasets demonstrates 23% latency reduction compared to default planners while maintaining 94% constraint satisfaction. The student planner achieves 89% accuracy in replicating optimal plans with 15x faster inference time. Our single-file implementation enables reproducible big-data analytics on resource-limited machines and is publicly available at https://github.com/mahdinaser/agentic-kd-planner. --- ### Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents Authors: Ahmad Al-Tawaha, Shangding Gu, Peizhi Niu, Ruoxi Jia, Ming Jin Published: 2026-05-18 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2605.17830v1 Safety evaluations of memory-equipped LLM agents typically measure within-task safety: whether an agent completes a single scenario safely, often under adversarial conditions such as prompt injection or memory poisoning. In deployment, however, a single agent serves many independent tasks over a long horizon, and memory accumulated during earlier tasks can affect behavior on later, unrelated ones. Studying this regime requires evaluation along the temporal dimension across tasks: not whether an agent is safe at any single memory state, but how its safety profile changes as memory accumulates across many independent interactions. We call this failure mode temporal memory contamination. To isolate memory exposure from stream non-stationarity, we introduce a trigger-probe protocol that evaluates a fixed probe set against read-only memory snapshots at varying prefix lengths, together with a NullMemory counterfactual baseline for identifying memory-induced violations. We apply this protocol across three deployment scenarios spanning records, memos, forms, and email correspondence and eight memory architectures, and additionally on Claw-like AI agents, such as OpenClaw, using the platform's native memory mechanism. Memory-enabled agents consistently exceed the NullMemory baseline, and memory-induced violation rates show a robust upward trend with exposure length on both agent classes. Order-randomization experiments indicate that the effect is driven primarily by accumulated content rather than encounter order. Finally, a structural consequence of the event decomposition is that memory-induced risk is detectable from retrieval state before generation, which we confirm with a high-recall diagnostic monitor. Our results argue for treating memory safety as a longitudinal property that requires temporal evaluation, not a single-state property that can be captured by a snapshot. --- ### Interactive Evaluation Requires a Design Science Authors: Keyang Xuan, Peiyang Song, Pan Lu, Pengrui Han, Wenkai Li, Zhenyu Zhang, Zexue He, Wenyue Hua, Manling Li, Jiaxuan You, Adrian Weller, Yizhong Wang et al. Published: 2026-05-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.17829v1 AI evaluation is undergoing a structural change. Large language models (LLMs) are increasingly deployed as systems that act over time through tools, environments, users, and other agents, while many evaluation practices still inherit assumptions from response-centered benchmarks (e.g., fixed inputs, isolated outputs, and outcome judgments that can be made from a single response). The field has begun to build interactive benchmarks, but the resulting landscape is fragmented: benchmarks differ in what interaction artifacts they admit, how trajectories are scored, and what claims their results support. This position paper argues that interactive evaluation should be treated as a principled evaluation paradigm, not merely a new family of agent benchmarks. Simply adopting previous evaluation paradigms does not suffice. We define evaluation as an autonomous mapping from evidence to judgments, and show that interactive evaluation changes both sides of this mapping: the evidence becomes interaction-generated trajectories, while the evaluation procedure must assess process, recoverability, coordination, robustness, and system-level performance. Building on this definition, we propose a two-axis taxonomy, derive design principles and reporting standards, examine representative scenarios, and analyze how longstanding evaluation challenges reappear at the trajectory level. --- ### Content-Style Identification via Differential Independence Authors: Subash Timilsina, Hoang-Son Nguyen, Sagar Shrestha, Xiao Fu Published: 2026-05-18 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2605.17827v1 Generative analysis often models multi-domain observations as nonlinear mixtures of domain-invariant content variables and domain-specific style variables. Identifying both factors from unpaired domains enables tasks such as domain transfer and counterfactual data generation. Prior work establishes identifiability under (block-wise) statistical independence between content and style, or via sparse Jacobian assumptions on the nonlinear mixing function, but such conditions can be restrictive in practice. In this work, we introduce content-style differential independence (CSDI), an alternative structural condition requiring that infinitesimal variations in content and style induce orthogonal directions on the data manifold, thereby enabling identifiability even when content and style are dependent and the Jacobian is dense. We operationalize this condition through a blockwise orthogonality constraint on the Jacobian subspaces associated with content and style. To support high-dimensional generative models, we design a stochastic regularizer based on numerical Jacobian approximation, enabling scalable training in settings such as high-resolution image generation. Experiments across multiple datasets corroborate the identifiability analysis and demonstrate practical benefits on counterfactual generation and domain