论文检索

Reinforcement Learning · Multi-agent

Jiajun Wu, Xuefeng Du, Yuduo Zheng, Fengqi Li

Decentralized multi-agent reinforcement learning faces a persistent exploration–coordination tension: intrinsic rewards promote exploration under sparse feedback, yet effective cooperation requires agents’ behaviors to remain consistent over a limited communication graph. Existing methods often combine exploration bonuses and coordination regularizers with fixed-weight schedules, making them hard to tune and prone to either fragmented conventions or premature behavioral collapse. We propose the IEC (Isomorphic Exploration-Consensus) framework that couples exploration and coordination through a single constrained objective: maximize task return augmented with two complementary exploration signals, dynamics-based information gain and state-coverage novelty, while constraining graph-induced policy disagreement via a spectral smoothness penalty on neighboring agents, which can be interpreted as a Dirichlet-energy regularizer on the communication graph. IEC optimizes the resulting Lagrangian with a lightweight primal–dual update that adapts the consensus multiplier from observed constraint violations, yielding an automatic shift from diverse exploration to stable cooperative conventions. Across three distinct benchmarks, IEC achieves superior performance.

Applications · Computer Vision

Wenjun Miao, Mingda Li, Yanchao Hao, Zheng Wei

While large vision-language models (LVLMs) have shown remarkable adaptability to downstream applications, their fine-tuning process remains susceptible to bias under long-tailed data. Compared to zero-shot scenarios, fine-tuning LVLMs on imbalanced datasets often yields limited performance improvements on tail data. This is because LVLMs tend to rapidly overfit the head data at an early fine-tuning stage, thereby impairing the learning of the tail data while simultaneously failing to exploit their quantitative advantage. Furthermore, in many downstream LVLM scenarios, quantified long-tailed prior knowledge of data distribution is often unavailable, significantly limiting the applicability of traditional long-tailed techniques that rely heavily on such information. To address these issues, we propose the Adaptive Token Refinement (ATR), a novel framework that adaptively refines the learning process of LVLMs under long-tailed data. Specifically, ATR consists of two token-level operations applied to output and input tokens, respectively: 1) a bounded adaptive loss that dynamically filters and reweights output tokens to mitigate overfitting on head data, and 2) a visual token mask strategy that augments the probability paths of input tokens to enhance long-tailed performance. Extensive experiments across multiple benchmarks demonstrate that ATR consistently enhance both performance and generalization for long-tailed LVLMs fine-tuning in a distribution-agnostic manner.

Social Aspects · Safety

Chuancheng Shi, shangze li, Wenjun Lu, Wenhua Wu, Fei Shen, Cong Wang, Zifeng Cheng, Tat-Seng Chua

Despite their capabilities, large foundation models (LFMs) remain susceptible to adversarial manipulation. Current defenses predominantly rely on the ``locality hypothesis", suppressing isolated neurons or features. However, harmful semantics act as distributed, cross-layer circuits, rendering such localized interventions brittle and detrimental to utility. To bridge this gap, we propose \textbf{TraceRouter}, a path-level framework that traces and disconnects the causal propagation circuits of illicit semantics. TraceRouter operates in three stages: (1) it pinpoints a sensitive onset layer by analyzing attention divergence; (2) it leverages sparse autoencoders (SAEs) and differential activation analysis to disentangle and isolate malicious features; and (3) it maps these features to downstream causal pathways via feature influence scores (FIS) derived from zero-out interventions. By selectively suppressing these causal chains, TraceRouter physically severs the flow of harmful information while leaving orthogonal computation routes intact. Extensive experiments demonstrate that TraceRouter significantly outperforms state-of-the-art baselines, achieving a superior trade-off between adversarial robustness and general utility. Our code will be publicly released. WARNING: This paper contains unsafe model responses.

