arXiv subjects

Mengyue Yang

Publications and source records attributed to Mengyue Yang.

At least 37 records · Page 2Linked to original sources

arXiv2026

The Landscape of Agentic Reinforcement Learning for LLMs: A Survey

The emergence of agentic reinforcement learning (Agentic RL) marks a paradigm shift from conventional reinforcement learning applied to large language models (LLM RL), reframing LLMs from passive sequence generators into autonomous, decision-making agents embedded in complex, dynamic worlds. This survey formalizes this conceptual shift by contrasting the degenerate single-step Markov Decision Processes (MDPs) of LLM-RL with the temporally extended, partially observable Markov decision processes (POMDPs) that define Agentic RL. Building on this foundation, we propose a comprehensive twofold taxonomy: one organized around core agentic capabilities, including planning, tool use, memory, reasoning, self-improvement, and perception, and the other around their applications across diverse task domains. Central to our thesis is that reinforcement learning serves as the critical mechanism for transforming these capabilities from static, heuristic modules into adaptive, robust agentic behavior. To support and accelerate future research, we consolidate the landscape of open-source environments, benchmarks, and frameworks into a practical compendium. By synthesizing over five hundred recent works, this survey charts the contours of this rapidly evolving field and highlights the opportunities and challenges that will shape the development of scalable, general-purpose AI agents.

cs.AI↗

arXiv2026

Invariant Causal Routing for Governing Social Norms in Online Market Economies

Social norms are stable behavioral patterns that emerge endogenously within economic systems through repeated interactions among agents. In online market economies, such norms -- like fair exposure, sustained participation, and balanced reinvestment -- are critical for long-term stability. We aim to understand the causal mechanisms driving these emergent norms and to design principled interventions that can steer them toward desired outcomes. This is challenging because norms arise from countless micro-level interactions that aggregate into macro-level regularities, making causal attribution and policy transferability difficult. To address this, we propose \textbf{Invariant Causal Routing (ICR)}, a causal governance framework that identifies policy-norm relations stable across heterogeneous environments. ICR integrates counterfactual reasoning with invariant causal discovery to separate genuine causal effects from spurious correlations and to construct interpretable, auditable policy rules that remain effective under distribution shift. In heterogeneous agent simulations calibrated with real data, ICR yields more stable norms, smaller generalization gaps, and more concise rules than correlation or coverage baselines, demonstrating that causal invariance offers a principled and interpretable foundation for governance.

cs.LG↗

arXiv2026

Learning Generation Orders for Masked Discrete Diffusion Models via Variational Inference

Masked discrete diffusion models (MDMs) are a promising new approach to generative modelling, offering the ability for parallel token generation and therefore greater efficiency than autoregressive counterparts. However, achieving an optimal balance between parallel generation and sample quality remains an open problem. Current approaches primarily address this issue through fixed, heuristic parallel sampling methods. There exist some recent learning based approaches to this problem, but its formulation from the perspective of variational inference remains underexplored. In this work, we propose a variational inference framework for learning parallel generation orders for MDMs. As part of our method, we propose a parameterisation for the approximate posterior of generation orders which facilitates parallelism and efficient sampling during training. Using this method, we conduct preliminary experiments on the GSM8K dataset, where our method performs competitively against heuristic sampling strategies in the regime of highly parallel generation. For example, our method achieves 33.1\% accuracy with an average of only only 4 generation steps, compared to 23.7-29.0\% accuracy achieved by standard competitor methods in the same number of steps. We believe further experiments and analysis of the method will yield valuable insights into the problem of parallel generation with MDMs.

cs.LG↗

arXiv2026

Seeking Necessary and Sufficient Information from Multimodal Medical Data

Learning multimodal representations from medical images and other data sources can provide richer information for decision-making. While various multimodal models have been developed for this, they overlook learning features that are both necessary (must be present for the outcome to occur) and sufficient (enough to determine the outcome). We argue learning such features is crucial as they can improve model performance by capturing essential predictive information, and enhance model robustness to missing modalities as each modality can provide adequate predictive signals. Such features can be learned by leveraging the Probability of Necessity and Sufficiency (PNS) as a learning objective, an approach that has proven effective in unimodal settings. However, extending PNS to multimodal scenarios remains underexplored and is non-trivial as key conditions of PNS estimation are violated. We address this by decomposing multimodal representations into modality-invariant and modality-specific components, then deriving tractable PNS objectives for each. Experiments on synthetic and real-world medical datasets demonstrate our method's effectiveness. Code will be available on GitHub.

cs.CV↗

arXiv2026

Dynamic Correction of Erroneous State Estimates via Diffusion Bayesian Exploration

In emergency response and other high-stakes societal applications, early-stage state estimates critically shape downstream outcomes. Yet, these initial state estimates-often based on limited or biased information-can be severely misaligned with reality, constraining subsequent actions and potentially causing catastrophic delays, resource misallocation, and human harm. Under the stationary bootstrap baseline (zero transition and no rejuvenation), bootstrap particle filters exhibit Stationarity-Induced Posterior Support Invariance (S-PSI), wherein regions excluded by the initial prior remain permanently unexplorable, making corrections impossible even when new evidence contradicts current beliefs. While classical perturbations can in principle break this lock-in, they operate in an always-on fashion and may be inefficient. To overcome this, we propose a diffusion-driven Bayesian exploration framework that enables principled, real-time correction of early state estimation errors. Our method expands posterior support via entropy-regularized sampling and covariance-scaled diffusion. A Metropolis-Hastings check validates proposals and keeps inference adaptive to unexpected evidence. Empirical evaluations on realistic hazardous-gas localization tasks show that our approach matches reinforcement learning and planning baselines when priors are correct. It substantially outperforms classical SMC perturbations and RL-based methods under misalignment, and we provide theoretical guarantees that DEPF resolves S-PSI while maintaining statistical rigor.