translation. --- ### CounterCount: A Diagnostic Framework for Counting Bias in Vision Language Models Authors: Reem Alzahrani, Hassan Alshanqiti, Bushra Bin Hemid, Zaid Alyafeai, Abdelrahman Eldesokey, Bernard Ghanem Published: 2026-05-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2605.17826v1 Vision-Language Models (VLMs) excel at multimodal reasoning, yet it remains unclear whether their answers are grounded in visual evidence or driven by learned language and world priors. Counting provides a precise testbed: when visual evidence conflicts with canonical object knowledge, a model must rely on the image rather than a prototypical count. We introduce CounterCount, a diagnostic framework for counterfactual counting in VLMs, consisting of paired factual and counterfactual images with edited count-relevant attributes, verified answers, and localized evidence annotations. Evaluating recent VLMs, we find strong performance on factual images but consistent degradation under counterfactual attribute changes, indicating reliance on object-level priors even when contradictory visual evidence is present. Using localized annotations, we show that these failures are not solely due to missing or ambiguous visual evidence, but to models underweighting attention to count-relevant visual tokens. We introduce a unified inference-time attention modulation strategy that reweights selected visual tokens, improving counterfactual counting accuracy by up to 8% across multiple VLMs. Overall, CounterCount exposes prior-driven counting failures and provides diagnostic insights for designing future VLMs. --- ### Why We Look Where We Look: Emergent Human-like Fixations of a Foveated Visual Language Model Maximizing Scene Understanding Authors: Shravan Murlidaran, Ziqi Wen, Sana Shehabi, Miguel P. Eckstein Published: 2026-05-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2605.17823v1 When humans view scenes without a specific task (free-viewing), they initially direct their eye movements toward the scene center and then fixate on people, text, objects being gazed at or grasped, and semantically meaningful regions. What these signature fixation patterns reflect and whether they optimize an underlying perceptual task remain unknown. We show that a computational agent with simulated foveation, trained to optimize scene comprehension, exhibits emergent human fixation signature patterns. In contrast, versions of the agent trained to search or classify scenes, or equipped with peripheral vision that was better or worse than human vision, predicted human fixation patterns less accurately. Thus, human free-viewing fixation patterns may emerge as a functional byproduct of optimizing scene comprehension under the biological constraints of foveated vision. --- ### Unleashing the Representational Power of Fourier Shapes for Attacking Infrared Object Detection Authors: Yixing Yong, Jian Wang, Ming Lei, Lijun He, Fan Li Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17822v1 Infrared object detection is crucial for perception in autonomous driving and surveillance but remains vulnerable to physical adversarial attacks. Unlike in the RGB domain, where attacks rely on color texture, infrared attacks must manipulate thermal signatures, making the geometry shape of heat-blocking materials the primary adversarial information carrier. Current shape-based methods suffer from a fundamental trade-off between representational capability and optimization power, limiting their attack effectiveness.In this work, we overcome this dilemma by introducing learnable Fourier shapes to the infrared domain. We utilize an end-to-end differentiable framework where a compact set of Fourier coefficients, defining the shape boundary, is analytically mapped to a pixel-space mask via the winding number theorem. This enables efficient gradient-based optimization to generate potent shapes that cause human targets to evade detection. Extensive digital and physical experiments provide a comprehensive evaluation and validate our superior performance. Our resulting physical patch achieves striking robustness, successfully evading detectors across diverse distances, angles, poses, and individuals, and achieves over 88% attack success rate at distances greater than 25m (conf.=0.5). Code is available at https://github.com/Yongyx99/Fourier-shape-attack. --- ### TierCheck: Tiered Checkpointing for Fault Tolerance in Large Language Model Training Authors: Shujie Han, Feng Jiang, Patrick P. C. Lee, Xiao Zhang, Zhijie Huang, Nannan Zhao, Xiaonan Zhao, Lichen Pan Published: 2026-05-18 Categories: cs.DC, cs.AI arXiv: https://arxiv.org/abs/2605.17821v1 Large Language Model (LLM) training is frequently interrupted by a heterogeneous spectrum of failures, from common GPU crashes to catastrophic cluster-wide outages. Existing checkpointing systems rely on monolithic, single-tier storage backend, forcing a trade-off between state-saving overhead and recovery speed. We propose TierCheck, a cluster-aware tiered checkpointing system that aligns storage placement with failure heterogeneity. TierCheck adopts a three-tier design that maintains lightweight differential checkpoints in local and peer memory for fast localized recovery, while asynchronously migrating heavyweight base checkpoints to remote persistent storage. It also ensures strict global consistency across tiers without stalling training, and achieves fast cluster-aware checkpoint restoration during recovery. Evaluations on models up to 40 billion parameters show that TierCheck achieves low training overhead, reduces end-to-end checkpointing time to under 10s, and supports high-frequency checkpointing, ultimately striking an optimal balance between low-overhead persistence and fast recovery. --- ### Evidence-Guided