论文检索

Deep Learning · Large Language Models

YIPING SONG, Jinyu You, Zhiliang Tian, Jinsong Su, Minlie Huang, Chenping Hou

Auto-Regressive (AR) models with Monte Carlo Tree Search (MCTS) are a dominant paradigm for achieving “System 2” reasoning. However, this approach suffers from significant latency due to the serial, token-by-token generation mechanism of AR models. In contrast, Diffusion Large Language Models (dLLMs) offer inherent speed advantages via parallel sequence generation, yet they often struggle with accuracy in complex reasoning due to a lack of rigorous search, evaluation, and revision capabilities. Directly applying MCTS to diffusion models faces architectural barriers, since the denoising generation process lacks the discrete decision steps that naturally accommodate tree search. To retain efficiency while improving the reasoning ability, we propose DiffuReason, a Monte Carlo tree search reasoning algorithm for diffusion models. By modeling the generation process as a Markov Decision Process (MDP), DiffuReason discretizes the continuous diffusion flow into searchable thought blocks. During the reverse generation process, DiffuReason recursively performs four MCTS-style stages: select the best node (block), expand to obtain candidate nodes, simulate to evaluate node values, and revise the unsatisfactory nodes. Experiments on mathematical reasoning benchmarks demonstrate that DiffuReason significantly improves the reasoning ability of diffusion models, and achieves superior balance of accuracy and efficiency even compared with auto-regressive models.

General Machine Learning · Evaluation

Austin Letson, Leopoldo Sarra, Auguste Poiroux, Oliver Dressler, Paul Lezeau, Dhyan Aranha, Frederick Pu, Aaron Hill, Miguel Hidalgo, Julian Berman 等

We present SorryDB, a dynamically-updating benchmark of open Lean tasks drawn from 78 real world formalization projects on GitHub. Unlike existing static benchmarks, often composed of competition problems, hillclimbing the SorryDB benchmark will yield tools that are aligned to the community needs, more usable by mathematicians, and more capable of understanding complex dependencies. Moreover, by providing a continuously updated stream of tasks, SorryDB mitigates test-set contamination and offers a robust metric for an agent's ability to contribute to novel formal mathematics projects. We evaluate a collection of approaches, including generalist large language models, agentic approaches, and specialized symbolic provers, over a selected snapshot of 1000 tasks from SorryDB. We show that current approaches are complementary: even though an agentic approach based on Gemini Flash is the most performant, it is not strictly better than other off-the-shelf large-language models, specialized provers, or even a curated list of tactics.

Deep Learning · Robustness

Yao Ni, Jeremie Houssineau, Yew Soon ONG, Piotr Koniusz

Deep neural networks achieve impressive results across diverse applications, yet their overconfidence on unseen inputs necessitates reliable epistemic uncertainty modelling. Existing methods for uncertainty modelling face a fundamental dilemma: Bayesian approaches provide principled estimates but remain computationally prohibitive, while efficient second-order predictors lack rigorous derivations connecting their specific objectives to epistemic uncertainty quantification. To resolve this dilemma, we introduce **D**irichlet-**a**pproximated **p**ossibilistic **p**osterior p**r**edictions (DAPPr), a principled framework leveraging possibility theory. We define a possibilistic posterior over parameters, projects this posterior to the prediction space via supremum operators, and approximates the projected posterior using learnable Dirichlet possibility functions. This projection-and-approximation strategy yields a simple training objective with closed-form solutions. Extensive experiments across diverse benchmarks demonstrate that our approach achieves competitive or superior uncertainty quantification performance compared to state-of-the-art evidential deep learning methods while maintaining both principled derivation and computational efficiency. Code will be available upon acceptance.