Applications · Computer Vision

Guangze Shi, Yingjie Mi, Jia Shen, Feixue Shao, Jiarui Cao, Yexin Lai, Xueyu Liu, Rui Wang, Yongfei Wu, Mingqiang Wei

Optimizing prompts for foundation models like SAM represents a challenging high-dimensional black-box optimization problem, fundamentally plagued by the credit assignment ambiguity. To address this, we introduce PromptPilot, a task-agnostic reinforcement learning framework that structurally decomposes the search space into orthogonal semantic and spatial subspaces. Specifically, a centralized manager orchestrates two specialized agents, a feature agent ensuring semantic coherence and a physical agent maximizing spatial coverage, to navigate conflicting optimization objectives. Crucially, our reward mechanism synergizes global segmentation feedback with an efficient approximation of Shapley values, enabling fine-grained attribution of performance gains to individual prompt actions. PromptPilot functions as an inference-time optimization strategy without parameter updates. Extensive experiments demonstrate that our game-theoretic approach significantly improves segmentation performance and generalization, offering a principled solution for automated prompt engineering.

Social Aspects · Accountability, Transparency, and Interpretability

Xinting Huang, Aleksandra Bakalova, Satwik Bhattamishra, William Merrill, Michael Hahn

Recent work has shown that the computations of Transformers can be simulated in the RASP family of programming languages. These findings have enabled improved understanding of the expressive capacity and generalization abilities of Transformers. In particular, Transformers have been suggested to length-generalize exactly on problems that have simple RASP programs. However, it remains open whether trained models actually implement simple interpretable programs. In this paper, we present a general method to extract such programs from trained Transformers. The idea is to faithfully re-parameterize a Transformer as a RASP program and then apply causal interventions to discover a small sufficient sub-program. In experiments on small Transformers trained on algorithmic and formal language tasks, we show that our method often recovers simple and interpretable RASP programs from length-generalizing transformers. Our results provide the most direct evidence so far that Transformers internally implement simple RASP programs.

Deep Learning · Large Language Models

Ankur Samanta, Akshayaa Magesh, Ayush Jain, Kavosh Asadi, Youliang Yu, Daniel Jiang, Boris Vidolov, Kaveh Hassani, Paul Sajda, Jalaj Bhandari 等

Self-correction in language models remains elusive. In this work, we explore whether language models can explicitly localize errors in incorrect reasoning, as a path toward building AI systems that can effectively correct themselves. We introduce a prompting method that structures reasoning as discrete, semantically coherent thought steps, and show that models are able to reliably localize errors within this structure, while failing to do so in conventional, unstructured chain-of-thought reasoning. Motivated by how the human brain monitors errors at discrete decision points and resamples alternatives, we introduce Iterative Correction Sampling of Thoughts (Thought-ICS), a self-correction framework. Thought-ICS iteratively prompts the model to generate reasoning one discrete and complete thought at a time—where each thought represents a deliberate decision by the model—creating natural boundaries for precise error localization. Upon verification, the model localizes the first erroneous step, and the system backtracks to generate alternative reasoning from the last correct point. When asked to correct reasoning verified as incorrect by an oracle, Thought-ICS achieves 20-40\% self-correction lift. In a completely autonomous setting without external verification, it outperforms contemporary self-correction baselines.

Optimization · Zero-order and Black-box Optimization

Yan Zhang, Xuefeng Liu, Sipeng Chen, Sascha Ranftl, Chong Liu, Shibo Li

Standard Bayesian Optimization (BO) assumes uniform smoothness across the search space—an assumption violated in multi-regime problems such as molecular conformation search through distinct energy basins or drug discovery across heterogeneous molecular scaffolds. A single GP either oversmooths sharp transitions or hallucinates noise in smooth regions, yielding miscalibrated uncertainty. We propose RAMBO, a Dirichlet Process Mixture of Gaussian Processes that automatically discovers latent regimes during optimization, each modeled by an independent GP with locally-optimized hyperparameters. We derive collapsed Gibbs sampling that analytically marginalizes latent functions for efficient inference, and introduce adaptive concentration parameter scheduling for coarse-to-fine regime discovery. Our acquisition functions decompose uncertainty into intra-regime and inter-regime components. Experiments on synthetic benchmarks and real-world applications—including molecular conformer optimization, virtual screening for drug discovery, and fusion reactor design—demonstrate consistent improvements over state-of-the-art baselines on multi-regime objectives.

