论文检索

Deep Learning · Large Language Models

Jinwoo Kim, Taylor Berg-Kirkpatrick, Loris D'Antoni

Diffusion language models offer a promising alternative to autoregressive models due to their global, non-causal generation process, but their continuous latent dynamics make discrete constraints---e.g., the output should be a JSON file that matches a given schema---difficult to impose. We introduce a training-free guidance method for steering continuous diffusion language models to satisfy formal syntactic constraints expressed using regular expressions. Our approach constructs an analytic score estimating the probability that a latent state decodes to a valid string accepted by a given regular expression, and uses its gradient to guide sampling, _without_ training auxiliary classifiers. The denoising process targets the base model conditioned on syntactic validity. We implement our method in Diffinity on top of the PLAID diffusion model and evaluate it on 180 regular-expression constraints over JSON and natural-language benchmarks. Diffinity achieves 68-96\% constraint satisfaction while incurring only a small perplexity cost relative to unconstrained sampling, outperforming autoregressive constrained decoding in both constraint satisfaction and output quality.

Reinforcement Learning · Batch/Offline

Meng XU, Zhongying Chen, Weiwei Fu, Yan Li, Shuguang Wang, Jianping Wang

Offline Federated Deep Reinforcement Learning (FDRL) methods aggregate multiple client-side offline Deep Reinforcement Learning (DRL) models, each trained locally, to facilitate knowledge sharing while preserving privacy. Existing offline FDRL methods assign client weights during global aggregation using either simple averaging or Q-values, but they neglect the combined consideration of Q-values and policy inconsistency, the latter of which reflects the distributional discrepancy between the learned policy and the policy from offline data. This causes clients with no significant advantages in one aspect but obvious disadvantages in the other to disproportionately affect the global model, thereby degrading its capabilities in that aspect. During local training, clients in existing methods are compelled to fully adopt the global model, which negatively impacts clients when the global model is weak. To address these limitations, we propose a novel Federated Learning (FL) framework that can be seamlessly integrated into current offline FDRL approaches to improve their performance. Our method considers both policy inconsistency and Q-values to determine the weights of client models, with the latter adjusted by a scaling factor to avoid significant numerical discrepancies with the former. The aggregated global model is then distributed to clients to facilitate their learning from the global model. The impact of the global model on the local models is reduced when a client's model performance exceeds that of the global model, thereby mitigating the influence of a weaker global model. Experiments on the Datasets for Deep Data-Driven Reinforcement Learning (D4RL) demonstrate that our method enhances six state-of-the-art (SOTA) offline FDRL methods in terms of return and D4RL score.

Deep Learning · Large Language Models

Qi Liu, Xinhao Zheng, Renqiu Xia, Xingzhi Qi, Qinxiang Cao, Junchi Yan

Large language models (LLMs) have achieved remarkable progress in mathematical reasoning, yet persistently suffer from hallucinations and erroneous logic. While formal theorem proving (FTP) shows promise in process-level reliability, it is limited to _verification_ (checking known propositions). This leaves constructive problem-solving (finding unknown terms that satisfy specific conditions) underexplored and disconnected from process-level verifiability. To bridge this gap, we introduce **FPS** (_**F**ormal **P**roblem-**S**olving_), a principled framework to encompass the end-to-end problem-solving process in Lean 4. In FPS, the answer is an unknown metavariable coupled with a proof obligation, forcing it to be mathematically derived and verified. We further present **D-FPS** (_**D**eductive **FPS**_), which enforces a rigorous chain-of-thought structure, aligning formal derivation with human reasoning steps. To support this direction, we construct three benchmarks via the manual refactoring of over 1,000 problems: **FormalMath500**, **MiniF2F-Solving**, and **PutnamBench-Solving**. We further propose **RPE** (_**R**estricted **P**ropositional **E**quivalence_), a symbolic metric that evaluates semantic correctness beyond brittle string matching. Extensive experiments with state-of-the-art provers reveal that solving is significantly harder than proving, highlighting the ``alignment tax'' required to transition from loose validity checking to constructive, human-aligned reasoning.