Probabilistic Methods · Bayesian Models and Methods

Qian Xie, Linda Cai, Alexander Terenin, Peter Frazier, Ziv Scully

In automated machine learning, scientific discovery, and other applications of Bayesian optimization, deciding when to stop evaluating expensive black-box functions in a cost-aware manner is an important but underexplored practical consideration. A natural performance metric for this purpose is the cost-adjusted simple regret, which explicitly captures the trade-off between solution quality and cumulative evaluation cost. Existing stopping rules for Bayesian optimization are either heuristic, or are theoretically grounded but designed to optimize simple regret without accounting for evaluation costs; as a result, they provide no guarantees against unnecessary evaluations when costs are high. We propose a principled cost-aware stopping rule for Bayesian optimization that adapts to varying evaluation costs without heuristic tuning. Our rule is grounded in a theoretical connection to state-of-the-art cost-aware acquisition functions, namely the Pandora's Box Gittins Index (PBGI) and log expected improvement per cost (LogEIPC). When paired with either acquisition function, we prove that the resulting policy satisfies a theoretical guarantee bounding the expected cost-adjusted simple regret. Across synthetic tasks and empirical benchmarks including hyperparameter optimization and neural architecture size search, pairing our stopping rule with PBGI or LogEIPC usually matches or outperforms other acquisition-function--stopping-rule pairs in terms of cost-adjusted simple regret.

Applications · Robotics

Kangye Ji, Yuan Meng, Jianbo Zhou, Ye Li, Hanyun Cui, Zhi Wang

Diffusion Policy has dominated action generation due to its strong capabilities for modeling multi-modal action distributions, but its multi-step denoising processes make it impractical for real-time visuomotor control. Existing caching-based acceleration methods typically rely on $\textit{static}$ schedules that fail to adapt to the \textit{dynamics} of robot-environment interactions, thereby leading to suboptimal performance. In this paper, we propose $\underline{\textbf{S}}$parse $\underline{\textbf{A}}$ction$\underline{\textbf{G}}$en $(\textbf{SAG})$ for extremely sparse action generation. To accommodate the iterative interactions, SAG customizes a rollout-adaptive prune-then-reuse mechanism that first identifies prunable computations globally and then reuses cached activations to substitute them during action diffusion. To capture the rollout dynamics, SAG parameterizes an observation-conditioned diffusion pruner for environment-aware adaptation and instantiates it with a highly parameter- and inference-efficient design for real-time prediction. Furthermore, SAG introduces a one-for-all reusing strategy that reuses activations across both timesteps and blocks in a zig-zag manner, minimizing the global redundancy. Extensive experiments on multiple robotic benchmarks demonstrate that SAG achieves up to 4$\times$ generation speedup without sacrificing performance. Project Page: https://sparse-actiongen.github.io/.

Deep Learning · Large Language Models

Haolin Deng, Xin Zou, Zhiwei Jin, Chen Chen, Haonan Lu, Xuming Hu

Multimodal hallucination remains a persistent challenge for Vision-Language Models (VLMs). Standard textual Direct Preference Optimization (DPO) often fails to mitigate it due to a lack of explicit visual supervision. While existing works introduce visual preference DPO by contrasting original images against negative ones, they suffer from a theoretically inconsistent objective caused by partition function mismatches and relies on coarse-grained negatives that could enable shortcut learning. In this work, we propose In-Context Visual Contrastive Optimization (IC-VCO). By placing contrastive images within a shared multi-image context, IC-VCO ensures a mathematically rigorous objective. Furthermore, we introduce Visual Contrast Distillation (VCDist), a mechanism which transfers the superior discriminatory power of the multi-image context to the single-image policy via reliability-gated self-distillation, enforcing consistent visual grounding. Finally, we propose a contrastive sample editing strategy that generates hard negatives via precise semantic perturbations. Experiments on five benchmarks demonstrate IC-VCO's superior performance and the effectiveness of our sample editing strategy.