Applications · Computer Vision

Xuyue Huang, Zhe Chen, Wang Shen, Xiao-Ping Zhang

Diffusion Transformers (DiTs) have driven substantial progress in image and video generation but suffer from prohibitive computational costs. Feature caching accelerates inference by reusing intermediate representations. Existing methods rely on historical features for implementation simplicity, yet suffer from severe error accumulation at high acceleration ratios. To address this limitation, we investigate the nature of the requisite feature correction. We demonstrate that the optimal calibration update is characterized by a shared low-rank subspace across diverse prompts. Guided by this structural insight, we propose LearniBridge, a learnable calibration mechanism for feature caching that bridges multiple timesteps through lightweight LoRA updates. This mechanism enables effective calibration requiring only $3-5$ training samples. Extensive experiments on image and video generation show that LearniBridge achieves up to $5.87\times$, $5.75\times$, and $4.10\times$ acceleration on FLUX, HunyuanVideo, and WAN 2.1, respectively. On WAN 2.1, it improves VBench by 1.28\% over the previous SOTA at $4.10\times$ acceleration. Our code is included in the supplementary material and will be released on GitHub.

General Machine Learning · Causality

Anahita Haghighat, Dominik Janzing

Building on recent formalizations of root cause analysis for rare events (“outliers”) in structural equation models, we propose a formal definition of a causal pathway and discuss its testable implications. We identify conditions under which these implications depend only on a causal abstraction defined by the pathway of rare events, rather than on the full causal graph of the underlying system. Accordingly, we introduce an abstraction of causal structure to pathways of rare events that bridges simple verbal causal explanations and detailed causal modeling.

Deep Learning · Large Language Models

Tingting Chen, Beibei Lin, Zifeng Yuan, Qiran Zou, Hongyu He, Anirudh Goyal, Yew Soon ONG, Dianbo Liu

Many scientific problems are underdetermined: multiple distinct hypotheses are equally consistent with the same observations. In such settings, effective inference requires not only producing valid explanations, but also systematically exploring and covering the admissible hypothesis set. We introduce HypoSpace, a benchmark that treats large language models (LLMs) as samplers over finite hypothesis spaces and evaluates them on three metrics: Validity, Uniqueness, and Recovery. HypoSpace spans three structured domains (causal graph inference, gravity-constrained 3D voxel reconstruction, and Boolean genetic interaction modeling) with deterministic validators and exactly enumerable solution spaces, plus real-world anchored case studies. Empirically, frontier LLMs exhibit a consistent failure mode: high Validity but sharp degradation in Uniqueness and Recovery as hypothesis spaces grow. We further show that stratified decoding partially mitigates this collapse, demonstrating HypoSpace's utility as a diagnostic benchmark for set-valued inference.

Applications · Computer Vision

Yitong Yang, Xuexin Liu, Yinglin Wang, Jing Wang, Hao Dou, Changshuo Wang, Shuting He

3D style transfer enables the creation of visually expressive 3D content, enriching the visual appearance of 3D scenes and objects. However, existing VGG- and CLIP-based methods struggle to model multi-view consistency within the model itself, while diffusion-based approaches can capture such consistency but rely on denoising directions, leading to unstable training. To address these limitations, we propose DiffStyle3D, a novel diffusion-based paradigm for 3DGS style transfer that directly optimizes in the latent space. Specifically, we introduce an Attention-Aware Loss that performs style transfer by aligning style features in the self-attention space, while preserving original content through content feature alignment. Inspired by the geometric invariance of 3D stylization, we propose a Geometry-Guided Multi-View Consistency method that integrates geometric information into self-attention to enable cross-view correspondence modeling. Based on geometric information, we additionally construct a geometry-aware mask to prevent redundant optimization in overlapping regions across views, which further improves multi-view consistency. Extensive experiments show that DiffStyle3D outperforms state-of-the-art methods, achieving higher stylization quality and visual realism.

