Computer Vision
Triangular Resampling for Long-Horizon Motion Generation
Kunhang Li, Yiyi Cai, Xiangyue Zhang et al. (+5 more)
We introduce Triangular Resampling (TR), a post-training method for mitigating long-horizon error accumulation in motion
diffusion models. Built on FloodDiffusion's triangular denoising schedule, TR addresses the mismatch between ground-truth-derived training windows and model-generated inference states. Replacing only completed motion history leaves this mismatch unresolved in partially denoised states within the active window. TR therefore extends rollout-based training to these states, using ground-truth clamping to limit excessive drift. For each replayed sample, TR draws one denoising threshold, shared across latent positions and replay updates, and replays multi-step triangular denoising without gradient tracking. After each update, states below the threshold are replaced with noise-matched ground truth, while those at or above it retain model predictions. The resulting latent window enters the standard training update. This rollout construction supports both supervised training (TR) and distribution matching (TR-DMD). On 120-second motion generation from HumanML3D test
prompts, TR and TR-DMD achieve state-of-the-art FID AUC within their respective non-DMD and DMD comparison groups. Supervised TR reduces FID AUC by 40.9% and FID degradation slope by 55.3% relative to matched post-training without replay.
Read the full abstract ▾
All authors Kunhang Li, Yiyi Cai, Xiangyue Zhang, Fangyuan Tu, Yuhan Wu, Zhixiang Wang, Kaipeng Zhang, Haiyang Liu
Robotics
Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control
Yisheng Zhang, Tao Wang and Sicun Gao
Reward shaping is fundamental to modern robotic control with deep reinforcement learning (RL), yet practitioners still rely heavily on heuristic principles borrowed from classical optimal control and trajectory optimization. Existing methods rarely distinguish reward terms that are intrinsic to the control objective from numerical regularizers, leading to brittle hyperparameter tuning. To determine which quantities a reward must contain, we study the stabilization control problem with a focus on zeroth-order (configuration) and first-order (velocity) information. We theoretically and empirically demonstrate that policy gradient methods can successfully solve stabilization tasks without first-order reward terms, adding such terms can instead introduce severe sensitivity as their scale grows. Conversely, our findings confirm that reward functions must be zeroth-order complete over goal-relevant coordinates, while the first-order state remains necessary in the policy observation under our low-dissipation assumptions. Overall, these results provide actionable and principled guidance for reward design in robotic RL.
Read the full abstract ▾
Machine Learning
Predicting Delayed Train Trajectories on the Dutch Railway Network: Explainable AI Evaluation of Topological, Operational and Weather Features with Tree Based Ensemble Methods
Jia Long Bao, Ali Mohammed Mansoor Alsahag and Seyed Sahand Mohammadi Ziabari
The reliable prediction of passenger train delays is a critical component of railway management. While contemporary research frequently attempts to maximize absolute accuracy by deploying opaque deep learning architectures, the underlying data mechanics driving longitudinal predictive decay remain underexplored. Consequently, this study provides an explainable temporal robustness analysis of network-wide railway delay prediction. Focusing on the Dutch railway network, this research utilizes interpretable tree-based ensembles to integrate granular topological, environmental, and operational features. The overarching finding establishes that while feature-rich tree-based models improve simultaneous (within-month) prediction, predictive performance systematically degrades when evaluated across non-simultaneous (future) months. Furthermore, multi-horizon SHAP and dispersion analyses explicitly link this degradation to environmental feature
volatility and instability within the statistical target definition. Ultimately, this study demonstrates that richer feature sets alone are insufficient to resolve long-term
forecasting constraints, underscoring the necessity to transition toward dynamic, season-aware architectures anchored by absolute operational boundaries.
Read the full abstract ▾
Computation and Language
Using LLMs to Detect LLM-Generated Texts: A Cross-Generation Analysis
Haiyue Yuan, Jie Guo, Weidong Qiu et al. (+3 more)
Automated detection of LLM-generated texts (LGTs) is critical, yet dedicated detectors often struggle to generalize across domains and models. While general-purpose LLMs offer flexible zero-shot authorship classification with explanatory rationale, their detection behavior, especially regarding self-detection versus cross-detection across model generations, remains poorly understood. We systematically evaluate 15 LLMs spanning three model generations as both generators and detectors. Using a
benchmark of 1,000 human-written texts and 15,000 LGTs (1,000 per model), we collected over 233,000 binary classifications alongside natural-
language explanations. Our results reveal that detection efficacy is primarily driven by detector capability rather than generator provenance, although outputs from newer generators remain notably harder to detect. Crucially, statistical comparisons show no systematic advantage or disadvantage for self-detection across models. Error analysis further exposes generational bias shifts: first-generation detectors under-detect LGTs (high false-negative rates), second-generation detectors over-flag human texts (high false-positive rates), and the latest models achieve balanced trade-offs. Finally, we highlight significant inconsistencies in how different LLMs apply textual cues to justify their decisions. Code: https://github.com/hyyuan/detect-llm-generated-texts.
