论文检索

Reinforcement Learning · Batch/Offline

Jiaxin Zhao, Weihang Pan, xun liang, Binbin Lin

Offline policy improvement faces an inherent conflict between maximizing value and fitting the data distribution. While in-sample weighted regression is stable, it suffers from over-conservatism that suppresses high-value actions in the distribution tail; conversely, gradient-based approaches often exhibit a fitting-optimization conflict of gradients, which drive the policy off the data manifold. To address this, we propose Support-Preserving Action Rectification (SPAR), which reframes global learning as a local residual rectification anchored to a frozen pure behavior cloning policy. This framework performs fine-grained fitting and local policy improvement in the residual space, thereby contracting the search space. We further introduce Latent Self-Imitation, utilizing a latent-sampling weighted-regression mechanism to address fitting-improvement gradient conflict in the residual space. Theoretically, we prove this mechanism eliminates the manifold-normal drift of standard value gradients, while extensive D4RL experiments show SPAR extracts significant gains from suboptimal baselines to achieve state-of-the-art performance.

Reinforcement Learning · Multi-agent

Hao Xiang Li, Michael Amir, Amanda Prorok

The agent-environment co-design paradigm jointly optimises agent policies and environment configurations in search of improved system performance, promising to fundamentally reshape how we deploy multi-agent systems in domains such as warehouse logistics and windfarm management. However, current co-design methods collapse under high dimensional environment design spaces and suffer from sample inefficiency when addressing moving targets inherent to joint optimisation. We address this by developing **Diffusion Co-Design** (DiCoDe), a scalable and sample-efficient co-design framework incorporating two core innovations. We introduce Projected Universal Guidance (PUG), enabling exploration of constraint-satisfying reward-maximising environments, and devise a critic distillation mechanism to transfer knowledge from the reinforcement learning loop to a guided diffuision model. Together, these improvements lead to superior environment-policy pairs when validated on challenging multi-agent co-design benchmarks, for example, exceeding state-of-the art in a warehouse setting with 39% higher rewards and 66% fewer simulation steps.

Applications · Everything Else

Marcio Monteiro, Weichen Li, Puyu Wang, Marius Kloft, Sophie Fellenz

Selecting a pretrained large language model (LLM) to fine-tune for a task-specific dataset can be time-consuming and costly. With several candidate models available to choose from, varying in size, architecture, and pretraining data, finding the best model for a specific task often involves extensive trial and error. In addition, the "best" model may not necessarily be the one with the lowest test loss, as practical considerations such as deployment costs, inference throughput, and limited search budgets might also play crucial roles. To address this, we introduce LAMPS (LAnguage Model Pareto Selection), a novel and open-source multi-objective AutoML framework that meta-learns a resource allocation policy to efficiently identify (or approximate) the Pareto front of candidate LLMs for a task-specific dataset. It is based on two key ideas: (1) landmark fine-tuning, which generates early performance indicators of the candidate models, and (2) meta-learning via reinforcement learning, which learns an effective selection policy from historical performance data (a meta-dataset). Our results show that, on held-out datasets, LAMPS reduces search time by an average of 73\% compared to exhaustive search, while still covering more than 99\% of the optimal target space hypervolume.

Deep Learning · Large Language Models

Bowei He, Minda Hu, Zenan Xu, Hongru WANG, Licheng Zong, Yankai Chen, Chen Ma, Xue Liu, PlutoZhou, Irwin King

Search-integrated reasoning enables language agents to transcend static parametric knowledge by actively querying external sources. However, training these agents via reinforcement learning is hindered by the *multi-scale credit assignment* problem: existing methods typically rely on sparse, trajectory-level rewards that fail to distinguish between high-quality reasoning and fortuitous guesses, leading to redundant or misleading search behaviors. To address this, we propose Search-R2, a novel Actor–Refiner collaboration framework that enhances reasoning through targeted intervention, with both components jointly optimized during training. Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories, and a Meta-Refiner, which selectively diagnoses and repairs flawed steps via a ``cut-and-regenerate'' mechanism. To provide fine-grained supervision, we introduce a hybrid reward design that couples outcome correctness with a dense process reward quantifying the information density of retrieved evidence. Theoretically, we formalize the Actor–Refiner interaction as a smoothed mixture policy, proving that selective correction yields strict performance gains over strong baselines. Extensive experiments across various general and multi-hop QA datasets demonstrate that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales, achieving superior reasoning accuracy with minimal overhead.