Unknown Rejection for High-Confidence Near-Known Unknowns Authors: Xi Chen, Yingjun Xiao, Gang Fang Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17818v1 Open-set recognition systems face a neglected failure mode: high-confidence near-known unknowns, which lie outside the known label set but are close enough to known classes that a closed-set classifier accepts them with high confidence. We show that this failure is widespread across scalar-threshold methods, including recent post-hoc detectors, and that stronger encoders can amplify rather than remove the risk. We propose EGUR-A, which changes the decision from ``is this sample's score high enough?'' to ``does this predicted known class have sufficient evidence to accept this sample?'' EGUR-A combines class-conditional local acceptance evidence with global residual evidence, and selects their relative weight from known-sample statistics without unknown validation data. Across CUB, FGVC-Aircraft, and ImageNet-hard, EGUR-A substantially reduces high-confidence false known acceptance at matched known-rejection operating points. The result is not a stronger threshold; it is a different question: whether a known class is entitled to accept a sample. --- ### Virtues of Ordered Chaos: Planning with Topple Actions in Tabletop Stack Rearrangement Authors: Hao Lu, Rahul Shome Published: 2026-05-18 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2605.17815v1 Efficient object manipulation strategies have significant impact in automation applications. In this work, the stack rearrangement in tabletop settings is studied, with a focus on augmenting the task planning domain with richer nonprehensile aggregating actions, in particular the toppling of objects from a stack to the table. Toppling can compress long sequences of intermediate relocations. Computed plans need to interleave pick-and-place actions with topple throughout its plan based on the problem. In order to generate the task plan and model an abstraction to compute solutions that include both pick-and-place and topple actions, a novel aggregating gadget for topple is introduced. Using this directed graphical abstraction, candidate task plan computation becomes a variant of the pebble motion problem, treating objects as pebbles. Benchmarks are then reported in a IsaacSim-based physics simulation. Results highlight clear benefits of achieving faster execution than solely using pick-and-place actions. Though this work primarily investigates the topple action, we demonstrate that similar abstractions can model other aggregating actions of interest, like scoop. The current work provides a preliminary, strong indication of the promising benefits of abstractions for rich object interactions in manipulation applications. --- ### Going Headless? On the Boundaries of Vertical AI Firms Authors: Muhammad Zia Hydari, Farooq Muzaffar Published: 2026-05-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.17812v1 Vertical AI firms in accounting, law, healthcare, procurement, and similar domains historically bundled workflow, domain logic, and accountability into a single application. General-purpose AI agents are now unbundling that package, prompting founders and investors to advocate "going headless": cede the workflow and interface to agents and expose domain expertise as callable services. This article argues that going headless is correct for some firms and destructive for others, and that the latter often cede their value capture inadvertently through architectural choices that look like interface decisions. This is a boundary question, and the answer turns on distinguishing the interface boundary, which can often move, from the accountability boundary, which often must not. Drawing on Coase's theory of the firm, Eisenmann, Parker, and Van Alstyne's platform envelopment framework, and Teece's analysis of complementary assets and appropriability, the article shows that orchestrators operating through open protocols acquire envelopment power even as technical interoperability improves, and that durable value capture concentrates in cospecialized accountability assets: professional signoff, regulated workflows, evidence trails, and trusted systems of record. The article proposes a three-position taxonomy (component, integrated software platform, dual-track) determined not by sector but by task-accountability regime, and formalizes the construct of rule debt: the future governance, maintenance, and accountability burden that accrues to customer organizations when business rules and professional standards migrate from governed systems into prompts and agent instructions. Four principles follow: decompose by accountability not interface, invert the edges while retaining the core, position rule debt as the customer cost the integrated platform prevents, and avoid single-orchestrator dependence. --- ### One Model, Two Roles: Emergent Specialization in a Shared Recurrent Transformer Authors: Jucheng Shen, Barbara Su, Anastasios Kyrillidis Published: 2026-05-18 Categories: cs.LG, cs.AI, math.OC arXiv: https://arxiv.org/abs/2605.17811v1 Can a shared-weight recurrent Transformer develop distinct internal roles without being partitioned into separate modules? We study this in Asymmetric Input Recurrence (AIR), a minimal two-state reasoning architecture in which the same Transformer model is reused for both updates (per literature, L and H) and the only built-in difference in the update rule is that the encoded input is injected during L-updates but not H-updates. Across Sudoku-Extreme and Maze, decoded rollouts reveal a consistent split: $\zH$ behaves like a fully committed proposal state, whereas $\zL$ retains local uncertainty and shifting intermediate structure. Freeze