Read the full abstract ▾
All authors Haiyue Yuan, Jie Guo, Weidong Qiu, Zheng Huang, Ruizhe Li, Shujun Li
Computer Vision
Unified Trajectory Matching Policy Optimization: Diverse T2I Generation and VLA Generalization
Zhiyuan Ma, Jiaming Li, Lingzhen Li et al. (+7 more)
Reward-maximizing reinforcement learning (RL) is widely used to post-train
stochastic diffusion and flow policies for text-to-image (T2I) generation. However, reward-maximizing RL causes policy mode collapse even under reference KL or
entropy regularization, reducing the policy to a single high-reward mode. In T2I, this produces similar images and reward hacking. When extended to
vision-
language-action (VLA) models, the same collapse removes alternative successful strategies and weakens task and scene generalization. To address this limitation, we introduce Unified Trajectory Matching Policy Optimization (Uni-TMPO), a unified RL post-training framework for
diffusion and flow policies. First, Uni-TMPO converts standardized rewards into a target distribution within each trajectory group and derives the policy distribution from trajectory log probabilities. Then, forward Kullback-Leibler optimization matches the two distributions instead of maximizing expected reward. A progress-conditioned coarse-to-fine scheduler efficiently constructs T2I trajectories. Within the unified framework, feedback-conditioned sampling uses updated observations to construct VLA trajectories. Extensive experiments show that Uni-TMPO achieves higher T2I rewards and VLA ID success rates than the strongest baselines. More importantly, it achieves the best T2I reward-diversity-efficiency trade-off and VLA generalization to held-out tasks and scenes, while real-
robot evaluation demonstrates the value of multiple action strategies when the higher-reward target is blocked.
Read the full abstract ▾
All authors Zhiyuan Ma, Jiaming Li, Lingzhen Li, Yu Liu, Xuekai Zhu, Dingkang Liang, Kaiyan Zhang, Jianjun Li, Bowen Zhou, Xiang Bai
Artificial Intelligence
VCN-Bench: A Video-Contextualized Navigation Benchmark for Spatial Reasoning over Prior Visual Experience
Siqi Zhang, Meng Wei, Chenyang Wan et al. (+6 more)
Spatial
reasoning is fundamental to
embodied agents, yet it remains unclear whether spatial understanding can be carried forward to guide sequential interactions. Existing spatial-
reasoning benchmarks typically terminate at offline predictions, while
navigation benchmarks evaluate spatial
reasoning as part of instruction following and exploration. We introduce VCN-Bench, a
Video-Contextualized
Navigation benchmark for probing closed-loop spatial
reasoning over prior visual experience in MLLMs. Given a prior
video covering both the initial location and destination, the
agent is tasked with
reasoning out the instruction-specified target and navigating toward it with the inferred spatial context. Built on Matterport3D, VCN-Bench contains five instruction types, 100k training episodes, and 1,250 evaluation episodes.
Navigation serves as the primary evaluation, while diagnostic goal identification helps distinguish destination-resolution errors from subsequent
navigation failures. We further propose MV-DualVLN, a planning-oriented baseline that jointly leverages prior
video and in-episode observations. Experiments reveal limited
navigation performance, a substantial destination-resolution-to-
navigation gap, and frequent
navigation failures even after correct destination identification.
Read the full abstract ▾
All authors Siqi Zhang, Meng Wei, Chenyang Wan, Shaohao Zhu, Shufan Shen, Xihui Liu, Zhihua Wei, Tai Wang, Jiangmiao Pang
Artificial Intelligence
Jailbreak Context Lingers: Divergent Safety Routing and Its Cross-Task Predictability in Tool Agents
Xi Wang, Songlei Jian, Yiming Zhang et al. (+5 more)
As large
language models increasingly operate as tool-using
agents, post-jailbreak safety feedback is often assumed to serve as a reliable safeguard; however, how lingering jailbreak context shapes subsequent
agent behavior remains largely unexplored. To systematically examine this dynamic, we introduce a paired continuation framework across 192 parent tasks spanning 42 domains, evaluating 12,148 analyzed continuation pairs (curated from a 12,288-pair initially design) across eight diverse
agents. We find that identical safety feedback induces sharply model-dependent behavioral routing rather than uniform protection: redirecting unsafe trajectories toward legitimate completion (rescue), sustaining unauthorized execution (persistent unsafe), or triggering over-refusal on benign tasks (collateral loss). Through layer-wise activation patching, we discover a shared late-commit pattern where causal intervention effects surge sharply near the final layers (relative depths of 0.958–0.984) despite an over 30-fold variation in peak magnitude across architectures. Crucially, critical-layer representations correlate with macroscopic routing outcomes, and intervening at these layers causally alters concrete next-step tool actions. Building on this causal foundation, we test whether localized intervention-derived features can serve as predictive proxies for full-trajectory routing outcomes on unseen parent tasks under leave-one-parent-task-out evaluation, finding that they provide viable predictive signals in responsive
agents with peak ROC AUCs reaching 0.675 for rescue, 0.777 for collateral loss, and 0.702 for persistent unsafe. These findings establish a mechanistic lens and a predictive baseline for anticipating the safety and utility trade-offs of post-jailbreak feedback in autonomous
agents.
Read the full abstract ▾
All authors Xi Wang, Songlei Jian, Yiming Zhang, Bin Ji, Zhaoye Li, Ma Jun, Baosheng Wang, Jie Yu
Robotics
Natural State-Prediction Accuracy can Hide Weak Controlled Responsiveness in VLA Readouts
Hyungjoon Kim, Wonbin Son, Mi Young Lee et al. (+2 more)
Accurately decoding object states from the internal representations of
vision-
language-action (VLA) models does not establish that the predictions respond faithfully to changes in the target physical state. In natural observations, object state,
robot configuration, occlusion, and task progress vary together, allowing contextual cues to contribute to prediction. In this paper, we introduce an evaluation framework that separates prediction accuracy, target-state responsiveness, and context stability using physically validated observations that cross target coordinates with
robot contexts. We demonstrate that high natural-trajectory accuracy can coexist with weak controlled target-state responsiveness in fixed representation-readout pairs. Comparisons and interventions involving representations, readouts, and training data show that the three properties provide distinct diagnostic information. Furthermore, adding responsiveness and context sensitivity to a failure predictor based on initial state error and physical variables reduces policy-failure prediction error on new initializations relative to the specified baseline while same-observation controlled MAE is also informative. These findings motivate evaluating target-state responsiveness and context stability alongside natural prediction accuracy, and examining their relationship to actual policy behavior and task outcomes.