Deep Learning · Large Language Models

Kongcheng Zhang, QI YAO, Shunyu Liu, Wenjian Zhang, Cen, Yang Zhou, Wenkai Fang, Yiru Zhao, Baisheng Lai, Mingli Song

Reinforcement Learning (RL) has shown promise for aligning Large Language Models (LLMs) to follow instructions with various constraints. Despite the encouraging results, RL improvement inevitably relies on sampling successful, high-quality responses; however, the initial model often struggles to generate responses that satisfy all constraints due to its limited capabilities, yielding sparse or indistinguishable rewards that impede learning. In this work, we propose ***H**indsight **i**nstruction **R**eplay* (HiR), a novel sample-efficient RL framework for complex instruction following tasks, which employs a *select*-then-*rewrite* strategy to *replay failed attempts as successes* based on the constraints that have been satisfied in hindsight. We perform RL on these replayed samples as well as the original ones, theoretically framing the objective as dual-preference learning at both the instruction- and response-level to enable efficient optimization using only a binary reward signal. Extensive experiments demonstrate that the proposed HiR yields promising results across different instruction following tasks, while requiring less computational budget. Our code and dataset are available at anonymous url.

Reinforcement Learning · Batch/Offline

Nazim Bendib, Nicolas Perrin-Gilbert, Olivier Sigaud

Offline zero-shot reinforcement learning (RL) aims to learn agents that optimize unseen reward functions without additional environment interaction. The standard approach to this problem trains task-conditioned policies by sampling task vectors that define linear reward functions over learned state representations. In most existing algorithms, these task vectors are randomly sampled, implicitly assuming this adequately captures the structure of the task space. We argue that doing so leads to suboptimal zero-shot generalization. To address this limitation, we propose extracting task vectors directly from the offline dataset and using them to define the task distribution used for policy training. We introduce a simple and general reward function extraction procedure that integrates into existing offline zero-shot RL algorithms. Across multiple benchmark environments and baselines, our approach improves zero-shot performance by an average of 20%, highlighting the importance of principled task sampling in offline zero-shot RL.

General Machine Learning · Evaluation

Archana Warrier, Dat Nguyen, Michelangelo Naim, Moksh Jain, Yichao Liang, Karen Schroeder, Cambridge Yang, Josh Tenenbaum, Sebastian Vollmer, Kevin Ellis 等

World models are central to building AI agents capable of flexible reasoning and planning. Yet current evaluations (i) test only properties measurable from observed interactions within an environment, such as next-frame prediction or task return, and (ii) do not test whether a learned model supports diverse queries about the same environment. In contrast, humans build \textit{general-purpose} models that can answer many different questions about an environment---including questions that require understanding global structure and counterfactual consequences. We propose *WorldTest*: a protocol for evaluating agents' ability to learn general-purpose models. A *WorldTest* benchmark pairs environments with multiple *environment-level queries*---properties of the full environment---rather than objectives defined only on observed trajectories. Individually, these queries can target global and counterfactual properties (e.g., reachability or the effects of interventions) that are not determined by any single rollout distribution. Collectively, they assess model generality across query types. We instantiate WorldTest as *AutumnBench*, a minimal yet expressive benchmark of 43 interactive grid-world environments and 129 tasks across three query families for both humans and learning agents. AutumnBench supports diverse environments and evaluations, including queries to evaluate prediction, counterfactual reasoning, and long-horizon planning. Experiments with 517 human participants and five frontier models show that humans substantially outperform these models, a gap we attribute to differences in exploration and belief updating. *WorldTest* and *AutumnBench* provide a rigorous framework for evaluating world-model learning and expose critical limitations in current approaches.

Theory · Reinforcement Learning and Planning

Safwan Labbi, Paul Mangold, Daniil Tiapkin, Eric Moulines

In this paper, we study the role of the critic in actor-critic for entropy-regularized, finite, discounted environments. We establish that, when the critic is exact, using the latter as a baseline is an actual variance-reduction method. In this case, actor-critic with stochastic gradients matches the sample complexity of deterministic policy gradient, reaching an $\epsilon$-optimal regularized value with $\tilde{O}(\log(1/\epsilon))$ samples. In practice, the critic is learned alongside the actor: the variance of the actor update is then influenced by the critic's variance and bias. Specifically, when the critic has a sufficiently small error, the variance reduction and rapid convergence are preserved. This suggests to learn the critic first, keeping it up to date after each actor update, underscoring the pivotal role of accurate critic estimation in actor-critic methods.