Social Aspects · Robustness

Konstantin Kaulen, Hadar Shavit, Holger Hoos

Deep neural networks achieve strong performance on many supervised learning tasks but remain vulnerable to adversarial perturbations. Neural network verification provides mathematically rigorous robustness guarantees, yet at substantial computational cost. To mitigate this, certified training techniques optimise for verifiable robustness during training, typically inducing a trade-off between natural and certified accuracy controlled by method-specific hyperparameters. Because these metrics are inherently conflicting, the common practice of reporting a single configuration is problematic: it can mislead conclusions about overall performance and prevents unbiased assessments of the state of the art. We address this by evaluating certified training methods via Pareto front comparisons over the natural--certified accuracy trade-off. To enable fair, method-agnostic comparisons, we perform efficient automated multi-objective hyperparameter optimisation to identify a set of Pareto-optimal configurations for each method. This approach often uncovers substantial undertuning in previously reported configurations, yielding superior performance and establishing a new state of the art. Leveraging these fronts, we present the first comprehensive multi-objective comparison of certified training approaches, showing that prior advancements are less pronounced than assumed and revealing previously unreported performance complementarities.

Deep Learning · Large Language Models

Tsz Ting Chung, Lemao Liu, Mo Yu, Dit-Yan Yeung

In-context learning (ICL) adapts large language models (LLMs) to new tasks by conditioning on demonstrations in the prompt without parameter updates. With long-context models, many-shot ICL can use dozens to hundreds of examples and achieve performance comparable to fine-tuning, yet current understanding of its scaling behavior is largely derived from non-reasoning tasks. We study many-shot chain-of-thought in-context learning (CoT-ICL) for reasoning and show that standard many-shot rules do not transfer. Across non-reasoning and reasoning-oriented LLMs and across non-reasoning and reasoning tasks, we find: (i) a setting-dependent scaling effect, where increasing the number of CoT demonstrations is unstable for non-reasoning LLMs and benefits mainly reasoning-oriented LLMs; (ii) similarity-based retrieval helps on non-reasoning tasks but fails on reasoning, since semantic similarity poorly predicts procedural (i.e., CoT) compatibility; and (iii) an order-scaling effect, where performance variance grows with more CoT demonstrations. We interpret these behaviors by viewing many-shot CoT-ICL as in-context test-time learning rather than scaled pattern matching, and suggests two principles: (i) demonstrations should be easy for the target model to understand, and (ii) they should be ordered to support a smooth conceptual progression. Guided by the principle, we propose Curvilinear Demonstration Selection (CDS), a simple ordering method that yields an average 3.81\% gain across math and narrative reasoning tasks. Overall, our results reframe the long context window from a retrieval buffer into a structured curriculum for in-context test-time learning.

Applications · Neuroscience, Cognitive Science

Dennis Wu, Yi-Chun Hung, Braden Yuille, James Fitzgerald, Han Liu

Neural population geometry shapes downstream inference. Recent findings in neurobiology suggest that a hyperbolic structure underlies population activity. However, a theoretical framework for this phenomenon is still lacking. Here, we propose a plausible construction of hippocampal tuning curves that statistically induce hyperbolic geometry. Next, we establish a connection between neural decoding and associative memory by demonstrating that the Modern Hopfield Network update rule computes the optimal squared loss estimator under hyperbolic geometry. Furthermore, we introduce a novel associative memory model defined in hyperbolic space that yields significantly larger capacity than existing models. Our results suggest that animals encode spatial information as a latent hyperbolic cognitive map, which enhances both memory capacity and decoding accuracy.

Reinforcement Learning · Multi-agent

Qinhong Zhou, Chuang Gan, Anoop Cherian

Embodied agents operating in decentralized and partially observable environments have attracted growing attention in recent years. However, existing large language model (LLM)–based agents often exhibit behaviors that are misaligned with their partners or inconsistent with the environment state, leading to inefficient cooperation and poor task success. To address this challenge, we propose a novel framework, Learning Laws for Cooperation (LLawCo), that enables embodied agents to autonomously align with both their partners and task objectives. Our framework allows agents to reflect on past failures to extract misaligned behavioral patterns, which are used to derive high-level behavioral laws (e.g., “Talk when necessary”, “Wait for partner”). These laws are explicitly incorporated into the agents’ chains of thought via supervised fine-tuning, aligning their reasoning with task requirements and the behavior of other agents. To evaluate our approach, we introduce PARTNR-Dialog, a large-scale multi-agent communicative and cooperative planning benchmark built on the PARTNR environment. Experiments on existing tasks and our new benchmark demonstrate significant improvements in cooperative efficiency and task success rates. Across four backbone LLMs, our method achieves average success rate improvements of 4.5% on the PARTNR-Dialog benchmark and 6.8% on the TDW-MAT benchmark over state-of-the-art open-source communicative agent frameworks.