Reinforcement Learning · Policy Search

Yihan Wang, Peiyu Liu, Runyu Chen, Wei Xu

Text-to-SQL has recently achieved impressive progress, yet remains difficult to apply effectively in real-world scenarios. This gap stems from the reliance on single static workflows, fundamentally limiting scalability to out-of-distribution and long-tail scenarios. Instead of requiring users to select suitable methods through extensive experimentation, we attempt to enable systems to adaptively construct workflows at inference time. Through rigorous theoretical and empirical analysis, we demonstrate that optimal dynamic policies consistently outperform the best static workflow, with performance gains fundamentally driven by heterogeneity across candidate workflows. Motivated by this, we propose SquRL, a reinforcement learning framework that enhances LLMs' reasoning capability in adaptive workflow construction. We design a rule-based reward function and introduce two effective training mechanisms: dynamic actor masking to encourage broader exploration, and pseudo rewards to improve training efficiency. Experiments on widely-used Text-to-SQL benchmarks demonstrate that dynamic workflow construction consistently outperforms the best static workflow methods, with especially pronounced gains on complex and out-of-distribution queries.

Ran Xin, Zeyu Zheng, Yanchen Nie, Kun Yuan, Xia Xiao

The integration of Large Language Models (LLMs) with automated theorem proving has shown immense promise, yet is constrained by challenges in scaling up both training-time reinforcement learning (RL) and inference-time compute. This paper introduces BFS-Prover-V2, a step-level theorem proving system designed to address this dual scaling problem. We present two primary innovations. The first is a novel multi-turn off-policy RL framework for continually improving the performance of the LLM step-prover at training time. This framework, inspired by the principles of AlphaZero, utilizes a multi-stage expert iteration pipeline featuring adaptive tactic-level data filtering and periodic retraining to surmount the performance plateaus that typically curtail long-term RL in LLM-based agents. The second innovation is a planner-enhanced multi-agent system that scales reasoning capabilities at inference time. This architecture employs a general reasoning model as a high-level planner to iteratively decompose complex theorems into a sequence of simpler subgoals. This hierarchical approach substantially reduces the search space, enabling a team of parallel prover agents to collaborate efficiently by leveraging a shared proof cache. We demonstrate that this dual approach to scaling yields state-of-the-art results on established formal mathematics benchmarks. BFS-Prover-V2 achieves 95.08% and 41.4% on the miniF2F and ProofNet test sets respectively. While demonstrated in the domain of formal mathematics, the RL and inference techniques presented in this work are of broader interest and may be applied to other domains requiring long-horizon multi-turn reasoning and complex search.

Deep Learning · Foundation Models

Guangyu Zhao, Kewei Lian, Haoxuan Ru, Borong Zhang, Haowei Lin, Zhancun Mu, Haobo Fu, Qiang Fu, Shaofei Cai, Zihao Wang 等

Goal-conditioned policies enable decision-making models to execute diverse behaviors based on specified goals, yet their downstream performance is often highly sensitive to the choice of instructions or prompts. To bypass the limitations of discrete text prompts, we formulate post-training adaptation as a latent control problem, where the goal embedding serves as a continuous control variable to modulate the behavior of a frozen policy. We propose Preference Goal Tuning (PGT), a framework that optimizes this latent control variable to align the induced trajectory distribution with task preferences. Unlike standard fine-tuning that updates policy parameters, PGT keeps the policy frozen and updates only the latent goal using a trajectory-level preference objective. This approach essentially searches for the optimal conditioning input that maximizes the likelihood of preferred behaviors while suppressing undesirable ones. We evaluate PGT on the Minecraft SkillForge benchmark across 17 tasks. With minimal data, PGT achieves average relative improvements of 72.0\% and 81.6\% on two foundation policies, consistently outperforming expert-crafted prompts. Crucially, by decoupling task alignment (latent goal) from physical dynamics (frozen policy), PGT surpasses full fine-tuning by 13.4\% in out-of-distribution settings, demonstrating superior robustness and generalization.