experiments show that this split is, in practice, related to the model's state dynamics: in Sudoku, freezing $\zH$ reduces $\zL$'s content changes whereas freezing $\zL$ increases $\zH$'s, while in Maze, freezing either state increases content changes in the other state. Ablations show that to induce specialization, the shared model needs to be able to tell the two update types apart, either from input injection asymmetry or from a separate level token. Mechanistically, attention analysis shows that L-updates are consistently more local than H-updates in both Sudoku and Maze. Together, these results show that, in a two-state recurrent setting, a clear state-identity signal can induce stable, related functional roles inside a shared-parameter recurrent Transformer. Code is available at \href{https://github.com/juchengshen/air}{\textcolor{blue}{https://github.com/juchengshen/air}}. --- ### Accelerating AI-Powered Research: The PuppyChatter Framework for Usable and Flexible Tooling Authors: Chun-Hsiung Tseng, Hao-Chiang Koong Lin, Andrew Chih-Wei Huang, Yung-Hui Chen, Jia-Rou Lin Published: 2026-05-18 Categories: cs.AI, cs.IR arXiv: https://arxiv.org/abs/2605.17809v1 This research addresses the challenges inherent in developing Artificial Intelligence (AI) applications, particularly those leveraging Large Language Models (LLMs). While AI vendors provide Application Programming Interfaces (APIs) and Software Development Kits (SDKs) to facilitate developer interaction, the former often requires intricate manual request construction, and the latter can lead to significant vendor lock-in. Furthermore, existing model abstraction frameworks, though mitigating vendor dependency, introduce an additional layer of complexity and potential security concerns. To reconcile these conflicting factors, the study introduces PuppyChatter, a novel software framework designed to preserve the intuitive simplicity of vendor-specific SDKs while simultaneously adhering to the vendor-neutrality principles characteristic of model abstraction, thereby offering a more streamlined and flexible development paradigm. --- ### A Unified Framework for Data-Free One-Step Sampling via Wasserstein Gradient Flows Authors: Chenguang Wang, Tianshu Yu Published: 2026-05-18 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2605.17808v1 We develop a unified theoretical framework for data-free one-step sampling from unnormalized target distributions based on Wasserstein gradient flows. For a broad class of standard f-divergence objectives, we show that the induced velocity field admits the universal form $\mathbf{V}(x)=w(r(x))\,β(x)$, where $β(x)=\nabla \log (p(x)/q(x))$ is shared across objectives and $w$ is determined solely by the choice of divergence. This decomposition shows that standard f-divergence drifts share the same asymptotic target distribution $p$ and differ primarily in how they redistribute transient repair effort across under-covered regions. To formalize this distinction, we derive a one-step regional-response theory for a soft under-coverage functional and obtain a compression--elasticity identity that links divergence choice to the geometry of mass transport into under-covered regions. We further extend the framework beyond the f-divergence family to the Log-Variance (LV) divergence, analyze how the reference distribution alters the resulting drift structure, and motivate a practical LV-inspired surrogate for data-free training. Based on this theory, we instantiate the framework with a KDE-based implementation and describe a complementary normalizing-flow route, enabling one-step inference after training. Experiments on multimodal Gaussian-mixture benchmarks are consistent with the theoretical predictions and demonstrate effective one-step sampling on these targets. --- ### Curriculum Group Policy Optimization: Adaptive Sampling for Unleashing the Potential of Text-to-Image Generation Authors: Baoteng Li, Xianghao Zang, Xinran Wang, Xiangyu Na, Zhixiang He, Hao Sun, Chi Zhang, Zhongjiang He, Tianwei Cao, Kongming Liang, Zhanyu Ma Published: 2026-05-18 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2605.17807v1 Text-to-Image (T2I) generation has achieved remarkable progress in recent years. Meanwhile, reinforcement learning methods, particularly those based on Group Relative Policy Optimization (GRPO), have attracted widespread attention and been successfully applied to T2I tasks. However, the uniform sampling strategy commonly used during training often ignores the match between sample difficulty and the model's current learning capability, leading to low training efficiency. We argue that improving training efficiency requires continuously prioritizing prompts that match the model's evolving capability and remain actively learnable. To this end, we propose Curriculum Group Policy Optimization (CGPO), an adaptive curriculum training framework. During training, each prompt produces a group of images scored by a reward model. We use the variance of group rewards as an online proxy for prompt inconsistency. A higher variance suggests that the model has partially captured the prompt requirements but has not yet achieved stable mastery. Such prompts are more likely to provide useful learning signals, so we increase their sampling probabilities accordingly. Additionally, to address data imbalance in multi-category datasets, we design a category calibration method based on proportional fairness optimization, which balances training difficulty across categories. Experiments on GenEval, T2I-CompBench++, and DPG Bench demonstrate that our framework effectively improves generation performance. --- ### AMO: Operator-level Adaptive Muon Orthogonalization Authors: Xinlin Zhuang, Panyi Ouyang, Yichen