Theory · Reinforcement Learning and Planning

David Bossens, Atsushi Nitanda

Safety is an essential requirement for reinforcement learning systems. The newly emerging framework of robust constrained Markov decision processes allows learning policies that satisfy long-term constraints while providing guarantees under epistemic uncertainty. This paper presents mirror descent policy optimisation for robust constrained Markov decision processes, making use of policy gradient techniques to optimise both the policy (as a maximiser) and the transition kernel (as an adversarial minimiser) on the Lagrangian representing a constrained Markov decision process. Our proposed algorithm obtains an $\tilde{\mathcal{O}}\left(1/T^{1/3}\right)$ convergence rate in the sample-based robust constrained Markov decision process setting. The paper also contributes an algorithm for approximate gradient descent in the space of transition kernels, which is of independent interest for designing adversarial environments in general Markov decision processes. Experiments confirm the benefits of mirror descent policy optimisation in constrained and unconstrained optimisation, and significant improvements are observed in robustness tests when compared to baseline policy optimisation algorithms.

Deep Learning · Sequential Models, Time series

Matteo Pinna, Giacomo Lagomarsini, Andrea Ceni, Claudio Gallicchio

Reservoir Computing (RC) has established itself as an efficient paradigm for temporal processing. However, its scalability remains severely constrained by (i) the necessity of processing temporal data sequentially and (ii) the prohibitive memory footprint of high-dimensional reservoirs. In this work, we revisit RC through the lens of structured operators and state space modeling to address these limitations, introducing Parallel Echo State Network (ParalESN). ParalESN enables the construction of high-dimensional and efficient reservoirs based on diagonal linear recurrence in the complex space, enabling parallel processing of temporal data. We provide a theoretical analysis demonstrating that ParalESN preserves the Echo State Property and the universality guarantees of traditional Echo State Networks while admitting an equivalent representation of arbitrary linear reservoirs in the complex diagonal form. Empirically, ParalESN matches the predictive accuracy of traditional RC on time series benchmarks, while delivering substantial computational savings. On 1-D pixel-level classification tasks, ParalESN achieves competitive accuracy with fully trainable neural networks while reducing computational costs and energy consumption by orders of magnitude. Overall, ParalESN offers a promising, scalable, and principled pathway for integrating RC within the deep learning landscape.

Applications · Social Sciences

Maximilian Puelma Touzel, Sneheel Sarangi, Aurélien Bück-Kaeffer, Zachary Yang, Jean-François Godbout, Reihaneh Rabbany

LLM-based social simulations—in which many language model agents interact over multiple turns—are rapidly proliferating across policy analysis, epidemiology, and computational social science. Yet the field lacks consensus on how to validate these simulations, with evaluation methods that are sparse, inconsistent, and rarely shared across disciplinary silos. We argue this creates a serious risk: premature deployment of unvalidated simulators in high-stakes domains. Our position is that the field must pivot from expansion to consolidation, prioritizing methodological standardization—shared benchmarks, open data, and reproducible evaluation protocols grounded in social science and complex systems research. We outline a concrete research program organized around specific learning problems/benchmarks, providing a path toward answering the fundamental question: when are LLM social simulations useful modelling objects?

Social Aspects · Safety

Khang Tran, Yazan Boshmaf, Issa Khalil, Hai Phan, Ting Yu, Md Rizwan Parvez

Code Large Language Models (CLLMs) serve as the core of modern code agents, enabling developers to automate complex software development tasks. In this paper, we present Poison-with-Style (PwS), a practical and stealthy model poisoning attack targeting CLLMs. Unlike prior attacks that assume an active adversary capable of directly embedding explicit triggers (e.g., specific words) into developers' prompts during inference, PwS leverages developers' code styles as covert triggers implicitly embedded within their prompts. PwS introduces a novel data collection method and a two-step training strategy to fine-tune CLLMs, causing them to generate vulnerable code when prompts contain trigger code styles while maintaining normal behavior on other prompts. Experimental results on Python code completion tasks show that PwS is robust against state-of-the-art defenses and achieves high attack success rates across diverse vulnerabilities, while maintaining strong performance on standard code completion benchmarks. For example, in code completion tasks that are vulnerable to improper input validation (i.e., CWE-20), the poisoned model generates insecure code up to 95\% of the cases when the trigger code style is used, with less than $5\%$ drop in pass@1 performance on the HumanEval and MBPP benchmarks.