Applications · Computer Vision

Mohammed Asad Karim, Vinay Kumar Verma

In-context localization (ICL) seeks to localize a target object specified by a small set of support examples in a query image, operating on the fly without training or parameter updates. Despite rapid advances in vision–language models (VLMs), achieving category-agnostic and visually grounded ICL remains an open problem, even though it is essential for applications such as image editing, personalized visual search, and retrieval. Existing methods are fragile and rely on explicit category supervision, which not only limits applicability in realistic settings with unnamed or instance-specific objects but also introduces category bias that steers predictions toward semantic priors rather than visual evidence. We introduce a two-stage training framework that explicitly optimizes in-context attention between support bounding boxes and query images without category supervision. We further refine localization via reinforcement learning using Group Relative Policy Optimization (GRPO) to directly minimize localization error. This formulation enforces visual correspondence over semantic priors, yielding robust instance-level localization. Empirically, a 7B-parameter model trained with our objectives outperforms models up to 72B parameters, demonstrating that context-aware localization objectives can surpass scaling alone. Comprehensive ablations validate the contribution of each component.

Reinforcement Learning · Policy Search

Jiaqi Wang, Haoge Deng, Ting Pan, Yang Liu, Chengyuan Wang, Fan Zhang, Yonggang Qi, Xinlong Wang

Uniform Discrete Diffusion (UDM) has recently emerged as a promising paradigm for discrete generative modeling; however, its integration with reinforcement learning remains largely unexplored. We observe that naively adapting GRPO to UDM leads to unstable training and marginal performance. To address this, we propose \Ours, the first framework that integrates UDM with RL. Our method is guided by two key insights: (i) treating the final clean sample, rather than intermediate predicted sample, as the action provides more accurate and stable optimization signals; and (ii) adopting the forward process to reconstruct the training trajectories helps the model learn probability paths that are more consistent with pretraining. For efficiency, we introduce Reduction-Step and CFG-Free training strategies. \Ours significantly improves the performance of the base model across multiple T2I tasks. Notably, GenEval accuracy improves from $69\%$ to $96\%$ and PickScore increases from $20.46$ to $23.81$, achieving state-of-the-art performance in both continuous and discrete settings. On the OCR benchmark, accuracy improves from $4\%$ to $57\%$, further validating the effectiveness and generalization capability of our method.

Reinforcement Learning · Policy Search

Álvaro Serra-Gómez, Daniel Jarne Ornia, Dhruva Tirumala, Thomas M Moerland

Effective exploration remains a key challenge in model-based reinforcement learning (MBRL), especially in high-dimensional continuous control tasks where sample efficiency is critical. Recent work addresses this by using learned policies as proposal distributions for Model-Predictive Path Integral (MPPI) planning. Early approaches update the sampling policy independently of the planner, typically via deterministic policy gradients with entropy regularization. However, since the data distribution is induced by the MPPI planner, misalignment between the policy and planner degrades value estimation and long-term performance. To address this, recent methods explicitly align the policy with the planner by minimizing KL divergence to the planner distribution or by incorporating planner-guided regularization. In this work, we unify these approaches under the Policy Optimization–Model Predictive Control (PO-MPC) framework, a family of KL-regularized MBRL methods that treat the planner’s action distribution as a prior in policy optimization. We show how existing methods emerge as special cases of this family and explore previously unstudied variants. Experiments demonstrate that these variants yield significant performance gains, advancing the state of the art in MPPI-based RL.