Read the full abstract ▾
All authors Hyungjoon Kim, Wonbin Son, Mi Young Lee, Jun Young Lee, Seungmin Rho
Machine Learning
AgentPerfBench: A Benchmarking and Evaluation Suite for Inference Performance of Agentic LLMs
Cheuk Hang Lau, Zeyu Cao, Kevin Wong Cheuk Yin et al. (+6 more)
The optimization of LLM serving engines, such as vLLM and SGLang, is largely
benchmark-driven: optimizations, scheduling policies, hardware and system designs are all selected based on representative workloads. However, a significant mismatch has emerged in the agentic era. Existing
benchmarks primarily focus on simple single-turn chatbot workloads. LLM applications are increasingly agentic: coding
agents, terminal execution systems, and tool-use
agents issue multi-turn requests with growing context lengths. We introduce AgentPerfBench, a
benchmark suite for agentic inference. It uses real traces from agentic
benchmarks, such as SWE-Bench and TerminalBench, alongside standard chat baselines. This enables benchmarking of models on multi-turn tasks involving tool calling, skill utilization, and increasing context lengths. AgentPerfBench also samples from empirical distributions of input length, output length, and turn count derived from the real traces, generating representative synthetic profiles for cheap and accurate measurements on new hardware. In addition, we further find that several existing
benchmarks fail to accurately reflect real hardware performance for two key reasons: 1) they do not account for realistic context-length growth, and 2) they measure inference performance without operating at hardware saturation. We discuss these issues in detail and provide rich kernel-level Nsight Compute (NCU) traces to construct a new multi-dimensional roofline model that captures hardware-system limitations in both
memory bandwidth and
memory capacity footprint. The benchmarking suite then includes automated scripts to identify potential bottleneck conditions on emerging hardware when evaluated with diverse agentic traces. Together, these contributions quantify the chat-to-agentic gap in current inference
benchmarks and characterise per-kernel GPU resource utilisation via roofline analysis.
Read the full abstract ▾
All authors Cheuk Hang Lau, Zeyu Cao, Kevin Wong Cheuk Yin, Yao Lai, Haoran Wu, Nicholas D. Lane, Robert D. Mullins, Ilia Shumailov, Yiren Zhao
Computer Vision
V-Gym: Enhancing Agentic Visual Reasoning via Skill-Data Co-Evolution
Bei Yan, Yuecong Min, Jie Zhang et al. (+3 more)
Advances in
multimodal understanding,
reasoning, and tool use enable
agents to tackle increasingly complex visual
reasoning tasks. By distilling past execution experience into reusable skills,
agents can transfer lessons from both successes and failures into future
reasoning, reducing repeated errors and improving capabilities. However, limited experience may produce unreliable, poorly generalizable skills, while static
datasets may lack the targeted and diverse practice needed for refinement. To address this gap, we introduce V-Gym, an autonomous framework that iteratively co-evolves procedural skills and
multimodal practice data from execution trajectories. During skill
evolution, V-Gym analyzes trajectories to distill and refine hierarchical skills, updating procedural guidance and applicability conditions while retaining an update only if it improves validation performance. During data
evolution, V-Gym selects generation seeds by balancing data utility and exploration, then translates trajectory-identified bottlenecks into diverse, targeted practice data that expand the data bank after quality checks. The resulting practice outcomes feed back into subsequent skill updates, closing the loop for continual skill refinement. Experiments across diverse
multimodal reasoning benchmarks show substantial improvements over baselines with multiple backbone models. Its evolved skills generalize across domains and models, while evolved data support more effective skill refinement, enabling autonomous diagnosis, targeted practice, and continual self-improvement.
Read the full abstract ▾
All authors Bei Yan, Yuecong Min, Jie Zhang, Junqi Yang, Shiguang Shan, Xilin Chen
Machine Learning
SOLAR: A State-Driven Online Learning Rate Scheduler for LLM Pretraining
Qiulin Shang, Binyu Wang, Yongqi Qiao et al. (+3 more)
Learning-rate (LR) scheduling plays a central role in large
language model (LLM) pretraining, yet current practice still relies heavily on hand-crafted heuristics such as Warmup-Cosine-Decay and Warmup-Stable-Decay. Because these schedules are fixed in advance, they cannot adapt to evolving optimization dynamics. Online learned scheduling within the Learning to Optimize (L2O) framework offers a dynamic alternative, but remains brittle at LLM scale due to noisy signals, delayed feedback, and the risk of catastrophic divergence. We propose
SOLAR (State-driven Online Learning rAte scheduleR), a stabilized framework for reliable online LR adaptation.
SOLAR uses a base schedule as a reference and learns bounded, state-dependent residual corrections for individual parameter groups. Each correction re-anchors to the base at every step, allowing the policy to adapt the LR without relearning the warmup-decay profile. A lightweight state representation and progress-aware reward guide online learning, while a
Circuit-Breaker restores training after rare unsafe actions. Across autoregressive
language-model pretraining,
SOLAR improves final perplexity over tuned static schedules and automatic LR tuners for dense models from 60M to 1B, AdamW and
Muon, and two MoE settings up to 3B. Matched 130M controls show that adding base anchoring and action bounds improves a global PPO controller from 27.09 to 23.74 final PPL, while group-wise control reaches 22.87 on the same two seeds. A residual policy trained on a 60M proxy can also be frozen and reused at larger dense scales without target PPO updates, remaining effective across a fourfold base-LR range. These results establish
SOLAR as a practical learned LR controller for LLM pretraining.
