论文检索

Reinforcement Learning · Multi-agent

lei yuan, Ruiqi Xue, Yang Yu

Offline meta reinforcement learning (RL) enables agents to learn a unified policy from multi-task offline data to support generalization in out-of-distribution (OOD) tasks. Recent approaches in single-agent RL tackle this by learning an efficient task representation to distinguish between tasks, showing promising adaptation ability. However, when extended to multi-agent settings, these methods struggle with decentralized task identification due to limited global information, and suffer from inefficient knowledge transfer in the absence of role information. To address this, we propose D$^2$TR, a novel context-based meta RL framework with efficient decentralized and disentangled task-role identification. Specifically, D$^2$TR first introduces mutual information knowledge distillation to align decentralized task representations with centralized task representations inferred from global trajectories, enabling efficient decentralized team-centric information identification. Next, D$^2$TR leverages a large language model to assign semantic roles to trajectories in offline data, and achieves effective individual-centric information inference by learning decentralized role representations. Extensive experiments conducted on commonly used multi-agent environments, including CN, SMAC, and SMACv2, demonstrate that D$^2$TR exhibits strong generalization performance to unseen tasks, outperforming prior multi-agent multi-task and context-based meta RL baselines.

Deep Learning · Other Representation Learning

Bianca Lamm, Janis Keuper

This paper introduces mSOP-765k, a large-scale benchmark for the evaluation of multi-modal Structured Output Prediction (mSOP) pipelines. Besides novel evaluation metrics, the benchmark provides combined training and test datasets with over 765,000 images taken from real-world product advertisements. Each of these images contains product visualizations, textual information like product name or brand, and numerical data such as product weight, price, and discount. All images are annotated with the corresponding structured information in form of dictionaries containing key-value pairs. An initial baseline evaluation, including various LLMs and VLMs, as well as multi-modal RAG approaches, shows that the proposed benchmark provides a challenging problem which can not yet be solved completely by state-of-the-art mSOP methods. The benchmark and dataset are available under a creative-commons license: https://www.msop-765k.org/

Deep Learning · Large Language Models

Tianyu Xie, Yuexiao Ma, Yuhang Wu, Wang Chen, Jiayi Ji, Tat-Seng Chua, Xiawu Zheng, Rongrong Ji

Building interactive omni-modal assistants often relies on end-to-end multimodal alignment to fuse heterogeneous modalities, which incurs substantial data and compute costs and limits extensibility. We present Training-Free Large Language Model Orchestration (LLM Orchestration), a training-free orchestration framework that integrates off-the-shelf modality experts into a unified multimodal input--output system without additional gradient-based training for integration. LLM Orchestration comprises three components: (1) an LLM controller that infers user intent and emits explicit control tokens for expert selection and sequencing, enabling protocol-constrained and auditable routing; (2) a text-centric cross-modal memory that compresses multimodal evidence into structured records for lightweight retrieval and reuse, reducing redundant expert invocations across turns; and (3) a unified interaction layer that executes routing and memory decisions to support consistent modality transitions, full-duplex streaming, and interruption-aware dialogue. Across diverse multimodal benchmarks, LLM Orchestration achieves strong performance under standard evaluation constraints while maintaining low orchestration overhead and modular upgradeability, providing a practical alternative to costly joint training for omni-modal systems.

General Machine Learning · Transfer, Multitask and Meta-learning

Jiangtao Kong, Peijun Zhao, Chun-Fu (Richard) Chen, Youngwook Do, Shaohan Hu, Tianyi Zhou, Huajie Shao