Deep Learning · Generative Models and Autoencoders

Parsa Rahimi, Sébastien Marcel

Synthetic data generation is increasingly used in machine learning for **training and data augmentation**. Yet, many current strategies rely on external foundation models or datasets, which can be restricted by policy or legal constraints, especially for sensitive modalities such as human face images and videos. We propose **ScoreMix**, a **self-contained data augmentation** method to boost recognition performance by leveraging score compositionality in class-conditioned diffusion models. ScoreMix mixes class-conditioned scores along reverse diffusion trajectories, yielding domain-specific hard augmentations without external resources. We systematically study class-selection strategies and find that mixing classes that are distant in the discriminator embedding space yields larger gains, providing **up to 3\% additional average improvement across benchmarks** over proximity-based selection. Interestingly, we observe that learned condition and embedding spaces are largely uncorrelated under standard alignment metrics, and that condition-space distances are weakly correlated to downstream gains. Across **8 public face recognition benchmarks**, ScoreMix improves accuracy by **up to 7 percentage points** without hyperparameter search, highlighting robustness and practicality. Code and dataset will be made publicly available.

Optimization · Zero-order and Black-box Optimization

Yanchi Li, Jiao Liu, Wenyin Gong, Qiong Gu, Yue Zhao, Yew Soon ONG

Evolutionary multi-tasking accelerates black-box optimization via knowledge transfer but falters in scenarios involving many low-similarity tasks. We identify this scalability barrier as the *Multi-Task Curse*, driven by evaluation budget dispersion and negative transfer. To overcome this, we propose MES-RET (*M*any-task *E*volution *S*trategy with *R*eward-weighted *E*valuation and *T*ransfer), which combats budget dispersion via a reward-weighted evaluation scheme that guarantees superior expected improvement, while simultaneously mitigating negative transfer through a robust reward-weighted aggregation of mean and covariance statistics, ensuring a safe fallback to independent evolution. Furthermore, to handle neural dimensional mismatches in many-task policy search, we introduce a semantic parameter alignment strategy that bridges heterogeneous state-action spaces. Extensive experiments on synthetic benchmarks, real-world engineering problems, and reinforcement learning tasks demonstrate that MES-RET consistently outperforms state-of-the-art methods, notably enabling skill transfer across morphologically distinct policies.

Deep Learning · Large Language Models

Jiawei Wang, Jiacai Liu, Yuqian Fu, Yingru Li, Xintao Wang, Yuan Lin, Lin Zhang, YuYue, Yang Wang, WANG KE

In long-horizon tasks, recent agents based on Large Language Models (LLMs) face a significant challenge that sparse, outcome-based rewards make it difficult to assign credit to intermediate steps. Previous methods mainly focus on creating dense reward signals to guide learning, either through traditional reinforcement learning techniques like inverse reinforcement learning or by using Process Reward Models for step-by-step feedback. In this paper, we identify a fundamental problem in the learning dynamics of LLMs: the magnitude of policy gradients is inherently coupled with the entropy, which leads to inefficient small updates for confident correct actions and potentially destabilizes large updates for uncertain ones. To resolve this, we propose Entropy-Modulated Policy Gradients (EMPG), a framework that recalibrates the learning signal based on step-wise uncertainty and the final task outcome. EMPG amplifies updates for confident correct actions, penalizes confident errors, and attenuates updates from uncertain steps to stabilize exploration. We further introduce a bonus term for future clarity that encourages agents to find more predictable solution paths. Through comprehensive experiments on three challenging agent tasks, WebShop, ALFWorld, and Deep Search, we demonstrate that EMPG achieves substantial performance gains and significantly outperforms strong policy gradient baselines.

Reinforcement Learning · Policy Search

Jonathan Colaco Carr, Prakash Panangaden, Doina Precup, Benjamin Van Roy

Agents that can beat or tie any other under a model of pairwise preference have strong guarantees for both user satisfaction and overall social welfare. However, searching for these agents in long-term decision problems is not computationally tractable with current approaches, which require the size of an agent's policy to increase with the problem length. We introduce the \textit{Markov decision contest}, a model of learning from general preferences in long-term (infinite-horizon) decision problems. Within this model, we prove that agents only need a stationary Markov policy in order to be optimal (that is, to beat or tie any agent with a history-dependent policy); that the problem of finding an optimal policy is in P; and that a simple iterative algorithm (which we call Hedged Policy Iteration) converges to an optimal policy at a sublinear rate. In a suite of high-dimensional experiments, we demonstrate that Hedged Policy Iteration scales well to function approximation. Lastly, we present a near approximation of Hedged Policy Iteration, called HPI-Clip, which both matches the performance of Proximal Policy Optimization on reward-based tasks while also outperforming it on tasks with non-transitive preferences. These results show that learning from pairwise preferences in long-term decision problems can be far more tractable than what is known from prior work.