Applications · Neuroscience, Cognitive Science

Chengrui Li, Yunmiao Wang, Yule Wang, Weihan Li, Dieter Jaeger, Anqi Wu

Low-rank recurrent neural networks (lrRNNs) are a class of models that uncover low-dimensional latent dynamics underlying neural population activity. Although their functional connectivity is low-rank, it lacks independence interpretations, making it difficult to assign distinct computational roles to different latent dimensions. To address this, we propose the Factored Recurrent Neural Network (FacRNN), a generative lrRNN framework that assumes group-wise independence among latent dynamics while allowing flexible within-group entanglement. These independent latent groups allow latent dynamics to evolve separately, but are internally rich for complex computation. We reformulate the lrRNN under a variational autoencoder (VAE) framework, enabling us to introduce a partial correlation penalty that encourages independence between groups of latent dimensions. Experiments on synthetic, monkey M1, and mouse voltage imaging data show that FacRNN consistently improves the disentanglement and interpretability of learned neural latent trajectories in low-dimensional space and low-rank connectivity over baseline lrRNNs that do not encourage group-wise independence.

Deep Learning · Large Language Models

Hen Davidov, Nachshon Cohen, Oren Kalinsky, Yaron Fairstein, Guy Kushilevitz, Ram Yazdi, Patrick Rebeschini

Large language models (LLMs) using chain-of-thought reasoning often waste substantial compute by producing long, incorrect responses. Abstention can mitigate this by withholding outputs unlikely to be correct. While most abstention methods decide to withhold outputs before or after generation, dynamic mid-generation abstention considers early termination of unpromising reasoning traces at each token position. Prior work has explored empirical variants of this idea, but principled guidance for the abstention rule remains lacking. We present a formal analysis of dynamic abstention for LLMs, modeling abstention as an explicit action within a regularized reinforcement learning framework. An abstention reward parameter controls the trade-off between compute and information. We show that abstaining when the value function falls below this reward strictly outperforms natural baselines under general conditions. We further derive a principled and efficient method to approximate the value function. Empirical results on mathematical reasoning tasks support our theory and demonstrate improved selective accuracy over existing methods.

Lin Wang, Zhichao Wang, Ye Shi, Sai Praneeth Reddy Karimireddy, Xiaoying Tang

Federated Learning (FL) often suffers from a trade-off between global model performance and client-level fairness due to data heterogeneity, which often leads to inconsistent performance of the globally trained models, resulting in unfair outcomes among users. Existing fair FL algorithms face a trade-off: they either sacrifice global model performance to promote fairness or fall short of achieving optimal fairness. In this paper, we propose a novel framework that bridge this trade-off by integrating information-theoretic principles with model alignment. Specifically, we leverage the Maximum Entropy Principle to derive an analytic, closed-form solution for fair aggregation weights, ensuring significant fairness enhancements with minimal computational overhead. To maintain the global model performance, we further employ a step-wise model alignment strategy that synchronizes gradient directions across heterogeneous clients, effectively mitigating the drift induced by local updates. Theoretical analysis proves that our method guarantees convergence even in non-convex settings. Importantly, we push the theoretical frontier of federated fairness by extending performance variance analysis to generalized regression, providing broader guarantees. Extensive experiments on five datasets demonstrate that our approach consistently outperforms state-of-the-art methods, achieving superior fairness without sacrificing global accuracy.