Applications · Robotics

Kunjal Panchal, Saayan Mitra, Sunav Choudhary, Victor Bursztyn, Somdeb Sarkhel, Hui Guan

LLM-based multi-agent embodied planning remains impractical due to prohibitively high execution latency. We identify failed actions as the dominant bottleneck, stemming from two core challenges: inaccurate state tracking under partial observability and inefficient coordination that produces redundant or conflicting actions. We introduce Mosaic, a runtime-efficient multi-agent planning framework that addresses both challenges. Mosaic maintains accurate yet lightweight state tracking through agent-centric semantic memory that stores objects in relative coordinates, enabling geometric transformations and coordination. It ensures efficient coordination through Integer Linear Programming that allocates actions at every planning step, enforcing physical feasibility and inter-agent coordination constraints. Across AI2-THOR and search-and-rescue benchmarks, Mosaic achieves 27–32% faster execution, 30–33% fewer LLM calls, 25–31% fewer steps, and 4–10% points higher success rates. These results demonstrate that efficient memory and constraint-guided coordination are critical for scalable, low-latency multi-agent planning.

General Machine Learning · Evaluation

Jinxiang Meng, Shaoping Huang, Fangyu Lei, Jingyu Guo, Haoxiang Liu, Jiahao Su, Sihan Wang, Yao Wang, Enrui Wang, Ye Yang 等

Real-world data visualization (DV) requires native environmental grounding, cross-platform evolution, and proactive intent alignment. Yet, existing benchmarks often suffer from code-sandbox confinement, single-language creation-only tasks, and assumption of perfect intent. To bridge these gaps, we introduce DV-World, a benchmark of 260 tasks designed to evaluate DV agents across real-world professional lifecycles. DV-World spans three domains: DV-Sheet for native spreadsheet manipulation including chart and dashboard creation as well as diagnostic repair; DV-Evolution for adapting and restructuring reference visual artifacts to fit new data across diverse programming paradigms and DV-Interact for proactive intent alignment with a user simulator that mimics real-world ambiguous requirements. Our hybrid evaluation framework integrates Table-value Alignment for numerical precision and MLLM-as-a-Judge with rubrics for semantic-visual assessment. Experiments reveal that state-of-the-art models achieve less than 50\% overall performance, exposing critical deficits in handling the complex challenges of real-world data visualization. DV-World provides a realistic testbed to steer development toward the versatile expertise required in enterprise workflows. Data and code are available at \url{https://anonymous.4open.science/r/DV-World-50D2}.

Deep Learning · Algorithms

Casey Mogilevsky, Kimberly Liang

Classical dynamic programming algorithms such as Smith-Waterman, edit distance, and constituency parsing solve structured combinatorial problems using hard constraints. Soft relaxations replace hard or operators with differentiable probabilistic models whose gradients correspond to alignment or parsing marginals, but existing approaches typically treat the DP algorithm itself as fixed, relying on hand-tuned gap penalties, edit costs, span penalties, and temperatures. We show how to learn these parameters directly from data: the marginals these algorithms produce are themselves first-order derivatives, so learning them is inherently a second-order problem. We derive efficient Hessian-vector products and cross-Jacobians for twelve dynamic programming algorithms spanning alignment, edit distance, and parsing; both derivative families admit closed-form covariance expressions under the induced Gibbs distribution. We implement these operators as fused CUDA kernels in dp, achieving -- speedups over standard PyTorch and making end-to-end parameter learning practical at modern scales. We then demonstrate that learning these parameters is critical in practice. In protein structure alignment, freezing gap penalties collapses performance from to (and to at lower encoder capacity), while jointly learning them recovers biologically meaningful gap regimes and reaches and lDDT, % of the TM-align lDDT ceiling. The same machinery transfers to constituency parsing, where a structured CKY CRF matches dense per-span supervision to within on English with no direct parse-tree structural supervision.