Incremental Learning (IL) for Open-ended Image-to-Text Generation (OpenITG) enables models to continuously generate accurate, contextually relevant text for new images while preserving previously acquired knowledge. Unlike prior studies, this paper addresses a more practical scenario in which the predominant category of visual data shifts over time as environments evolve. In this context, we introduce a new notion of continual alignment, which incrementally adapts the alignment module within pre-trained VLMs to preserve high-quality cross-modal representations. Based on this idea, we propose **E**fficient **C**ontinual **A**lignment (ECA), a novel exemplar-free IL approach for OpenITG. The key challenge is enabling the model to acquire new, task-specific features while minimizing interference with the established alignment without accessing raw data from previous tasks. To address this, ECA employs three core mechanisms: a **M**ixture **o**f **Q**uery (MoQ) module that adapts task-specific query tokens, a **F**ish**e**r **D**ynamic **Ex**pansion (FeDEx) that dynamically expands model structure based on a Fisher Information Matrix (FIM)-based metric, and an embedding dictionary with **D**ictionary **R**eplay (DR) to retain past knowledge. To evaluate ECA's performance, we construct four new IL OpenITG benchmarks that better reflect real-world scenarios. Experimental results demonstrate that ECA significantly mitigates catastrophic forgetting and improves IL performance compared to baseline methods. Benchmarks are available at https://anonymous.4open.science/r/ECA-ToS-Benchmarks-FB17.

Deep Learning · Large Language Models

Julian Skifstad, Xinyue Annie Yang, Glen Chou

Inference-time LLM alignment methods, particularly activation steering, offer an alternative to fine-tuning by directly modifying internal activations during generation. Existing methods, however, often rely on non-anticipative interventions that ignore how perturbations propagate through transformer layers and lack online error feedback, resulting in suboptimal, open-loop control. To address this, we show empirically that, despite the nonlinear structure of transformer blocks, layer-wise dynamics across multiple LLM architectures and scales are well-approximated by locally-linear models. Exploiting this property, we model LLM inference as a linear time-varying dynamical system and adapt the classical linear quadratic regulator to compute feedback controllers using layer-wise Jacobians, steering activations toward desired semantic setpoints in closed-loop with minimal computational overhead and no offline training. We also derive theoretical bounds on setpoint tracking error, enabling formal guarantees on steering performance. Using a novel adaptive semantic feature setpoint signal, we show that our method achieves robust, fine-grained behavior control across models, scales, and tasks, including modulation of toxicity, truthfulness, and arbitrary concepts, surpassing baseline steering methods.

Reinforcement Learning · Multi-agent

Chang Yao, Youfang Lin, Shoucheng Song, Hao Wu, Shengkun Yang, Yuqing Ma, Kai Lv

Achieving cross-task generalization remains a critical challenge in Multi-Agent Reinforcement Learning (MARL), fundamentally relying on effective inductive biases. However, existing entity-level biases often overlook collaborative patterns, whereas task-level biases lack sufficient coverage for novel scenarios. To address this, we introduce a role-level inductive bias as an intermediate abstraction that integrates entity-level flexibility with task-level inter-agent collaboration. To instantiate this, we propose Gaussian-mixture-model-based Transferable Role discovery (GTR). Specifically, GTR constructs a structured role space to ensure diverse role assignment, further achieves role decoupling via regularization, and ultimately utilizes these roles for efficient generalization. Empirical results demonstrate that GTR achieves superior zero-shot and few-shot transfer performance on unseen tasks compared to state-of-the-art methods.

Deep Learning · Generative Models and Autoencoders

Jing Xu, Yuexiao Ma, Songwei Liu, Xuzhe Zheng, Shiwei Liu, Chenqian Yan, Xiawu Zheng, Rongrong Ji, Fei Chao, WANG

Autoregressive video generation paradigms offer theoretical promise for long video synthesis, yet their practical deployment is hindered by the computational burden of sequential iterative denoising. While cache reuse strategies can accelerate generation by skipping redundant denoising steps, existing methods rely on coarse-grained chunk-level skipping that fails to capture fine-grained pixel dynamics. This oversight is critical: pixels with high motion require more denoising steps to prevent error accumulation, while static pixels tolerate aggressive skipping. We formalize this insight theoretically by linking cache errors to residual instability, and propose $\textbf{MotionCache}$, a motion-aware cache framework that exploits inter-frame differences as a lightweight proxy for pixel-level motion characteristics. MotionCache employs a coarse-to-fine strategy: an initial warm-up phase establishes semantic coherence, followed by motion-weighted cache reuse that dynamically adjusts update frequencies per token. Extensive experiments on state-of-the-art models like SkyReels-V2 and MAGI-1 demonstrate that MotionCache achieves significant speedups of $\textbf{6.28}\times$ and $\textbf{1.64}\times$ respectively, while effectively preserving generation quality (VBench: 1%$\downarrow$ and 0.01%$\downarrow$ respectively). The code is available at https://anonymous.4open.science/r/MotionCache.