Read the full abstract ▾
All authors Qiulin Shang, Binyu Wang, Yongqi Qiao, Songde Rao, Zhoutong Wu, Kun Yuan
Machine Learning
QuantForge: Discovering Residual Decompositions for MXFP4 Post-Training Quantization
Qiulin Shang, Zhoutong Wu, Jie Hu et al. (+1 more)
Four-bit post-training quantization can reduce the
memory demands of large
language models, but preserving accuracy under strict MXFP4 W4A4 requires coordinating several design choices. Coordinate transforms change block-encoding errors, which in turn affect the residuals propagated through the network. The useful algorithmic decomposition is therefore not fully known before search. LLM-driven program
evolution offers a way to explore these choices, but performance scores alone do not explain which design should change next. We introduce QuantForge, a PTQ discovery system that records competing explanations, selects controls that distinguish them, and checks that successor code implements the resulting conclusions. This residual compilation guides program revisions while retaining useful programs even when their original explanations are rejected. Remeasuring the revised program reveals the next error to address. This process discovers HiRes, a fixed MXFP4 quantizer that shapes coordinates, refines legal code assignments, and recovers errors along
attention and MLP paths. Each stage acts on residuals measured after the preceding stage has executed. Across seven tasks, HiRes achieves the lowest seven-model Robust Fit (0.09300) and the lowest quantized Fit-7 at 32B. In matched-budget comparisons of LLM-driven program
evolution, each with 240 evaluator calls, QuantForge reaches a held-out transfer target in six of eight runs, compared with three each for textual
memory and reflection
memory, and one for score-only
evolution, despite evaluating fewer new programs. These results show that QuantForge improves the discovery of transferable PTQ
algorithms by turning controlled evidence into subsequent program changes. We have open-sourced the QuantForge search framework.
Read the full abstract ▾
All authors Qiulin Shang, Zhoutong Wu, Jie Hu, Kun Yuan
Machine Learning
From Preference to Reciprocity: Decentralized Matching with Empirically Grounded LLM-agent Based Modeling
Wangxuan Fan, Xiaoyu Nie, Zhoutian Shi et al. (+5 more)
Bipartite matching is a fundamental problem in game theory and
market design. Classical approaches such as Gale–Shapley assume complete preferences and centralized computation, whereas many real-world matching processes are decentralized, asynchronous, and shaped by sequential interaction under limited information. We propose a dynamic bipartite matching framework that combines large
language model (LLM)
agents with contextual bandits. In a simulated Chinese marriage
market, economically grounded LLM
agents evaluate locally encountered candidates, while
agent-specific Logistic-UCB models learn reciprocal acceptance from realized proposal outcomes. The mechanism therefore separates two decisions—whom do I like? and who is likely to like me back?—without requiring ex ante
market-wide preference rankings. We first validate LLM-induced mate preferences against the empirical conditional-logit reference across multiple LLM backbones. In the $50\times50$ matching experiment, Bandit-UCB achieves the highest mean mutual welfare (56.01 versus 54.87 for Gale–Shapley), a smaller gender rank gap than the classical baselines, and the fewest blocking pairs among the LLM-ABM policies. Learned acceptance models show economically interpretable gender-differentiated associations, while counterfactual setups reveal no systematic unilateral advantage from prior search knowledge. Overall, these results support the advantages of decentralized matching with LLM-based behavioral modeling and online learning under incomplete information for economic simulation and computational social science research.
Read the full abstract ▾
All authors Wangxuan Fan, Xiaoyu Nie, Zhoutian Shi, Xiangcheng Meng, Shipei Zeng, Pin Gao, Yan Hu, Zhongxiang Dai
Computation and Language
Fair Fact-Checking: Closing the Cross-Lingual Gap in LLM Factual Judgement with RoSh
Muhammad Ahmad, Fatemeh Seyedin, Adrian Weller et al. (+2 more)
Misinformation on social media remains a critical problem, and more and more people settle it by asking a
language model instead of a fact checker. Whether models judge such claims reliably is debated; whether they judge them equally well in every
language people ask in has gone almost unasked. We test eight models from five families, 3B to 70B, on 1,500 encyclopedic factual claims that exist in identical form in eight
languages. English is judged better than every other
language on every model, and the gap is widest on the smallest ones, where Llama-3B on Arabic is no better than guessing. Existing remedies retrain on more
multilingual data or fit an unconstrained map between
language representations, and neither asks whether the model already holds the answer and simply fails to say it. It largely does: a linear probe recovers the truth from the very activations the model fails to express. We propose RoSh, a per-
language shift and rotation of the residual stream, computed in closed form at three layers, with no training and no weight modified. It improves every model and closes 75% of the gap on average, helping most where the model was worst: Arabic on Llama-3B goes from chance to nearly the English level, and a fifth fewer of the claims answered correctly in English are lost in translation. What remains is no longer a read-out failure: afterwards the head recovers as much of what is encoded outside English as it does in English. An unconstrained map fitted on the same pairs falls below the untouched baseline, so the orthogonality constraint is doing the work, and every model clears a scrambled-correspondence control and ten further controls. On the two
benchmarks of the closest inference-time method, latent-space intervention, run with its own data and metric code, RoSh's gains are five to thirteen times larger.