Janghwan Lee, Sihwa Lee, Jinseok Kim, Yongjik Kim, Jieun Lim, Jinwook Oh, Jungwook Choi

Large Reasoning Models (LRMs) achieve strong problem-solving through long chain-of-thought, but their deployment is constrained by the high cost of full-precision inference and growing KV cache footprints. Microscaled FP4 formats enable efficient FP4 deployment; however, fully quantizing weights, activations, and KV caches (W4A4KV4) causes severe reasoning degradation that existing PTQ and QAT fail to recover. We identify that FP4 failures concentrate on low-entropy tokens—precise symbolic commitments such as digits and operators—where quantization noise inflates sampling errors that cascade through reasoning traces. Based on this insight, we propose ReQAT, a reasoning-centric FP4 training framework with three components: (i) Trace-Aligned QAT (TAQ), which revisits identical reasoning traces to focus updates on critical low-entropy decisions; (ii) Selective Entropy Minimization (SEM), which reinforces confidence at low-entropy positions; and (iii) Q-FIT, a quantization-friendly initialization that jointly calibrates RoPE-consistent KV cache transformations to stabilize QAT. Under the same training budget, ReQAT not only recovers but surpasses BF16 fine-tuning accuracy—achieving while delivering up to $3.9\times$ throughput speedup on NVIDIA DGX Spark and $3.1\times$ on B200. This is the first demonstration that FP4 QAT can exceed full-precision accuracy for LRMs with over 3× speedup on production hardware.

Deep Learning · Foundation Models

Jingru Fei, Kun Yi, Alex Wang, Qingsong Wen, Xiangxiang Zhu, Wei Fan

Time series foundation models rely on large-scale pretraining over diverse datasets across domains, yet their heterogeneity in temporal patterns could hinder the effectiveness of training and learning transferable time series representations. Inspired a fundamental concept, normalized power spectral density (PSD) in signal processing, we assume harmonizing datasets via PSDs in the spectral domain could reduce mismatches and enhance pretraining. We then go beyond the direct intractable minimization optimization and innovatively reformulate it as a principled \textit{harmonization} approach. Specifically, we propose \textit{Harmonizer}, a module that reshapes spectral structures and implicitly harmonizing PSDs across datasets, which theoretically corresponds to a shared reparameterization of second-order temporal correlations. Our theoretical analysis further reveals token interactions with Harmonizer can be efficiently mediated by a compact set of resonators, motivating a \textit{HarmonicAttention} design that performs self-attention in a low-dimensional interaction space. Then, we propose \textit{Olivia}, a novel time series foundation model built upon these harmonization mechanisms. Extensive experiments on two large-scale benchmarks (TSLib and GIFT-Eval) and extra 6 datasets from GluonTS, demonstrate Olivia consistently achieves state-of-the-art performance under zero-shot, few-shot, and full-shot forecasting scenarios.

Deep Learning · Large Language Models

Wonjin Lee, Soomi Jeong, Kwang In Kim

Decoder-only large language models (LLMs) struggle with table reasoning because tables must be serialized, obscuring row- and column-level structure. Prior graph and hypergraph approaches encode structure with an external encoder, but their gains are often inconsistent under autoregressive decoding. We analyze how tabular structure is represented inside decoder-only LLMs and find that row and column relations concentrate in a small subset of layers and attention heads. Based on this observation, we propose HInT, which injects hypergraph-derived structural features directly into these structural layers. HInT constructs a table hypergraph over cells and headers, performs lightweight message passing, and fuses the resulting features with token hidden states via token-level gated fusion, while preserving standard autoregressive computation. Experiments across diverse table reasoning tasks show consistent improvements over text-only baselines and prior (hyper)graph-based methods.