Deep Learning · Large Language Models

Qingchuan Ma, Yuexiao Ma, Yongkang Xie, Tianyu Xie, Xiawu Zheng, Rongrong Ji

Abstract reasoning ability reflects the intelligence and generalization capacity of LLMs to extract and apply abstract rules. However, accurately measuring this ability remains challenging: existing benchmarks either rely on expensive manual annotation, limiting their scale, or risk measuring memorization rather than genuine reasoning. To address this, we introduce an automated pipeline named A$^2$RBench, encompassing generation, expansion, evaluation, and analysis. Specifically, in the generation stage, LLMs create diverse tasks demanding genuine reasoning; in the expansion stage, LLMs reuse validated rules and expand new input spaces to generate task variations, achieving scaling. However, such a process may cause hallucinations. To eliminate it, we further establish a theoretical framework and prove that programmatic verification—testing whether the inverse operation perfectly reverses the forward operation (cycle consistency)—guarantees a unique solution. Through extensive evaluations on mainstream LLMs, we find: (1) Current LLMs exhibit fundamental deficiencies in abstract reasoning, with top models significantly underperforming humans on a representative subset (39.8\% vs. 68.5\%). (2) Current LLMs fall far short of 2D and 1D in the complexity of generated 3D tasks, revealing their lack of understanding of high-dimensional tasks. (3) Counterintuitively, inputs with higher information complexity can simplify the reasoning process. Code and data are available at: https://anonymous.4open.science/r/A2Rbench.

Reinforcement Learning · Multi-agent

Beyazit Yalcinkaya, Marcell Vazquez-Chanlatte, Ameesh Shah, Hanna Krasowski, Sanjit Seshia

We study learning multi-task, multi-agent policies for cooperative, temporal objectives, under centralized training, decentralized execution. In this setting, using automata to represent tasks assigned to agents enables breaking down a team-level objective into simpler, smaller sub-tasks. However, existing approaches remain sample-inefficient and are limited to the single-task case, requiring retraining policies for each new task. In this work, we present Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning (ACC-MARL), a framework for learning task-conditioned, decentralized team policies. We identify the main challenges to the feasibility of ACC-MARL, propose solutions, and prove that our approach is optimal. We further show that learned value functions can be used to assign tasks optimally at test time. Experiments demonstrate emergent task-aware, multi-step coordination among agents, such as pressing a button to unlock a door, holding the door, and short-circuiting tasks.

Probabilistic Methods · Everything Else

Ege Onur Taga, Samet Oymak, Shubhanshu Shekhar

We develop horizon-aware anytime-valid tests and confidence sequences for bounded means under a strict deadline $N$. Using the betting/e-process framework, we cast horizon-aware betting as a finite-horizon optimal control problem with state space $(t, \log W_t)$, where $t$ is the time and $W_t$ is the test martingale value. We first show that in certain interior regions of the state space, policies that deviate significantly from Kelly betting are provably suboptimal, while Kelly betting reaches the threshold with high probability. We then identify sufficient conditions showing that outside this region, more aggressive betting than Kelly can be better if log-wealth is low or time is short, and less aggressive can be better if log-wealth is high. Taken together these results suggest a simple phase diagram in the $(t, \log W_t)$ plane, delineating regions where Kelly, fractional Kelly, and aggressive betting may be preferable. Guided by this phase diagram, we introduce a Deep Reinforcement Learning approach based on a universal Deep Q-Network (DQN) agent that learns a single policy from synthetic experience and maps simple statistics of past observations to bets across horizons and null values. In limited-horizon experiments, the learned DQN policy outperforms state-of-the-art baselines.