Read the full abstract ▾
All authors Muhammad Ahmad, Fatemeh Seyedin, Adrian Weller, Dongwon Lee, Mahmoudreza Babaei
Machine Learning
Learning What to Recall: Adaptive Multi-Cue Episodic Memory for World Models
Beomsu Kim, Chieh-Hsin Lai, Bac Nguyen et al. (+3 more)
World models predict future observations from current experience and actions, yet prediction can depend on observations seen far in the past. Episodic
memory preserves past observations for later recall; however, as
memory accumulates, it raises a fundamental question: which
memories are useful for the current prediction, and which available
retrieval cues should be trusted to find them? This is challenging because fixed criteria based on recency, pose overlap, or visual similarity can be unreliable across environments and queries. We propose Future-Aware Recall (FAR), a framework that learns episodic recall from future-aware predictive supervision and adaptive multi-cue scoring. During training, FAR measures predictive utility by the conditional log-likelihood of the realized future given recalled context, approximated by negative
diffusion prediction loss, and uses it to train a retriever that remains future-blind at inference. The retriever learns cue-specific relevance and automatically determines which available
retrieval cues, such as time, pose,
vision, and audio, to trust for each query when selecting
memories. Across three complementary settings, FAR outperforms hand-designed recall even with the same
retrieval cues, automatically adapts which available cues to trust, and recalls the right history as the world changes. Together, these results establish FAR as a flexible, principled approach to episodic
memory access in world models.
Read the full abstract ▾
All authors Beomsu Kim, Chieh-Hsin Lai, Bac Nguyen, Amir Bar, Jong Chul Ye, Yuki Mitsufuji
Robotics
HOI-Retarget: Contact-Centric Retargeting for Human-Object Interaction
Jihwan Shin, Adrià López Escoriza, Junzhe He et al. (+2 more)
Learning from demonstration (LfD) has enabled
humanoid robots to acquire diverse whole-body skills, but extending this paradigm to human-object interaction (HOI) is limited by the availability of
robot-compatible interaction references. We present HOI-Retarget, a contact-centric retargeting method that transfers HOI onto a
humanoid robot for large-scale motion-data generation. Its windowed trajectory optimization uses every labeled contact as a target in the object frame, balancing body tracking, foot support and smoothness under the
robot's kinematic limits. The method can augment a single demonstration across object sizes, absorb contacts reconstructed from
monocular video, and extend to several
robots manipulating one object. We publicly release the code and the retargeted motion
dataset.
Read the full abstract ▾
All authors Jihwan Shin, Adrià López Escoriza, Junzhe He, Matthias Heyrman, Marco Hutter
Machine Learning
FestDPO: Few-step Generator Alignment with Direct Preference Optimization
Jaewoo Lee, Kyuil Sim, Hyeongyu Kang et al. (+3 more)
Few-step generative models can generate high-fidelity samples within a few function evaluations. Despite this efficiency, generated samples may not exhibit desirable properties. When these properties are difficult to encode as an explicit reward function, direct preference optimization (DPO) can align generative models using pairwise preference feedback without training a separate reward model. However, extending DPO to few-step generative models is challenging because few-step generative models are generally implicit, making the likelihood evaluation required by DPO intractable. To address this challenge, we introduce Few-step DPO (FestDPO), an extension of DPO for few-step generative models that leverages nonparametric likelihood estimation from empirical samples. By exploiting the fast sampling capabilities of few-step generative models, our approach makes sample-based approximation of DPO loss computationally feasible. Furthermore, the sample-based formulation makes FestDPO agnostic to the model family and sampling procedure. Our toy experiment demonstrates that FestDPO matches the reward-tilted target distribution across four few-step generators. For real-world tasks, we evaluate FestDPO in two domains: text-to-image generation and
protein backbone generation. In text-to-image generation, FestDPO outperforms preference optimization baselines in both win rates against the base models and human evaluation scores. In
protein backbone generation, it achieves a higher $\beta$-sheet fraction and better structural designability than the baselines.
Read the full abstract ▾
All authors Jaewoo Lee, Kyuil Sim, Hyeongyu Kang, Kanghoon Lee, Woocheol Shin, Jinkyoo Park
Computer Vision
Evidence-Aligned Multimodal On-Policy Self-Distillation for Fine-Grained Visual Understanding
Nanxing Hu, Qiwei Yan, Jinchao Zhang et al. (+1 more)
Fine-grained visual understanding requires models to recognize small details within complex images.
Multimodal on-policy self-
distillation (OPSD) addresses this challenge by using a teacher conditioned on evidence-centered crops to supervise a student conditioned on original images along student-generated trajectories. Ideally, teacher corrections, the distributional changes from the student toward the privileged teacher, should be driven by task-relevant visual evidence. However, the designs that make the teacher effective also introduce other interference. Using a lagged or frozen teacher improves training stability but introduces a model-state gap from the evolving student, while cropping enhances task-relevant evidence but also loses the visual context. These two sources of interference make the teacher corrections not purely rely on the visual evidence. We introduce Evidence-Aligned
multimodal on-policy self-
Distillation (EAD), which retains the crop-conditioned teacher as the target but constructs a separate evidence reference for weighting the corrections. To exclude the effect of lagged model-state from this reference, EAD measures prediction changes using the current student. To avoid crop-induced context changes, EAD masks the evidence region in the original image while preserving the other visual context. The change from the student's masked-image prediction to its original-image prediction provides a controlled reference for the direction in which the visual evidence shifts the student's prediction. EAD weights each teacher correction by its cosine
alignment with the reference, i.e., retaining aligned corrections and downweighting the rest. Retaining only 6% of the supervision mass of dense OPSD, EAD consistently outperforms previous state-of-the-art methods.