Applications · Chemistry, Physics, and Earth Sciences

Liao Chang, Luotian Yuan, Yiping Ke, Ying Wei

Multi-step retrosynthesis planning is a fundamental challenge in organic chemistry, defined by its enormous search space. Existing methods typically formulate it as a Markov Decision Process (MDP) with a fixed choice of transition model (i.e., a single-step retrosynthesis model), and focus on improving *how to search* through better policies and value functions. However, *how the transition space itself is navigated* remains largely unexplored. This limitation is particularly urgent given our observation of pronounced *skill disparity* among single-step prediction models: different models exhibit substantially different performance across molecule states. Motivated by this observation, we introduce RetrOrchestrator, an LLM-powered agent that explicitly accounts for model skill disparity by reframing retrosynthesis planning as a Partially Observable Markov Decision Process (POMDP). By regarding each single-step prediction model as a tool, we further propose a scaffold-aware reinforcement learning algorithm to optimize navigation policy within the transition space. As a result, RetrOrchestrator jointly searches which molecule to expand and which single-step model to apply for the molecule at the current step. Empirically, RetrOrchestrator significantly outperforms static baselines on the Retro*-190 benchmark, achieving a state-of-the-art 94.21\% success rate as well as a Pareto front in both wallclock time and number of model queries.

Deep Learning · Large Language Models

Yash Akhauri, Mohamed Abdelfattah

Efficient LLM inference research has largely focused on reducing the cost of each decoding step (e.g., using quantization, pruning, or sparse attention), typically applying a uniform computation budget to every generated token. In practice, token difficulty varies widely, so static compression can over-compute on easy steps and under-compute on hard ones. We study **dynamic budget allocation** for autoregressive decoding: learning how much computation to spend **per token** from within a single model. Self-Optimizing Language Models (SOL) pair a frozen LLM with a lightweight policy network that reads the LLM hidden state and selects a discrete **efficiency action** at each decode step. Actions can jointly control (i) token-level attention sparsity, (ii) structured activation pruning in the MLP, and (iii) activation quantization bit-width, while leaving the base model weights unchanged. We train the policy with group-relative policy optimization on teacher-forced episodes: the token sequence is fixed, while we sample multiple compute schedules (i.e., “counterfactual” schedules that vary only the efficiency actions for the same token path) and compare their likelihoods under the same supervision. Our reward trades off language‑model quality against soft penalties that encourage episode‑average budget usage to match a requested target. Across model variants and compute regimes, SOL improves quality at matched budget over static allocation and strong random schedule search, offering a complementary axis for inference‑efficiency optimization. SOL discovers a better quality-efficiency pareto-front across all our experiments and improves MMLU accuracy by upto 7.3\% over uniform budget allocation strategies.

Deep Learning · Large Language Models

Zhiwei Ning, Xuanang Gao, Jiaxi Cao, Gengming Zhang, Shengnan Ma, Wenwen Tong, Hanming Deng, JIE YANG, Wei Liu

Multimodal large language models (MLLMs) have achieved remarkable success in general perception, yet complex multi-step visual reasoning remains a persistent challenge. Although recent agentic approaches incorporate tool use, they often neglect critical execution feedback. Consequently, they suffer from the imagination-action-observer (IAO) bias, a misalignment between prior imagination and observer feedback that undermines reasoning stability and optimality. To bridge this gap, we introduce V-ABS, an action-observer driven beam search framework that enables deliberate reasoning through thinker-actor-observer iterations. We also propose an entropy-based adaptive weighting algorithm to mitigate the IAO bias by dynamically balancing the confidence scores between the policy priors and the observational feedback. Moreover, we construct a large-scale supervised fine-tuning (SFT) dataset comprising over 80k samples to guide the model to assign higher prior confidence to correct action paths. Extensive experiments across eight diverse benchmarks show that V-ABS achieves state-of-the-art performance, delivering an average improvement of 19.7\% on the Qwen3-VL-8B baseline and consistent gains across both open-source and proprietary models.