Deep Learning · Foundation Models

Can Jin, Hongwu Peng, Mingcan Xiang, Qixin Zhang, Xiangchi Yuan, Amit Hasan, Ohi Dibua, Yifan Gong, Yan Kang, Dimitris Metaxas

Sparse Mixture-of-Experts architectures are essential for scaling model capacity efficiently, yet the standard Top-$k$ routing imposes a rigid sparsity pattern that ignores the intrinsic variance in token difficulty and layer-specific computational needs. While Top-$p$ routing offers a flexible alternative, we demonstrate that existing naive Top-$p$ implementations with fixed global probability thresholds provide only marginal gains over Top-$k$, suffer from hyperparameter sensitivity, and result in uncontrolled computational costs. In this paper, we propose $\texttt{DTop-}p$, a sparsity-controllable dynamic routing mechanism. To overcome the non-differentiability of the MoE sparsity level - the Top-$p$ threshold, we utilize a Proportional-Integral controller that dynamically learns the Top-$p$ probability threshold to align the running sparsity with a user-defined budget. Furthermore, we introduce dynamic routing normalization to adaptively rescale logits, enabling distinct expert selection patterns across layers under a global sparsity constraint. Extensive experiments on Large Language Models and Diffusion Transformers demonstrate that $\texttt{DTop-}p$ consistently outperforms both Top-$k$ and fixed Top-$p$ baselines while matching the average FLOPs of Top-$k$ MoE. Our analysis confirms that $\texttt{DTop-}p$ exhibits strong scaling properties across expert granularity, total expert capacity, model size, and dataset size, offering a robust and efficient MoE framework for foundation model pre-training.

Deep Learning · Large Language Models

Zihan Chen, Chengshuai Shi, Song Wang, Jundong Li, Cong Shen

Prompt optimization is a key way to steer large language models when fine-tuning is impractical. However, instruction optimization (IO) and in-context learning (ICL) demonstration selection are often optimized separately and combined post hoc, implicitly assuming that a "best'' instruction and a "best" demonstration set compose well. In practice, their interactions are strong, making such decoupled pipelines brittle. We propose SMILE, an efficient method that *jointly* selects instructions and demonstrations. Our key observation is that the ICL performance exhibits consistent diminishing returns across diverse instructions. Leveraging this structure, SMILE learns an instruction-conditioned surrogate aligned with LLM feedback and instantiates it as an Extended Deep Submodular Function that captures sample--sample coverage, sample--query relevance, and sample--instruction compatibility. SMILE then performs greedy, query-adaptive selection of the instruction--demonstration pair. Experiments on six datasets and multiple LLM backbones show that SMILE consistently outperforms IO-only, ICL-only, and existing joint baselines, supporting a context engineering view of prompting: jointly optimizing interacting components rather than tuning them in isolation.

Applications · Computer Vision

Bingxuan Zhao, Qing Zhou, Yu Wang, Chuang Yang, Qi Wang

While Diffusion Transformers (DiTs) have revolutionized high-fidelity image synthesis, the prohibitive computational costs of training at ultra-high resolutions necessitate robust inference-time extrapolation. Existing extrapolation methods typically operate under a *scale-agnostic* assumption, treating the denoising dynamics identically across resolutions. In this work, we identify a critical oversight in this paradigm: the spectral evolution of the diffusion process, transitioning from low-frequency structural construction to high-frequency texture refinement, is inherently scale-dependent. Consequently, applying a uniform strategy across scales causes a spectral misalignment, manifesting as *structural collapse* or *textural degradation*. To bridge this gap, we introduce **SigMa ($\sigma$)**, a training-free framework that utilizes Sigmoid Modulation for *scale-adaptive* calibration of the extrapolation process. SigMa orchestrates the spectral evolution via a parameterized schedule with two core mechanisms: *Decoupled Geometric Center Alignment*, which synchronizes the transition timing to secure global structure, and *Iso-Variance Rate Adaptation*, which scales the transition velocity to ensure a smooth feature handover. Extensive experiments demonstrate that SigMa effectively rectifies spectral deviations, enabling training-free extrapolation up to 16 megapixels and achieving state-of-the-art performance on standard benchmarks.