Li, Jiangming Shi, Yizhang Chen, Shuman Liu, Ying Qian, Weiyang Liu, Haibo Zhang, Imran Razzak Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17806v2 Muon has recently emerged as a competitive alternative to AdamW for large-scale pre-training, with orthogonalization via Newton-Schulz (NS) iteration as its core operation. Standard Muon applies a uniform NS schedule to all parameter matrices, overlooking possible differences in orthogonalization difficulty and its impact on performance. Through a systematic empirical study, we show that this per-matrix heterogeneity is pervasive and strongly associated with matrix geometry, which evolves dynamically across operator types, training stages, and network depths. Therefore, uniform NS schedules can lead to uneven orthogonalization quality across the model. Motivated by these findings, we propose Operator-level Adaptive Muon Orthogonalization (AMO), an observe-then-commit method that measures weight geometry by operator type early in training and then uses these signals to allocate the NS budget for the remainder of training. AMO delivers consistent improvements over uniform-schedule Muon across standard, prolonged, and continual pre-training, surpassing the strongest baseline by +0.76 on Llama3.1-1.4B and +0.51 on Qwen3-1.7B in average downstream performance of 12 evaluation tasks, with gains persisting at Llama3.1-4B scale. --- ### AMO: Adaptive Muon Orthogonalization Authors: Xinlin Zhuang, Panyi Ouyang, Yichen Li, Jiangming Shi, Yizhang Chen, Shuman Liu, Ying Qian, Weiyang Liu, Haibo Zhang, Imran Razzak Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17806v1 Muon has recently emerged as a competitive alternative to AdamW for large-scale pre-training, with orthogonalization via Newton-Schulz (NS) iterations as its core operation. Existing Muon variants apply a uniform NS schedule to all parameter matrices, overlooking possible differences in orthogonalization difficulty and its impact on performance. Through a systematic empirical study, we show that this per-matrix heterogeneity is pervasive and largely determined by matrix geometry, which evolves dynamically across operator types, training stages, and network depths. As a result, uniform NS schedules can lead to uneven orthogonalization quality across the model. Motivated by these findings, we propose Adaptive Muon Orthogonalization (AMO), an observe-then-commit method that measures weight geometry by operator type early in training and then uses these signals to allocate the NS budget for the remainder of training. AMO delivers consistent improvements over uniform-schedule Muon across standard, prolonged, and continual pre-training, surpassing the strongest baseline by +0.76 on Llama3.1-1.4B and +0.51 on Qwen3-1.7B in average downstream performance of 12 evaluation tasks. --- ### GenTS: A Comprehensive Benchmark Library for Generative Time Series Models Authors: Chenxi Wang, Xiaorong Wang, Peiyang Li, Yi Wang Published: 2026-05-18 Categories: cs.LG, eess.SP arXiv: https://arxiv.org/abs/2605.17804v2 Generative models have demonstrated remarkable potential in time series analysis tasks, like synthesis, forecasting, imputation, etc. However, offering limited coverage for generative models, existing time series libraries are mainly engineered for discriminative models, with standardized workflows for specific tasks, such as optimizing Mean Squared Errors for time series forecasting. This rigid structure is fundamentally incompatible with the distinct and often complex paradigms of generative models (e.g., adversarial training, diffusion processes), which learn the underlying data distribution rather than a direct input-output mapping. To this end, we proposed GenTS, a comprehensive and extensible benchmark library designed for systematic assessment on generative time series models. GenTS features a unified data preprocessing pipeline, a collection of versatile models, and panoramic evaluation metrics. Its modular design also enables the researchers to flexibly customize beyond our built-in datasets and models. Based on GenTS, we conducted benchmarking experiments under diverse tasks, accordingly offering suggestions for model selection and identifying potential directions for future research. Our codes are open-source at https://github.com/WillWang1113/GenTS. The official tutorials and document are available at https://willwang1113.github.io/GenTS/. --- ### Optimal Knock-Pick Planning for Tightly Packed Tabletop Blocks With Parallel Grippers Authors: Hao Lu, Rahul Shome Published: 2026-05-18 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2605.17800v1 Rearranging densely packed tabletop objects is challenging when parallel-gripper picks are infeasible without sufficient clearance around an object. This work studies the problem characteristics for practically motivated settings with uniformly sized blocks placed at planar tabletop grid locations. Since purely prehensile removal can become infeasible, a directional knock primitive is therefore introduced and the optimal knock-pick variant of the problem is formulated. The work proposes a series of abstractions wherein minimal constraining gadgets are covered to identify the necessary knocks. Utilizing a maximum-weight perfect matching on a graphical abstraction yields efficient polynomial-time computation of the optimal plan that minimizes the number of actions. Experiments are reported for increasing grid sizes in synthetic settings as well as in IsaacSim. The theoretical observations provide a promising stepping stone towards rigorously building efficient manipulation strategies that interleave prehensile and non-prehensile actions. --- ### Is Complex Training Necessary for Long-Tailed OOD Detection? A Re-think from Feature