Read the full abstract ▾
All authors Nanxing Hu, Qiwei Yan, Jinchao Zhang, Guoliang Kang
Machine Learning · Cross-listed from eess.SP
Hierarchical Clustering and Signal Denoising on Digraphs
Yi Wang, Sippanon Kitimoon, Hrushikesh N. Mhaskar et al. (+1 more)
In this paper, we propose a representation of a digraph (directed graph) as a Hermitian
matrix derived from its adjacency
matrix. This representation characterizes both the connectivity and the edge orientation of the digraph. Based on the spectral decomposition of the Hermitian
matrix, a digraph clustering
algorithm with $k$-means is introduced to produce a partition on the graph. Applying this
algorithm (bottom-up) recursively to a digraph with partially labeled vertices yields a spectral hierarchical digraph clustering (\myproj)
algorithm that produces consistent nested partitions of the digraph, or equivalently, a tree structure. Furthermore, based on the in-degree and out-degree of each cluster in the digraph clustering, a pair of hierarchical interval partitions (filtrations) can be derived in a top-down manner to produce a pair of nested
knot sequences. These
knot sequences facilitate the construction of multilevel spline quasi-interpolants, enabling a noisy graph signal to be decomposed into a coarse approximation and inter-level details, followed by adaptive thresholding and reconstruction. Experiments on synthetic and real-world digraphs demonstrate the superiority of our {\myproj}
algorithm for digraph clustering across diverse graph structural properties (homophily and heterophily) and supervision settings. Moreover, experiments on digraph signal processing using multilevel spline quasi-interpolants further demonstrate the effectiveness of signal recovery on digraphs in terms of RMSE and SNR.
Read the full abstract ▾
All authors Yi Wang, Sippanon Kitimoon, Hrushikesh N. Mhaskar, Xiaosheng Zhuang
Numerical Analysis
Fully Discrete Active Flux Methods for the Euler Equations: Exploring Different Reconstructions and New Multi-Dimensional Evolution Operators
Christiane Helzel, Mária Lukáčová-Medvid'ová, Amelie Porfetye et al. (+1 more)
We present new fully discrete Cartesian grid Active Flux methods for the two-dimensional Euler equations of gas dynamics. A key component of these methods is the
evolution operator used to evolve point-value degrees of freedom. Therefore, we derive several new, third-order accurate, multidimensional
evolution operators for point-value updates of the linearized Euler equations using the method of bicharacteristics. We also investigate how the reconstruction and choice of approximate
evolution operator affect the accuracy and stability of the resulting methods.
Read the full abstract ▾
All authors Christiane Helzel, Mária Lukáčová-Medvid'ová, Amelie Porfetye, Zhuyan Tang
Information Theory
Covert Communications over Enormous Fluid Antenna Systems (E-FAS): Impact of Channel Hardening on Detection
Farshad Rostami Ghadi, Damoon Shahbaztabar, Kai-Kit Wong et al. (+1 more)
Enormous fluid
antenna systems (E-FAS) enable guided surface wave (SW) propagation and induce an end-to-end channel that is conditionally complex Gaussian with a random covariance scale. This paper investigates the impact of this channel structure on covert communication in the presence of a passive warden. Under an equal power effective mode representation, the random channel scale follows a Gamma distribution, leading to a Bessel-K distribution for the effective channel power. We show that the optimal likelihood ratio test at the warden reduces to an energy detector and derive the corresponding false alarm and missed detection probabilities. These results are then used to characterize the achievable covert throughput under a prescribed detection error constraint. Numerical results show that, for finite observation intervals, finite mode E-FAS channels improve covertness relative to the fully hardened Rayleigh limit with the same average channel power. They further reveal an asymmetric role of channel hardening: stronger hardening can improve the legitimate link, whereas retaining channel scale fluctuations at the warden can enhance finite observation covertness.
Read the full abstract ▾
All authors Farshad Rostami Ghadi, Damoon Shahbaztabar, Kai-Kit Wong, Wei-Ping Zhu
Machine Learning · Cross-listed from stat.ML
Two-Timescale Fine-tuning Provably Learns New Features for Two-Layer ReLU Networks
Etienne Boursier and Nicolas Flammarion
Fine-tuning pre-trained models on specialized tasks with scarce data is central to modern deep learning. Despite its empirical success, theoretical understanding of fine-tuning remains limited. We introduce a Gaussian multi-index setting to study fine-tuning from pre-trained weights, where the teacher network has $m+1$ features, $m$ of which are learned during pre-training and one of which must be learned during fine-tuning. For two-layer ReLU networks, we show that two-timescale training, i.e., updating the outer weights infinitely faster than the hidden ones, learns the new task-specific feature while preserving the pre-trained ones in the model representation. Moreover, only $\mathcal{O}(d)$ fine-tuning samples are required for this recovery, independently of the number of pre-trained features. In contrast, with random initialization, the same number of samples is insufficient to recover the target parameters. Our results therefore demonstrate that pre-training can induce an implicit bias with a clear statistical advantage over random initialization, enabling feature learning from scarce fine-tuning data.
Read the full abstract ▾
Robotics
On the Numerical Reliability of Differentiable Physics-Based Optimization for Robotic Material Manipulation
Xintong Yang, Minglun Wei, Yu-Kun Lai et al. (+1 more)
Differentiable physics is increasingly used in robotic material manipulation for system identification, trajectory or skill optimization, demonstration generation, and
robot or end-effector design. These applications depend on gradients propagated through long, contact-rich simulation rollouts. We study the numerical reliability of those gradients using two Material Point Method (MPM) system-identification
benchmarks derived from elastoplastic and granular manipulation. The
benchmarks provide controlled cases for three effects that also arise in broader differentiable physics-based optimization. GPU many-to-one sums whose order depends on thread scheduling changed long-horizon gradients and reversed the sign of one parameter gradient relative to a deterministic reference. Finite-difference checks became less reliable for longer rollouts because repeated-run loss variation grew much faster than the loss change produced by the tested parameter perturbations. Observation and loss definitions changed optimization behaviour and the solution preferred by an independent metric. These results motivate reproducible accumulation, finite-difference validation that compares perturbation-induced loss changes with repeated-run variation, and explicit reporting of objective construction when differentiable simulation is used for robotic optimization.