Social Aspects · Safety

Seokil Ham, Jaehyuk Jang, Wonjun Lee, Changick Kim

Fine-tuning-as-a-Service (FaaS) enables personalization of large language models (LLMs) but poses significant safety risks, as fine-tuning user-provided data degrades the model's safety-alignment. Prior works addressing this issue typically rely on explicit regularization, which leads to practical limitations. In this paper, we propose a different paradigm that neutralizes harmful updates via harmful gradient saturation rather than explicit suppression. Our key observation is that, in a jailbroken LLM, safety-degrading gradients are largely saturated, while gradients unrelated to safety remain active during fine-tuning. Based on this insight, we introduce a **BufferLoRA-based fine-tuning framework**. BufferLoRA is a removable adapter that temporarily jailbreaks the model during user fine-tuning, saturating harmful updates while allowing a UserLoRA to learn user-specific tasks. After fine-tuning, BufferLoRA is removed to restore the base model’s original safety-alignment. To further reinforce safety, we additionally train a SafetyLoRA and integrate its safety components into the UserLoRA via QR decomposition-based merging strategy. Extensive experiments show that our framework achieves superior performance in both safety and utility, without requiring additional safety data during fine-tuning and with minimal computational cost.

Deep Learning · Other Representation Learning

Sigurd Gaukstad, Erik Hermansen, Valdemar Kargård Olsen, Melvin Vaupel, Benjamin Dunn

Deep neural networks learn representations where individual features often lack interpretable meaning; a single neuron may activate for scattered, unrelated inputs. We introduce coherence, a geometric property inspired by neural coding in the brain, where neurons like grid cells and head direction cells respond to contiguous regions of state space. A non-negative matrix is coherent if both each row (sample) attends to geometrically clustered columns (features) and, vice versa, each feature attends to clustered samples. We prove that coherent matrices induce a bounded interleaving between the Vietoris-Rips filtrations of samples and features, guaranteeing that both spaces share compatible topological structure. This geometric constraint facilitates interpretability. For example, if data lies on a circle, coherent features must tile that circle into contiguous arcs. We introduce COH, a differentiable regularizer based on Fréchet variance that enforces coherence during training. Unlike sparsity, which bounds how many samples a feature activates on, coherence bounds *which* samples, requiring geometric connectivity rather than only rarity. This yields not just interpretable features but an interpretable feature space. We validate COH using synthetic and rotated MNIST datasets.

Reinforcement Learning · Planning

Jiangweizhi Peng, Yuanxin Liu, Ruida Zhou, Charles Fleming, Zhaoran Wang, Alfredo Garcia, Mingyi Hong

Training LLMs as interactive agents for multi-turn decision-making remains challenging, particularly in long-horizon tasks with sparse and delayed rewards, where agents must execute extended sequences of actions before receiving meaningful feedback. Most existing reinforcement learning (RL) methods model LLM agents as *flat* policies operating at a single time scale, selecting an action at each turn. In sparse-reward settings, this forces the agent to infer long-range dependencies solely from distant end-of-trajectory signals, often leading to inefficient learning and unstable behavior in complex environments. We propose **HiPER**, a novel **Hi**erarchical **P**lan–**E**xecute **R**L framework that jointly models and optimizes high-level subgoal planning and low-level action execution for LLM agents to overcome flat RL's brittle long-horizon behavior and weak credit assignment under sparse outcome feedback. By maintaining persistent subgoals across multiple turns and explicitly deciding when to switch between them, HiPER introduces structured intermediate decision points that facilitate learning under sparse feedback, converting implicit multi-turn structure into learnable decisions at different time scales. To enable effective training, we introduce Hierarchical Advantage Estimation (HAE), a two-timescale policy gradient method that assigns credit to both action execution and subgoal transitions and achieves variance reduction relative to flat advantage estimation. Empirically, HiPER achieves state-of-the-art performance on challenging interactive benchmarks, reaching 97.4\% success on ALFWorld (+6.6\% over the best prior method) and 83.3\% on WebShop, with especially large gains on long-horizon tasks requiring multiple dependent subtasks. These results highlight the importance of explicit hierarchical decomposition for scalable RL training of multi-turn LLM agents.