Geometry Authors: Ningkang Peng, Xuanming Chen, Yanhui Gu Published: 2026-05-18 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2605.17799v1 Long-tailed out-of-distribution (LT-OOD) detection is often addressed with specialized training, including auxiliary out-of-distribution (OOD) data, abstention heads, contrastive objectives, energy losses, or gradient-conflict control. We show that these training mechanisms can obscure a simpler issue: frozen long-tailed representations may already contain useful OOD evidence, but raw Mahalanobis distance is distorted by frequency-coupled feature radius and poorly supported tail covariance. We propose Hyperspherical Pooled Mahalanobis (HPM), a post-hoc detector that normalizes features onto the unit sphere and replaces class-specific covariance with a pooled, ridge-regularized metric while keeping class means as semantic anchors. In CIFAR-LT experiments and an ImageNet-100-LT near-OOD boundary analysis, HPM improves raw Mahalanobis scoring; for Prior-Calibrated ERM (PC-ERM), it raises AUROC from 46.49 to 85.67 on CIFAR-10-LT and from 50.40 to 78.35 on CIFAR-100-LT. This simple PC-ERM+HPM pipeline also achieves the best Log Efficiency Score (LES; 3.08) on CIFAR-100-LT, retaining roughly 95% of the best CIFAR-100-LT AUROC observed among the compared post-hoc scores at substantially lower training-time cost. These results argue for evaluating representation quality, detector geometry, and training complexity as separate factors in LT-OOD detection. --- ### When Accuracy Is Not Enough: Uncertainty Collapse between Noisy Label Learning and Out-of-Distribution Detection Authors: Ningkang Peng, Jingyang Mao, Runhan Zhou, Peirong Ma, Yanhui Gu Published: 2026-05-18 Categories: cs.LG, cs.CV arXiv: https://arxiv.org/abs/2605.17795v1 Learning with noisy labels (LNL) is typically benchmarked by closed-set classification accuracy, yet deployment often requires classifiers to reject out-of-distribution (OOD) inputs. We present a learner-agnostic ACC-OOD benchmark that freezes LNL checkpoints and evaluates them with standardized near-/far-OOD routing and post-hoc scores across synthetic and real label noise. The benchmark reveals a recurring failure mode: high closed-set accuracy does not ensure OOD reliability, because low-confidence, misclassified in-distribution samples can overlap the score and feature regions occupied by OOD inputs under noisy training. We term this pathology uncertainty collapse. This structural overlap can make high-accuracy LNL methods lose separability at the ID-error/OOD interface under standard OOD scores. As an intervention, we study Virtual Margin Regularization (VMR), a lightweight repair probe demonstrated mainly with PSSCL that synthesizes boundary virtual outliers on trusted ID batches and widens the energy margin. VMR partially reduces the collapse-induced far-OOD failure without replacing the host objective or sacrificing closed-set accuracy in the tested settings. These results support LNL benchmarks that co-report closed-set generalization, open-world reliability, and structural overlap diagnostics. --- ### HydroAgent: Closing the Gap Between Frontier LLMs and Human Experts in Hydrologic Model Calibration via Simulator-Grounded RL Authors: Zhi Li, Songkun Yan, Jie Cao, Mofan Zhang, Anjiang Wei, Jinwoong Yoo, Yang Hong Published: 2026-05-18 Categories: cs.LG, physics.geo-ph arXiv: https://arxiv.org/abs/2605.17792v1 Calibrating distributed hydrologic models is a critical bottleneck across operational water resources management - streamflow prediction, reservoir operation, drought monitoring, infrastructure design, and flood forecasting all depend on it. Each basin demands an expert to translate hydrograph signatures into adjustments of a high-dimensional parameter vector, and the resulting workflow does not transfer between watersheds. We ask: can frontier large language model (LLM) agents replace the human hydrologic modeler, and if not, what would it take? We benchmark nine frontier LLM agents - Claude Opus 4.6/4.7, Sonnet 4.6, GPT-5/5.4/5.4-pro, and Gemini 2.5-pro/3.1-pro/3-flash - on the operational CREST distributed hydrologic model used by the U.S. National Weather Service for flash-flood forecasting. Best-of-twenty-rounds Nash-Sutcliffe Efficiency (NSE) across four held-out gauges spanning 329-40,792 km2 ranges from -0.16 (GPT-5.4) to 0.75 (Sonnet 4.6); the ceiling reproduces across all three vendors and capability tiers, with the strongest models concentrating in the 0.65-0.75 band, and no model reaches the human-expert reference except Opus-4.7 on one gauge. We argue this gap is not a parameter-count problem but a domain-grounding problem. We then propose HYDROAGENT, fine-tuning open-weight Qwen3-4B with supervised fine-tuning on 2,576 expert calibration trajectories and Group-Relative Policy Optimization using NSE as a verifiable reward from online CREST simulations - reinforcement learning with simulation feedback (RLSF). For Earth system science, a small domain-tuned policy with simulator-in-the-loop RL is a more compute-efficient and physically faithful path than scaling generic frontier models, and the multi-modal richness of Earth data - remote sensing, in-situ time series, and forecaster narrative - makes domain agents a leveraged direction for AI in physical science. --- ### STRIDE: A Self-Reflective Agent Framework for Reliable Automatic Equation Discovery Authors: Jiarui Su, Songjun Tu, Bei Sun, Xiaojun Liang Published: 2026-05-18 Categories: cs.AI arXiv: https://arxiv.org/abs/2605.17790v1 LLM-based equation discovery offers a promising route to recovering symbolic laws from data, but many systems still rely on generation-centered loops that propose candidates, fit parameters, score