Deep Learning · Foundation Models

Xuerui Qiu, Shaowei Gu, Peixi Wu, JiaKui Hu, Yaozhi Wen, Yuqi Pan, Xinhao Luo, Bo XU, Guoqi Li

Spiking Neural Networks (SNNs) offer an energy--efficient route to 3D spatio--temporal perception, yet they lag behind Artificial Neural Networks (ANNs) due to weak pretraining and heavy inference stacks, limiting generalization and multimodal reasoning (e.g., zero--shot 3D classification and open--world QA). We present a universal \textbf{S}pike--based \textbf{V}ision--\textbf{L}anguage pretraining framework (SVL) that equips SNNs with open--world 3D understanding while preserving end--to--end spike efficiency. SVL comprises two core components: (i) {Multi--scale Triple Alignment} (MTA), a label--free triplet contrastive objective aligning 3D, image, and text; and (ii) {Re--parameterizable Vision--Language Integration} (Rep--VLI), which converts offline text embeddings into lightweight weights for text--encoder--free inference. Moreover, we present the first fully spike--driven point Transformer, {Spike-driven PointFormer}, whose 3D spike--driven self--attention (3D-SDSA) reduces interactions to sparse additions, enabling faster, more efficient training. Extensive experiments show that SVL attains strong zero--shot 3D classification (85.4% top--1) and consistently outperforms prior SNNs on downstream tasks (e.g., +6.1% 3D cls, +2.1% DVS actions, +1.1% detection, +2.1% segmentation) while enabling open--world 3D question answering, sometimes outperforming ANNs. To the best of our knowledge, SVL represents the first scalable, generalizable, and hardware-friendly paradigm for 3D open-world understanding, effectively bridging the gap between SNNs and ANNs in complex open-world understanding tasks.

Deep Learning · Large Language Models

Janghwan Lee, Sihwa Lee, Jinseok Kim, Yongjik Kim, Jieun Lim, Jinwook Oh, Jungwook Choi

Large Reasoning Models (LRMs) achieve strong problem-solving through long chain-of-thought, but their deployment is constrained by the high cost of full-precision inference and growing KV cache footprints. Microscaled FP4 formats enable efficient FP4 deployment; however, fully quantizing weights, activations, and KV caches (W4A4KV4) causes severe reasoning degradation that existing PTQ and QAT fail to recover. We identify that FP4 failures concentrate on low-entropy tokens—precise symbolic commitments such as digits and operators—where quantization noise inflates sampling errors that cascade through reasoning traces. Based on this insight, we propose ReQAT, a reasoning-centric FP4 training framework with three components: (i) Trace-Aligned QAT (TAQ), which revisits identical reasoning traces to focus updates on critical low-entropy decisions; (ii) Selective Entropy Minimization (SEM), which reinforces confidence at low-entropy positions; and (iii) Q-FIT, a quantization-friendly initialization that jointly calibrates RoPE-consistent KV cache transformations to stabilize QAT. Under the same training budget, ReQAT not only recovers but surpasses BF16 fine-tuning accuracy—achieving while delivering up to $3.9\times$ throughput speedup on NVIDIA DGX Spark and $3.1\times$ on B200. This is the first demonstration that FP4 QAT can exceed full-precision accuracy for LRMs with over 3× speedup on production hardware.

Yulin He, Wei Chen, Zhikang Jian, Tianhang Guo, Wenjuan Zhou, Minglong Li, Shaowu Yang, Wenjing Yang

Reasoning segmentation is an emerging vision-language task that requires reasoning over intricate text queries to precisely segment objects. However, existing methods typically suffer from overthinking, generating verbose reasoning chains that interfere with object localization in multimodal large language models (MLLMs). To address this issue, we propose DR$^2$Seg, a self-rewarding framework that improves both reasoning efficiency and segmentation accuracy without requiring extra thinking supervision. DR$^2$Seg employs a two-stage rollout strategy that decomposes reasoning segmentation into multimodal reasoning and referring segmentation. In the first stage, the model generates a self-contained description that explicitly specifies the target object. In the second stage, this description replaces the original complex query to verify its self-containment. Based on this design, two self-rewards are introduced to mitigate overthinking and the associated attention dispersion. Extensive experiments conducted on 3B and 7B variants of Qwen2.5-VL, as well as on both SAM2 and SAM3, demonstrate that DR$^2$Seg consistently improves reasoning efficiency and overall segmentation accuracy. Codes are available in supplementary materials.