Social Aspects · Alignment

Matthew Bozoukov, Matthew Nguyen, Dani Roytburg, Jou Barzdukas, Mackenzie Puig-Hall, Narmeen Oozeer

Recent research has shown that large language models (LLM) favor own outputs when acting as judges, undermining the integrity of automated post-training and evaluation workflows. However, it is difficult to disentangle which evaluation biases are explained by narcissism versus general experimental confounds, distorting measurements of self-preference bias. We discover a core methodological confound which could reduce measurement error by **89.6%**. Specifically, LLM evaluators may deliver self-preferring verdicts when the judge responds to queries which they completed incorrectly themselves; this would be true regardless of whether one of their responses is their own. To decouple self-preference signals from noisy outputs on hard problems, we introduce an **Evaluator Quality Baseline**, which compares the probability that a judge incorrectly votes for itself against the probability that it votes for an incorrect response from another model. Evaluating this simple baseline on 37,448 queries, only **51%** of initial findings retain statistical significance. Finally, we turn towards characterizing the entropy of "easy" versus "hard" evaluation votes from LLM judges. Our corrective baseline enables future research on self-preference by eliminating noisy data from potential solutions. More widely, this work contributes to the growing body of work on cataloging and isolating judge-bias effects

Applications · Computer Vision

Jundong Zhou, Tianao Cai, Yujie Huang, Xinbing Wang, Guang-Zhong Yang, Nanyang Ye

Large Vision–Language Models (LVLMs) have demonstrated remarkable capabilities across diverse multimodal tasks, yet their high inference costs necessitate low-bit deployment. Existing post-training quantization (PTQ) pipelines primarily adopt methodologies from text-only LLMs by treating multimodal inputs as homogeneous sequences, overlooking the heterogeneous information density inherent in LVLMs. In this work, we present ImpQuant, an importance-aware PTQ framework tailored for LVLMs that mitigates low-bit accuracy degradation via fine-grained token-importance reweighted calibration and outlier-aware activation quantization. Our key insight is that quantization errors on decision-critical tokens disproportionately impact overall model behavior. Accordingly, we reweight the calibration loss using aggregated attention for textual tokens and a contextual redundancy metric for visual tokens, respectively. Across multiple LVLM backbones and diverse multimodal benchmarks, our approach consistently improves accuracy at low bitwidth and reduces quantization-induced object hallucinations compared to state-of-the-art PTQ baselines.

Deep Learning · Large Language Models

Shankar Padmanabhan, Mustafa Omer Gul, Tanya Goyal

Post-training endows pretrained LLMs with a variety of desirable skills, such as instruction-following, reasoning, and others. However, these post-trained LLMs only encode knowledge up to a cut-off date, necessitating continual adaptation. Unfortunately, existing solutions cannot effectively learn new knowledge from adaptation document corpora and simultaneously mitigate the forgetting of earlier learned capabilities. To address this, we introduce Distillation via Split Contexts (DiSC), a simple context-distillation based approach for continual knowledge adaptation. DiSC derives student and teacher distributions by conditioning on distinct segments of the training example and minimizes the KL divergence between them for the common tokens. This insight allows us to efficiently apply context-distillation without requiring explicit generation steps during training. We run experiments on three post-trained models and two adaptation domains. Compared to prior finetuning and distillation methods for continual adaptation, DiSC consistently reports the best trade-off between learning new knowledge and mitigating forgetting of previously learned skills like instruction-following and reasoning, or factual knowledge.