results, and reuse selected examples. Such loops can misjudge useful skeletons under unreliable fitting, discard near-correct equations that require repair, and accumulate redundant memories that provide limited guidance. We propose STRIDE, a self-reflective agent framework that improves reliability by coordinating data-aware generation, mixed-fitting evaluation, critic--executor repair, and diversity-preserving semantic memory. By turning fitted scores and candidate behavior into shared feedback, STRIDE enables equations to be proposed, assessed, refined, and reused within a closed-loop discovery process. Experiments on representative symbolic-regression benchmarks and LSR-Synth suites show that STRIDE improves accuracy, OOD robustness, and structural recovery across multiple LLM backbones, with ablations and analyses confirming the contribution of its core components. --- ### SocialMemBench: Are AI Memory Systems Ready for Social Group Settings? Authors: Olukunle Owolabi Published: 2026-05-18 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2605.17789v1 Memory systems for AI assistants were built for single-user dialogue and fail characteristically when applied to multi-party social group settings. This gap matters for the social assistants being built today: group-acting agents embedded in chat platforms, and proactive personal-assistant agents whose holistic model of a user must include their social context. Existing memory benchmarks evaluate dyadic or workplace dialogue; none targets multi-party social groups, where memory must anchor facts in shared history rather than professional roles, separate group norms from individual exceptions, and correctly attribute even after member departure. We introduce SocialMemBench, a benchmark of human-verified synthetic social group networks across five archetypes (close friends, family, recreational, interest community, acquaintance network) and three group-size tiers (4-30 members), with 430 personas and 7,355 conversation turns, yielding 1,031 QA pairs across nine question categories. Each category isolates an architectural capability, and the five failure modes (single-stream conflation, temporal-state overwrite, entity merging at scale, missing cross-persona knowledge, norm-individual conflation) are testable hypotheses; our two research probes Subject-Mem and SMG provide evidence on two, three remain open. A full-context Gemini 2.5 Flash reference reaches only 0.721 against a blind-critic reasoning-model mean of 0.98 on small networks, indicating the benchmark is genuinely difficult even with complete access to the conversation. Across all 43 networks, the four open-source memory frameworks evaluated (Mem0, LangMem, Graphiti, Cognee) cluster in the 0.12-0.18 question-weighted range with overlapping 95% CIs, well below an uncompressed retrieval reference of 0.345 and a matched-answerer full-context reference of 0.369 (GPT-4o-mini). Current memory systems show a measurable gap. --- ### Uncertainty-Calibrated Recommendations for Low-Active Users Authors: Bob Junyi Zou, Sai Li, Tianyun Sun, Wentao Guo, Qinglei Wang Published: 2026-05-18 Categories: cs.IR, cs.LG arXiv: https://arxiv.org/abs/2605.17788v2 A fundamental challenge in recommender systems is balancing reliability for Low-Active Users (LAUs) with diversity for High-Active Users (HAUs). The key to this balance lies in quantifying model uncertainty, which approximates the risk of prediction errors and reveals the limits of the model's current knowledge. On large-scale short-video and livestream platforms, model uncertainty can warn of low-quality recommendations that may lead to disengagement of LAUs and at the same time identify opportunities to diversify content recommendation for HAUs. To leverage this dichotomy, we introduce a unified, production-ready framework that calibrates uncertainty to drive differentiated strategies. Specifically, we implement a model-uncertainty-based risk-averse deboosting policy for LAUs to suppress unreliable recommendations, while employing a risk-seeking Upper Confidence Bound (UCB) strategy for HAUs to encourage exploration. Validated on a major livestream platform, our framework demonstrates significant improvements in retention (active hours) and satisfaction (quality watch time ratio) for LAUs as well as remarkable increases in interest diversity and category coverage for HAUs, proving the value of uncertainty-aware recommendation in industrial settings. --- ### Revisiting the Adam-SGD Gap in LLM Pre-Training: The Role of Large Effective Learning Rates Authors: Athanasios Glentis, Dawei Li, Chung-Yiu Yau, Mingyi Hong Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17787v1 It is widely believed that stochastic gradient descent (SGD) performs significantly worse than adaptive optimizers such as Adam in pre-training Large Language Models (LLMs). Yet the underlying reason for this gap remains unclear. In this work, we attribute a large part of the discrepancy to SGD's inability to sustain learning rates comparable to Adam's much larger effective learning rates. Through empirical and theoretical analysis of LLM pre-training dynamics, we identify that training is characterized by small gradient norms and large weight-to-gradient ratios, an effect that becomes more pronounced with larger batch sizes typical in pre-training, necessitating such large effective learning rates. However, we find that output-layer gradient magnitudes become highly uneven across token classes, and that large gradient spikes frequently occur during training. Together, these effects severely restrict the admissible learning rate of SGD. Guided by this understanding, we show that simple clipping mechanisms that stabilize SGD at large learning rates enable it to