Theory · Online Learning and Bandits

Shinsaku Sakaue, Han Bao, Yuzhou Cao

Online structured prediction, including online classification as a special case, is the task of sequentially predicting labels from input features. In this setting, the *surrogate regret*—the cumulative excess of the actual target loss (e.g., 0–1 loss) over the surrogate loss (e.g., logistic loss) incurred by the best fixed estimator—has gained attention because it admits a finite bound independent of the time horizon $T$. However, such guarantees break down in *non-stationary* environments, where every fixed estimator may incur surrogate loss that grows linearly with $T$. To address this limitation, we obtain an upper bound of $F_T + O(1 + P_T)$ on the cumulative target loss, where $F_T$ is the cumulative surrogate loss of any comparator sequence and $P_T$ is its *path length*. This bound depends on $T$ only through $F_T$ and $P_T$, thus offering stronger guarantees under non-stationarity. Our core idea is to combine the dynamic regret analysis of online gradient descent (OGD) with the *exploit-the-surrogate-gap* technique. This viewpoint sheds light on the usefulness of a Polyak-style learning rate for OGD, which systematically yields target-loss bounds and performs well empirically. We then extend our approach to broader settings beyond prior work via the *convolutional Fenchel–Young loss*. Finally, a lower bound shows that the dependence on $F_T$ and $P_T$ is tight.

General Machine Learning · Causality

Ayush Khot, Miruna Oprescu, Maresa Schröder, Ai Kagawa, Xihaier Luo

Causal inference in spatial domains faces two intertwined challenges: (1) unmeasured spatial factors, such as weather, air pollution, or mobility, that confound treatment and outcome, and (2) interference from nearby treatments that violate standard no-interference assumptions. While existing methods typically address one by assuming away the other, we show they are deeply connected: *interference reveals structure* in the latent confounder. Leveraging this insight, we propose the **Spatial Deconfounder**, a two-stage method that reconstructs a substitute confounder from local treatment vectors using a conditional variational autoencoder (C-VAE) with a spatial prior, then estimates causal effects via a flexible outcome model. We show that this approach enables nonparametric identification of both direct and spillover effects under weak assumptions—without requiring multiple treatment types or a known model of the latent field. Empirically, we extend ```SpaCE```, a benchmark suite for spatial confounding, to include treatment interference, and show that the Spatial Deconfounder consistently improves effect estimation across real-world datasets in environmental health and social science. By turning interference into a multi-cause signal, our framework bridges spatial and deconfounding literatures to advance robust causal inference in structured spatial data.

John Sweeney

Sequential fine-tuning on multiple datasets is ubiquitous, but the training order of sources can measurably change downstream performance; testing both orders roughly doubles compute. We model a single gradient step on a dataset as a nonlinear operator and show that non-commutativity induces order-dependent effects governed by a commutator (Lie-bracket) term. For two sources $A,B$ and target domain $E$, this yields a directional score $\sigma_{AB}^{(E)} = \langle g_E, H_B g_A - H_A g_B \rangle$ that predicts whether $A \to B$ or $B \to A$ yields lower $L_E$. We evaluate $g_E$ at a reference point capturing the shared drift of both orders (Trotter scoring) and develop a theory-driven $\eta$-autopilot that selects step sizes from pilot data by balancing signal-to-noise against higher-order stability constraints. On four LLMs and a diffusion UNet, our planner achieves 81–94% overall sign accuracy and 82–100% on highest-impact decisions, enabling practical transfer-order planning without manual hyperparameter tuning.

Deep Learning · Generative Models and Autoencoders

Dhruvesh Patel, Benjamin Rozonoyer, Gaurav Pandey, Tahira Naseem, Ramón Astudillo, Andrew McCallum

In many domains generating variable length sequences through insertions provides greater flexibility over autoregressive models. However, the action space of insertion models is much larger than that of autoregressive models (ARMs) making the learning challenging. To address this, we incorporate trainable order dynamics into the target rates for discrete flow matching, and show that with suitable choices of parameterizations, joint training of the target order dynamics and the generator is tractable without the need for numerical simulation. As the generative insertion model, we use a variable length masked diffusion model, which generates by inserting and filling mask tokens. On graph traversal tasks for which a locally optimal insertion order is known, we explore the choices of parameterization empirically and demonstrate the trade-offs between flexibility, training stability and generation quality. On de novo small molecule generation, we find that the learned order dynamics leads to a significant increase in validity and quality of the generated molecules, when compared to uniform order dynamics.