Applications · Language, Speech and Dialog

Kirandeep Kaur, Xingda Lyu, Chirag Shah

Generative AI agents equate *understanding* with resolving explicit queries, an assumption that confines interaction to what users can articulate. This assumption breaks down when users themselves lack awareness of what is missing, risky, or worth considering. In such conditions, proactivity is not merely an efficiency enhancement, but an epistemic necessity. We refer to this condition as *epistemic incompleteness*: where progress depends on engaging with unknown *unknowns* for effective partnership. Existing approaches to proactivity remain narrowly anticipatory, extrapolating from past behavior and presuming that goals are already well defined, thereby failing to support users meaningfully. However, surfacing possibilities beyond a user’s current awareness is not inherently beneficial. Unconstrained proactive interventions can misdirect attention, overwhelm users, or introduce harm. Proactive agents, therefore, require *behavioral grounding*: principled constraints on *when, how*, and to *what extent* an agent should intervene. We advance the position that **generative proactivity must be grounded both epistemically and behaviorally**. Drawing on the *philosophy of ignorance* and *research on proactive behavior*, we argue that these theories offer critical guidance for designing agents that can engage responsibly and foster meaningful partnerships.

Deep Learning · Everything Else

Yeonsang Shin, Insoo Kim, Bongkeun Kim, Keonwoo Bae, Bohyung Han

Transformer-based autoregressive models excel in data generation but are inherently constrained by their reliance on discretized tokens, which limits their ability to represent continuous values with high precision. We analyze the scalability limitations of existing discretization-based approaches for generating hybrid discrete-continuous sequences, particularly in high-precision domains such as logos, layouts, and semiconductor circuit designs, where precision loss potentially leads to visual artifacts, aesthetic degradation, and even functional failure. To address the challenge, we propose a novel unified framework that jointly models discrete and continuous values for variable-length sequences. Our approach employs a hybrid approach that combines categorical prediction for discrete values with diffusion-based modeling for continuous values, incorporating two key technical components: an end-of-sequence (EOS) logit adjustment mechanism that uses an MLP to dynamically adjust EOS token logits based on sequence context, and a length regularization term integrated into the loss function. Additionally, we present ContLayNet, a large-scale benchmark comprising 334K high-precision semiconductor layout samples with specialized evaluation metrics that capture functional correctness, where precision errors significantly impact performance. Experiments on semiconductor layouts (ContLayNet), graphic layouts, and SVGs demonstrate that our approach achieves higher-fidelity hybrid vector representations than discretization-based and fixed-schema baselines, while scaling to high-precision generation across multiple domains.

Applications · Computer Vision

Zhirui Chen, Ziwei Chen, Ling Shao

Multi-modal large language models (MLLMs) depend on in-context learning (ICL) for rapid task adaptation, but their scalability is severely limited by finite context windows and the growing cost of key–value (KV) caches in long multi-modal sequences. Existing memory compression approaches typically rely on rigid token removal or sample-dependent importance estimation, which introduces bias, disrupts semantic structure—particularly for visual representations—and yields static memories that cannot adapt to new queries. We introduce TASM (Task-Aware Structured Memory), a training-free framework that addresses these limitations through task-aware, structure-preserving, and dynamically accessible memory construction. TASM employs Task-Vector Guided Compression to replace sample-specific signals with a task-level direction that captures shared relevance across demonstrations. To preserve the underlying information manifold, it further applies Semantics-Aware Token Merging, formulating compression as a Bipartite Graph Matching problem that merges tokens without destructive pruning. Finally, TASM organizes compressed representations into a multi-resolution hierarchy consisting of a compact Core Memory and a Latent Bank, enabling Query-Adaptive Dynamic Activation and Dynamic Retrieval at inference time. Empirical evaluations show that TASM sustains strong multi-modal ICL performance under high compression ratios, demonstrating an effective balance between efficiency, adaptability, and semantic fidelity.