recover most of Adam's performance. In our large-scale experiments, the validation loss gap between large-learning-rate SGD and Adam shrinks from more than 50% to only about 3.5% when pre-training a 1B-parameter LLaMA model with a 1M-token batch size. --- ### Network Knowledge Prior Guided Learning for Data-Efficient Surface Defect Detection Authors: Hang-Cheng Dong, Guodong Liu, Dong Ye, Bingguo Liu Published: 2026-05-18 Categories: cs.CV arXiv: https://arxiv.org/abs/2605.17780v1 Deep learning-based methods have become the de facto standard for industrial defect detection. However, their data-hungry nature and inherent "black-box" characteristics often lead to performance bottlenecks and limited trustworthiness in real-world applications. To address these challenges, this paper proposes a novel knowledge-guided loss function that seamlessly integrates model interpretability into the training process without incurring any additional inference cost. Our method operates in two phases: first, a primary classification network is trained, and its explanations, in the form of saliency maps, are generated as prior knowledge. Second, a multi-task learning framework is established, where the main task performs classification, and an auxiliary task imposes consistency between the saliency maps of the final model and the primary model. This consistency is enforced by a dedicated knowledge-guided loss term, effectively acting as a powerful regularizer to steer the model towards robust feature representations. Extensive experiments on multiple public defect datasets demonstrate that our approach consistently enhances the performance of baseline models in terms of accuracy and AP. Moreover, visual analysis reveals that the proposed method yields more concentrated and human-intelligible saliency maps. This work presents a simple yet effective paradigm for bridging the gap between model performance and interpretability, paving the way for more reliable and high-performing vision systems in industrial quality inspection. --- ### Learning Variable-Length Tokenization for Generative Recommendation Authors: Minhao Wang, Bowen Wu, Wei Zhang Published: 2026-05-18 Categories: cs.LG arXiv: https://arxiv.org/abs/2605.17779v2 Generative recommendation reformulates recommendation as next-token prediction over discrete semantic identifiers (IDs). A fundamental yet unexplored design choice is that existing methods employ fixed-length tokenization for all items, implicitly assuming uniform encoding capacity regardless of item characteristics. Through systematic experiments across four datasets, we discover the Popularity-Length Paradox: popular items achieve optimal performance with short IDs, while tail items require substantially longer codes to capture discriminative semantics. This reveals a critical mismatch where popular items benefit from abundant collaborative signals and require minimal semantic detail, whereas tail items must rely on fine-grained content features due to sparse interaction data. To address this, we propose VarLenRec, a framework for learning variable-length tokenization. We develop Popularity-Weighted Information Budget Allocation (PIBA), an information-theoretic framework proving that optimal ID length should scale as a negative power of popularity. Directly implementing variable-length allocation faces two technical challenges: standard Euclidean residual quantization lacks geometric capacity to support diverse code lengths without distortion, and discrete length decisions are non-differentiable. We address these through Hyperbolic Residual Quantization, which leverages the exponential volume growth of the Poincaré ball to naturally stratify encoding capacity, and a Soft Length Controller, which enables differentiable length prediction via continuous layer retention probabilities regularized by PIBA-derived priors. Extensive experiments demonstrate that VarLenRec achieves significant improvements over state-of-the-art methods in recommendation accuracy and training/inference efficiency, revealing the importance of adaptive encoding capacity in generative recommendation. --- ### Self-Distillation is Optimal Among Spectral Shrinkage Estimators in Spiked Covariance Models Authors: Radu Lecoiu, Debarghya Mukherjee, Pragya Sur Published: 2026-05-18 Categories: math.ST, cs.LG, stat.ME, stat.ML arXiv: https://arxiv.org/abs/2605.17778v1 Self-distillation has emerged as a promising technique for improving model performance in modern machine learning systems. We develop the statistical foundations of self-distillation in spiked covariance models, by introducing and analyzing a broad class of estimators, namely spectral shrinkage estimators. We establish that for spiked covariance matrices with $s$ spikes, $s$-step self-distillation achieves optimal performance among spectral shrinkage estimators, outperforming well-known estimators in statistics and machine learning. Moreover, we show that $s$ steps are necessary for optimality: any $(s-k)$-step distilled estimator is strictly suboptimal for $1 \leq k \leq s$. For the special subclass of isotropic covariances, we show that optimally tuned Ridge regression performs best among spectral shrinkage estimators. We also study a federated approach where multiple data centers share spectral shrinkage estimators and a common server seeks to aggregate them to achieve optimal performance. In this case, we find that the best local rule again takes the form of self-distillation, though it differs from the optimal rule when data are hosted centrally on a single server. Together, our results elucidate why self-distillation improves predictive performance and provide a broader statistical framework connecting it with classical shrinkage-based methods. ---