Deep Learning · Attention Mechanisms

Lijie Yang, Zhihao Zhang, Arti Jain, Shijie Cao, Baihong Yuan, Yiwei Chen, Zhihao Jia, Ravi Netravali

Large reasoning models achieve strong performance through test-time scaling, but this incurs substantial computational overhead due to long decoding from short prompts. While sparse attention can reduce latency and memory usage, existing methods often degrade reasoning accuracy because selection errors accumulate over long generation horizons, or require costly retraining. We introduce LessIsMore, a training-free sparse attention mechanism for long-horizon reasoning. Our key insight is that token importance in reasoning is global and stable: critical tokens are largely shared across attention heads and remain stable over decoding steps. Guided by this structure, LessIsMore enforces cross-head unified token selection and preserves recent context via a stable recency window, yielding a globally consistent token set that can be reused across layers. Across multiple model families and challenging reasoning benchmarks, LessIsMore matches or improves accuracy while attending to substantially fewer tokens. With kernel-level optimizations, LessIsMore achieves up to $1.6\times$ end-to-end decoding speedup and up to $1.72\times$ faster sparse attention computation, with additional long-context results demonstrating the generality of our approach.

General Machine Learning · Evaluation

Jin Gao, Juntu Zhao, Zirui Zeng, Jiaqi Shen, Junhao Shi, Dukun Zhao, Yuming Lu, Dequan Wang

Existing benchmarks for biological language models (BLMs) inadequately capture the challenges of real-world applications, often lacking realistic out-of-distribution (OOD) scenarios, evolutionary depth, and consistency in measurement. To address this, we introduce TadABench-1M, a new benchmark based on a wet-lab dataset of over one million variants of the therapeutically relevant TadA enzyme, purpose-built to embody these three essential attributes. Generated across 31 rounds of wet-lab evolution, it offers unparalleled evolutionary depth and naturally presents a stringent OOD challenge. To ensure measurement consistency across this extensive campaign, we developed Seq2Graph, a scalable graph-based algorithm that systematically unifies multi-batch experimental data. Our high-fidelity benchmark highlights a critical finding: while state-of-the-art BLMs excel on a standard random split of the data (Spearman’s ρ ≈ 0.8), they fail dramatically on a realistic temporal prediction task (ρ ≈ 0.1). This stark performance gap validates the importance of our benchmark’s design principles and suggests that evolutionary depth is critical for building models with realistic utility.

Applications · Health / Medicine

Shuohao Gao, Xuanzhong Chen, Lingxiao Luo, Zilin Ding, Rong Han, Rui Jiang, Ting Chen

Diagnosing complex diseases is inherently a sequential and iterative medical investigation process, in which a clinician strategically requests multiple rounds of diagnostic tests to differentiate among similar diseases until reaching a definitive diagnosis. Although large language models show great potential as clinical assistants, they often struggle to navigate this complex interactive process, suffering from premature diagnostic closure. Furthermore, optimizing LLMs for such multi-round environments is frequently hindered by the challenge of reward sparsity and hacking. In this paper, we introduce $\textbf{CompDiag-Bench}$, a benchmark that formalizes diagnosis as a sequential decision-making process where a clinician must strategically request diagnostic tests from a dynamic environment in order to reach a definitive diagnosis. To address this task, we propose $\texttt{Salus}$, a multi-agent framework that decouples diagnostic reasoning into three specialized functional roles: a Differential Reasoner, a Strategic Controller, and a Workup Proposer. $\texttt{Salus}$ is optimized via multi-agent reinforcement learning employing structured rewards to calibrate strategic diagnostic behavior. Specifically, we leverage an LLM-as-a-Judge reward mechanism to provide dense, semantically-grounded feedback, designed to penalize premature closure and incentivize accurate differential diagnoses. Experimental results show that our model, $\texttt{Salus-7B}$, attains state-of-the-art Top-1 accuracy of $83.64\%$ on complex cases, outperforming DeepSeek-V3.2 ($71.38\%$) and achieving performance on par with GPT-5.2 ($80.30\%$).