Read the full abstract ▾
All authors Xintong Yang, Minglun Wei, Yu-Kun Lai, Ze Ji
Game Theory
Settling the Complexity Landscape of Multi-Agent Contracts with Binary Actions
Michal Feldman, Maya Schlesinger and Shay Shani
We study the computational complexity of optimal contract design in the multi-
agent binary-action model, focusing on gross-substitutes reward functions and related classes. While additive rewards admit an FPTAS and general submodular rewards admit only constant-factor approximation, the complexity within the intermediate class of gross substitutes has remained largely open. We uncover a fine-grained approximation landscape within this class. We first show that the optimal contract problem is APX-complete even for OXS rewards - a strict subclass of gross substitutes rewards, ruling out a PTAS for gross substitutes unless $\mathsf{P}=\mathsf{NP}$. In contrast, for weighted matroid rank functions (WMRFs) - another natural subclass of gross substitutes - we obtain an EPTAS and show that no randomized FPTAS exists in general. We further identify the special case of partition (weighted) matroid rank functions, for which we obtain an FPTAS. This stands in contrast to the multi-
agent multi-action setting, where no PTAS exists even for unweighted partition matroid rank functions. Finally, we consider the broader class of ultra reward functions. While ultra rewards retain the tractability of gross substitutes in a related combinatorial contract model with a single
agent, we show a sharp contrast in the multi-
agent model: no
polynomial-time randomized
algorithm using value queries can achieve a $2^{o(n)}$-approximation in expectation. Together, our results reveal several qualitatively distinct computational regimes within and beyond gross substitutes, and identify submodularity as a crucial ingredient for the approximability of multi-
agent contracts.
Read the full abstract ▾
Performance
Tool Waiting and Re-arrival in Compile-Time-Static LLM Serving: Cost Mechanisms and Configuration Selection
Dongkyeom Jang, In-Nea Wang and Junho Jeong
In agentic LLM services, a session calls an external tool, waits for it, and re-arrives to continue inference. Statically compiled NPU serving can fix the batch bucket set, the maximum batch size, and the number of KV cache slots at compile time. We define such an environment as a compile-time-static serving substrate and analyze the execution-time cost that tool waiting and re-arrival incur in it. On a single LLM instance, we run synthetic workloads following a measured tool waiting time distribution and compare, on the same inputs, a baseline configuration with settings {1, 2, 4, 8}, 8, and 8 against configurations that change some of them. Because re-arrival times differ across configurations, we build a simulator that replays request processing in time order, select the candidate with the lowest predicted cost among 2,077 configurations, and validate it on new inputs. We identify three mechanisms: discrete batch
alignment, KV cache survival, and prefill interference. At a concurrency of 6, absent from the bucket set, tool waiting lowered the padding ratio (0.235 to 0.120) yet increased decode execution time 1.51-fold, so padding alone did not indicate cost. On new inputs at a concurrency of 8, where the baseline reused KV in 9 of 24 re-arrivals, enlarging the maximum batch size alone cut execution cost by 8.25%, and the selected configuration, which also adjusted the bucket set, by 9.72%. Where 17 of 18 re-arrivals were already reused, the effect was 0.59%. Compile-time configurations should thus be selected by diagnosing KV reuse loss and the resulting change in execution.
Read the full abstract ▾
Sound
Unsupervised Speech Enhancement via Drifting
Diego Caviedes-Nozal, Liang Xu, Rasmus Kongsgaard Olsson et al. (+1 more)
This paper addresses unsupervised
speech enhancement in the unpaired setting using drifting methods, where training relies on separate collections of degraded and clean audio without corresponding pairs. While recent drifting approaches enable unpaired training, they do so at a heavy cost: because the objective optimizes only a marginal prior over clean
speech, the enhancer gradually loses the input's linguistic content and speaker identity. To fix this, we introduce input-conditioned drifting. We preserve the pull of the clean corpus while re-tethering the output to the degraded input via two mechanisms: an anchor encoder supplies the missing likelihood by pulling toward the input's features, and a key encoder conditions the prior by re-weighting retrieved frames. Neither requires labels or paired data. Using a training-free encoder selection criterion, Word Error Rate on VoiceBank-DEMAND falls to 10.1% (unprocessed: 11.7%), speaker similarity recovers from 0.490 to 0.879, and the recipe transfers in part to dereverberation on WSJ0-REVERB: content improves, rendering quality does not.
Read the full abstract ▾
All authors Diego Caviedes-Nozal, Liang Xu, Rasmus Kongsgaard Olsson, W. Bastiaan Kleijn
Software Engineering
Codoku: Renewable Program-Reasoning Challenges for Frontier Coding Agents
Cong Li, Hao Sun, Zenan Li et al. (+1 more)
Existing program-
reasoning benchmarks ask large
language models to predict a program's behavior on a given input. Coding
agents break two assumptions on which these
benchmarks rest: an
agent can recover the answer by executing the program instead of
reasoning about it, and fixed task sets drawn from existing programs are increasingly exposed to contamination, yet costly to renew. We introduce Codoku (code sudoku), a renewable
benchmark in which a solver fills typed
cells in a partial program to satisfy global static and dynamic constraints, such as a prescribed control-flow graph and execution path. Because a partial program cannot be executed and valid fillings are sparse in an exponentially large space of interdependent choices, neither tool use nor enumeration can substitute for program
reasoning. Puzzles are synthesized from scratch via semantic reification, so fresh puzzles of controllable complexity can be generated on demand, each with a witness that guarantees solvability. We evaluate five frontier models on 300 puzzles through a coding
agent free to use any tool within a fixed budget. Small puzzles already challenge open-weight models, whereas even proprietary models solve only about half of the large ones. Codoku thus offers a renewable testbed for program
reasoning that can keep pace with rapidly improving coding
agents. GitHub: https://github.com/connglli/Codoku.