Deep Learning · Large Language Models

Rubing Yang, Huajun Bai, Song Liu, Guanghua Yu, Runzhi Fan, Yanbin Dang, Zhang Jiejing, Kai Liu, Jianchen Zhu, Peng Chen

Despite their strong performance on reasoning tasks, large reasoning models (LRMs) often suffer from overthinking, producing unnecessarily long outputs and incurring high end-to-end latency, a significant limitation to their real-world deployment. To address overthinking, early-exit mechanisms have been proposed to terminate reasoning before typical completion, showing that this approach can effectively shorten generation length with minimal impact on accuracy. However, their reliance on probing mechanisms introduces a detection overhead that limits their end-to-end latency gains and compromises their generalizability across diverse problems. Inspired by the use of hidden states in speculative decoding, we propose **SpecExit**, a novel framework that predicts both future tokens and an early-exit signal directly from a lightweight draft model without probing overhead. Our method offers significant improvements, achieving up to 66\% generation length reduction and 2.5× end-to-end speedup compared with the speculative decoding baseline, without compromising accuracy. Our method leverages the inherent signals from hidden states to provide effective early-exit signals, suggesting broader use of hidden states for efficient reasoning. Our code is available at: https://anonymous.4open.science/r/SpecExit-B802.

Deep Learning · Large Language Models

Dmitrii Tarasov, Timofei Lashukov, Elizaveta Goncharova, Andrey Kuznetsov

Token cramming compresses sequences into learned embeddings with near-perfect reconstruction, but prior work used fixed token budgets and 99\% accuracy thresholds, obscuring whether residual errors reflect optimization failures or fundamental limits. We introduce progressive cramming, which grows the target prefix token-by-token and stops only when reconstruction is no longer achievable within a fixed optimization budget. Using progressive trajectories, we find that optimization paths occupy surprisingly low-dimensional structure in the embedding space. Attention analysis shows that compression embeddings often become attention sinks in specific intermediate layers, which correlates with both optimization difficulty and downstream degradation. On likelihood-based multiple-choice evaluation, prepending a crammed embedding drops accuracy to close to random guessing, even with the original prefix in context. These results suggest that perfect reconstruction can arise from brittle steering and attention hijacking, rather than a transferable semantic representation. Our results position progressive cramming as a tool for studying compression limits, while showing that perfect reconstruction is insufficient for meaningful compression.

Applications · Computer Vision

Keyou Zheng, Xuyang Su, Jiewu Leng

Parametric CAD is widely used in mechanical and product engineering, but current generative models still have difficulty producing assemblies that are both editable at the parameter level and consistent with inter-part constraints. Methods that generate meshes or history-free B-rep can represent multi-part shape, but they often lack the program structure and constraint logic needed for reliable downstream edits; in contrast, code-based CAD generation offers direct parametric control, yet most published settings and evaluations focus on single-part solids rather than constrained assemblies. We introduce SPADA (Self-testing Parametric Assembly Design Agent), a test-driven agent that synthesizes assembly code together with deterministic verification tests, and uses these tests as an executable contract for controllable generation. SPADA runs an iterative compile-test-repair loop with multimodal feedback, checking both specification logic and physical feasibility through programmatic constraints. To support evaluation, we release SPADA-Bench-Verified, a human-validated benchmark of real-world code-centric assemblies paired with deterministic tests and engineering-style constraints. Experiments show that SPADA could produces complex assemblies while maintaining geometric fidelity, supporting test-driven agents as a concrete path toward reliable, controllable CAD generation.

Vaibhav Singh, Oleksiy Ostapenko, Pierre-André Noël, Eugene Belilovsky, Torsten Scholak

Diffusion language models (DLMs) have emerged as a promising alternative to autoregressive (AR) generation, yet their reliance on Transformer backbones limits inference efficiency due to quadratic attention or KV-cache overhead. We introduce DiffuMamba, a masked diffusion language model built on a bidirectional Mamba backbone that combines the diffusion objective with linear-time sequence modeling, and DiffuMamba-H, a hybrid variant with interleaved attention. Across scales up to 1.3B parameters, our models match Transformer-based diffusion in downstream performance while achieving up to 8.2× and 4.3× higher inference throughput, respectively, on long sequences. We further present a systematic analysis of inference efficiency across modern DLM variants, combining asymptotic complexity with empirical measurements. Notably, cache-efficient block diffusion with Mamba mixers emerges as the only strategy that scales linearly with sequence length and achieves the strongest performance across all baselines, suggesting a promising direction for future diffusion-based generation systems.