Melody Guan, Miles Wang, Micah Carroll, Zehao Dou, Annie Wei, Marcus Williams, Benjamin Arnav, Joost Huizinga, Ian Kivlichan, Amelia Glaese 等

Safe deployment of increasingly capable AI agents may require visibility into how they make decisions. Chain-of-thought (CoT) monitoring can detect misbehavior in today’s reasoning models, but this “monitorability” may be fragile under different training procedures, data sources, or continued system scaling. We propose three evaluation archetypes (intervention, process, and outcome-property), a new monitorability metric, and a broad evaluation suite. We show CoT monitoring outperforms action-only monitoring in practical settings, and that frontier models are generally—but not perfectly—monitorable. We study scaling trends with pre-training model size and inference-time compute, finding longer CoTs are typically more monitorable. We find that, for a fixed capability level, using a smaller model at higher reasoning effort can yield higher monitorability, at greater inference compute cost. We further find that increasing a weak monitor’s test-time compute when monitoring a strong agent improves monitorability, and giving the monitor access to the CoT both boosts monitorability and steepens the compute–to-monitorability scaling trend. Finally, we show monitorability can be improved by asking follow-up questions and giving the follow-up CoT to the monitor.

Xinchang Wang, Yunhao Chen, Yuechen Zhang, Congcong Bian, Zihao Guo, Xingjun Ma, Hui Li

Recent image generators produce photo-realistic content that undermines the reliability of downstream recognition systems. As visual appearance cues become less pronounced, appearance-driven detectors that rely on forensic cues or high-level representations lose stability. This motivates a shift from appearance to behavior, focusing on how images respond to controlled perturbations rather than how they look. In this work, we identify a simple and universal behavioral signal. Natural images preserve stable semantic representations under small, structured perturbations, whereas generated images exhibit markedly larger feature drift. We refer to this phenomenon as \textbf{robustness asymmetry} and provide a theoretical analysis that establishes a lower bound connecting this asymmetry to memorization tendencies in generative models, explaining its prevalence across architectures. Building on this insight, we introduce Robustness Asymmetry Detection (RA-Det), a behavior-driven detection framework that converts robustness asymmetry into a reliable decision signal. Evaluated across 14 diverse generative models and against more than 10 strong detectors, RA-Det achieves superior performance, improving the average performance by 12.92\%. The method is data- and model-agnostic, requires no generator fingerprints, and transfers across unseen generators. Together, these results indicate that robustness asymmetry is a stable, general cue for synthetic-image detection and that carefully designed probing can turn this cue into a practical, universal detector.

Deep Learning · Large Language Models

Md Kowsher, Haris Mansoor, Nusrat Prottasha, Ozlem Garibay, Victor Zhu, Zhengping Ji, Chen Chen

MoE-PEFT methods combine Mixture of Experts with parameter-efficient fine-tuning for multi-task adaptation, but require separate adapters per expert—causing trainable parameters to scale linearly with expert count and limiting applicability to adapter-based architectures. We propose LiME (Lightweight Mixture of Experts), which achieves expert specialization through lightweight modulation rather than adapter replication. Instead of separate adapters, LiME uses a single shared PEFT module and modulates its output with lightweight expert vectors, reducing expert parameters while generalizing to any PEFT method. Notably, LiME introduces zero-parameter routing by leveraging existing frozen and adapted representations—eliminating learned router parameters typically required per layer. Theoretically, we prove that (i) more experts preserve more task-relevant information and (ii) modulation approximates full expert-specific PEFT with bounded error. LiME further incorporates n-gram windowed routing and adaptive expert selection (Auto Top-K) based on routing confidence. Experiments on MMT-47, a multimodal multi-task benchmark with 47 tasks spanning text, image, and video, demonstrate that LiME achieves competitive or superior performance while using up to 4× fewer trainable parameters and up to 29% faster training compared to corresponding MoE-PEFT baselines.