Read the full abstract ▾
All authors Cong Li, Hao Sun, Zenan Li, Zhendong Su
Computation and Language
Rewarding Novel Deductions: Solver-guided Process Supervision for Logical Reasoning
Muhammad Asif Ali, Wenqing Wang, Huan Wang et al. (+1 more)
Logical
reasoning remains a major challenge for large
language models (LLMs), particularly on structured problems that require precise constraint tracking, consistency preservation, and multi-step deduction. This challenge is especially acute for small-scale LLMs, which are more prone to producing inconsistent, redundant, or brittle
reasoning trajectories. Existing approaches for improving logical
reasoning largely optimize for final-answer correctness, providing only weak supervision over the intermediate
reasoning process. In this work, we propose SPRING: (Solver-guided Process Rewards for Novel LogIcal
ReasoNing Step Generation). SPRING uses SMT solver as a training-time verifier of intermediate
reasoning steps to provide process-level supervision. It introduces the notion of a novel
reasoning step, namely, a step that is logically valid, consistent with the evolving
reasoning state, and not already implied by previously accepted non-contradictory deductions. Based on this solver-based assessment, it designs process rewards that encourage novel inferential progress while penalizing contradictory and uninformative
reasoning steps. Evaluation across three logical
reasoning benchmarks, ZebraLogic, AR-LSAT, and Knights and Knaves, and four LLMs shows that SPRING consistently outperforms base LLMs, outcome-only reward baselines, and Logic-LM. On ZebraLogic, SPRING improves puzzle accuracy by up to 49.71 and 15.43 points over the base LLM and strongest outcome-only baseline, respectively. On AR-LSAT, it improves overall accuracy by up to 64.93 and 12.14 points, respectively. On Knights and Knaves, SPRING achieves up to 93.14 puzzle accuracy and 96.05 person accuracy.
Read the full abstract ▾
All authors Muhammad Asif Ali, Wenqing Wang, Huan Wang, Mohammad Raza
Computational Complexity · Cross-listed from quant-ph
Depth-Optimal Quantum Compilation
Francisca Vasconcelos
We achieve the first constant-depth
circuit for arbitrary single-
qubit gate synthesis. Unlike prior approaches, the construction is fully
unitary and requires no pre-supplied catalyst. For any constant $\delta>0$, it $\varepsilon$-approximates an arbitrary single-
qubit gate using $O(\log^{1+\delta}(1/\varepsilon))$ clean ancillae, Hadamard and $T$ single-
qubit gates, $O(\log(1/\varepsilon))$-width generalized Toffoli gates, and sublogarithmic-width Fan-Out gates. We further eliminate Fan-Out entirely, showing that Hadamard, $T$, and generalized Toffoli gates alone suffice for constant-depth synthesis. When restricted to the standard bounded-width gate model, our construction has depth $O(\log\log(1/\varepsilon))$, and we prove a matching $\Omega(\log\log(1/\varepsilon))$-depth lower bound. Overall, we establish that $\Theta(\log\log(1/\varepsilon))$-depth is unavoidable with only bounded-width gates, yet allowing even logarithmic-width multi-
qubit gates suffices to achieve constant-depth synthesis. These results also reveal new structure in shallow
quantum circuit complexity. We give a depth-preserving real simulation of bounded-error decision computation, showing that every depth-$d$ QAC
circuit can be simulated in depth $O(d)$ using only Hadamard, $X$, and generalized Toffoli gates. Thus arbitrary single-
qubit rotations and complex amplitudes do not increase the bounded-error decision power of QAC, even at constant depth. In particular, this reduces the long-standing
conjecture Parity$\notin$QAC$^0$ to proving a Parity lower bound against
circuits consisting only of Hadamard, $X$, and generalized Toffoli gates. More generally, this real normal form exposes a direct correspondence between the standard shallow-depth
quantum circuit hierarchy and a hierarchy of Forrelation
circuits with restricted oracle families.
Read the full abstract ▾
Computer Vision
CapField-OPD: Learning Continuous Capability Fields via Joint-Anchored Multi-Teacher On-Policy Distillation for Flow Models
Pengyang Ling, Jiazi Bu, Yujie Zhou et al. (+6 more)
Reward-specialized post-training produces strong experts for flow-based generative models, while multi-teacher on-policy
distillation (OPD) consolidates their capabilities into a single student. Existing methods, however, route each
prompt to a single teacher according to its semantic category, implicitly binding the desired capability to
prompt content. This coupling makes capability invocation vulnerable to
prompt perturbations and prevents users from explicitly adjusting the strength of the desired capability at inference time. In this work, we introduce CapField-OPD, an OPD framework that integrates multiple teachers into a continuous capability field through explicit capability coordinates. We use teacher models as anchors to construct this field, with the coordinates determining how their outputs are combined. Each capability configuration thus receives a unique supervision target, and capability control no longer depends on
prompt semantics. Since the training anchors may not be optimal at inference time, we further profile the learned field on a small calibration set. The coordinate with the highest mean reward serves as the recommended default, while coordinates that are frequently optimal offer a promising candidate set for test-time scaling. Extensive experiments on compositional generation, text rendering, and visual aesthetics demonstrate that CapField-OPD consolidates multiple specialized teachers into a single student while preserving or surpassing their performance, reliably invokes the desired capabilities under semantics-preserving
prompt variations, and supports continuous capability control and coordinate-based test-time scaling.
Read the full abstract ▾
All authors Pengyang Ling, Jiazi Bu, Yujie Zhou, Yibin Wang, Zeqiang Lai, Xiaoxiao Ma, Yi Jin, Huaian Chen, Yuhang Zang