Deep Learning · Generative Models and Autoencoders

Mingkun Lei, Tong Zhao, Liangyu Yuan, Chi Zhang

Step-level caching offers a promising avenue for accelerating diffusion models by exploiting temporal redundancy. However, existing strategies predominantly rely on heuristic, threshold-based metrics to trigger cache updates. This reactive paradigm is inherently myopic as it optimizes only for local feature consistency, and yields unpredictable deployment latency. In this work, we propose BudCache, a budget-constrained optimization framework that inverts this standard: instead of letting error thresholds dictate the cost, we enforce a strict computational budget and globally search for the caching policy that maximizes generation fidelity. To tackle the combinatorial complexity of step selection, we employ a hybrid strategy combining Simulated Annealing with deterministic Hill Climbing. This approach efficiently escapes local optima to locate globally optimized cache masks within minutes, incurring zero inference overhead. Crucially, to address the trajectory drift induced by aggressive caching, we introduce a cache-aware schedule alignment mechanism. By refining the time discretization via a lightweight, data-free distillation, we significantly enhance performance in low-NFE regimes. Extensive experiments on FLUX.1-dev and Wan2.1 demonstrate that BudCache consistently outperforms heuristic baselines, achieving superior perceptual quality under rigid latency constraints.

Reinforcement Learning · Batch/Offline

Hongqiang Lin, Pengfei Wang, Nenggan Zheng

Offline reinforcement learning (RL) aims to optimize policies from pre-collected datasets. A bottleneck of this paradigm is managing epistemic uncertainty, which arises from limited data coverage (sample-level) and the ambiguity in identifying transition dynamics from finite data (model-level). To provide a unified quantification of these uncertainties, Bayesian RL has been proposed by treating the dynamics model as a random variable and maintaining a corresponding belief. Despite its theoretical appeal, policy optimization in Bayesian RL remains computationally challenging as it requires solving composite objectives with expectations. Prior methods either employ search-based techniques with poor computational scalability or impose restrictive posterior assumptions that sacrifice the adaptability of Bayesian RL. To address these limitations, we propose Posterior Hybrid Bayesian Belief (PhyB), which reformulates the expectation as a convex combination over a subset of dynamics models. Theoretical analysis demonstrates that the objective discrepancy induced by this approximation remains bounded. Based on PhyB, we develop an iterative regularized policy optimization algorithm that provides metric-agnostic guarantees for monotonic improvement until convergence. Empirical results demonstrate that PhyB achieves state-of-the-art performance on various benchmarks.

Applications · Robotics

Kaixin Chai, Hyunjun Lee, Joseph Lim

Determining where to execute the manipulation policy is a fundamental challenge in mobile manipulation. Most approaches have formulated this as a geometric search problem, prioritizing physical reachability. However, given the high sensitivity of modern learning-based manipulation policies, geometric criteria alone are insufficient. Optimal performance requires base positioning that is aware of the policy's preference. While recent works have attempted to address this, they remain limited in practicality due to reliance on pre-built scene reconstruction and slow inference. In this work, we introduce N2M that systematically reformulates the approach to base positioning problem, naturally overcoming limitations of previous methods. Our key insight is that policy preferences are inherent to the local scene structure and can be effectively learned from the policy rollouts. Technically, we propose a novel _viewpoint augmentation_ strategy that enables the model to learn robust, viewpoint-invariant pose preferences with remarkable data efficiency. Extensive experiments demonstrate that N2M achieves state-of-the-art performance, outperforming both non-policy-aware baselines and recent policy-aware alternatives. Furthermore, we provide a comprehensive analysis highlighting N2M’s broad applicability, generalization capabilities, and data efficiency. Anonymized project website: https://nav2manip.github.io