cs.LG↗

arXiv2025

TriShGAN: Enhancing Sparsity and Robustness in Multivariate Time Series Counterfactuals Explanation

In decision-making processes, stakeholders often rely on counterfactual explanations, which provide suggestions about what should be changed in the queried instance to alter the outcome of an AI system. However, generating these explanations for multivariate time series presents challenges due to their complex, multi-dimensional nature. Traditional Nearest Unlike Neighbor-based methods typically substitute subsequences in a queried time series with influential subsequences from an NUN, which is not always realistic in real-world scenarios due to the rigid direct substitution. Counterfactual with Residual Generative Adversarial Networks-based methods aim to address this by learning from the distribution of observed data to generate synthetic counterfactual explanations. However, these methods primarily focus on minimizing the cost from the queried time series to the counterfactual explanations and often neglect the importance of distancing the counterfactual explanation from the decision boundary. This oversight can result in explanations that no longer qualify as counterfactual if minor changes occur within the model. To generate a more robust counterfactual explanation, we introduce TriShGAN, under the CounteRGAN framework enhanced by the incorporation of triplet loss. This unsupervised learning approach uses distance metric learning to encourage the counterfactual explanations not only to remain close to the queried time series but also to capture the feature distribution of the instance with the desired outcome, thereby achieving a better balance between minimal cost and robustness. Additionally, we integrate a Shapelet Extractor that strategically selects the most discriminative parts of the high-dimensional queried time series to enhance the sparsity of counterfactual explanation and efficiency of the training process.

cs.LG↗

arXiv2025

A Principle of Targeted Intervention for Multi-Agent Reinforcement Learning

Steering cooperative multi-agent reinforcement learning (MARL) towards desired outcomes is challenging, particularly when the global guidance from a human on the whole multi-agent system is impractical in a large-scale MARL. On the other hand, designing external mechanisms (e.g., intrinsic rewards and human feedback) to coordinate agents mostly relies on empirical studies, lacking a easy-to-use research tool. In this work, we employ multi-agent influence diagrams (MAIDs) as a graphical framework to address the above issues. First, we introduce the concept of MARL interaction paradigms (orthogonal to MARL learning paradigms), using MAIDs to analyze and visualize both unguided self-organization and global guidance mechanisms in MARL. Then, we design a new MARL interaction paradigm, referred to as the targeted intervention paradigm that is applied to only a single targeted agent, so the problem of global guidance can be mitigated. In implementation, we introduce a causal inference technique, referred to as Pre-Strategy Intervention (PSI), to realize the targeted intervention paradigm. Since MAIDs can be regarded as a special class of causal diagrams, a composite desired outcome that integrates the primary task goal and an additional desired outcome can be achieved by maximizing the corresponding causal effect through the PSI. Moreover, the bundled relevance graph analysis of MAIDs provides a tool to identify whether an MARL learning paradigm is workable under the design of an MARL interaction paradigm. In experiments, we demonstrate the effectiveness of our proposed targeted intervention, and verify the result of relevance graph analysis.

cs.AI↗

arXiv2025

Evolving LLMs' Self-Refinement Capability via Synergistic Training-Inference Optimization

Self-Refinement refers to a model's ability to revise its own responses to produce improved outputs. This capability can also serve as a fundamental mechanism for Self-Improvement, for example, by reconstructing datasets with refined results to enhance intrinsic model performance. However, our comprehensive experiments reveal that large language models (LLMs) show no clear evidence of inherent Self-Refinement and may even experience response quality degradation after Self-Refinement. To address this issue, we propose EVOLVE, a simple and effective framework for eliciting and tracking the evolution of Self-Refinement through iterative training. We first explore optimization methods during training to activate the model's Self-Refinement capability. Then, at inference, we investigate various generation strategies to further enhance and utilize Self-Refinement while supplying the necessary data for training. Through synergistic optimization of training and inference stages, we continually evolve the model's Self-Refinement ability, enabling it to better refine its own responses. Moreover, we demonstrate the potential of leveraging Self-Refinement to achieve broader Self-Improvement of intrinsic model abilities. Experiments show that the evolved Self-Refinement ability enables the Llama-3.1-8B base model to surpass GPT-4o, achieving 62.3% length-controlled and 63.3% raw win rates on AlpacaEval 2, and 50.3% on Arena-Hard. It also generalizes effectively to out-of-domain reasoning tasks, improving performance on mathematical reasoning benchmarks such as GSM8K and MATH.

cs.CL↗

arXiv2025

Curious Causality-Seeking Agents Learn Meta Causal World

When building a world model, a common assumption is that the environment has a single, unchanging underlying causal rule, like applying Newton's laws to every situation. In reality, what appears as a drifting causal mechanism is often the manifestation of a fixed underlying mechanism seen through a narrow observational window. This brings about a problem that, when building a world model, even subtle shifts in policy or environment states can alter the very observed causal mechanisms. In this work, we introduce the \textbf{Meta-Causal Graph} as world models, a minimal unified representation that efficiently encodes the transformation rules governing how causal structures shift across different latent world states. A single Meta-Causal Graph is composed of multiple causal subgraphs, each triggered by meta state, which is in the latent state space. Building on this representation, we introduce a \textbf{Causality-Seeking Agent} whose objectives are to (1) identify the meta states that trigger each subgraph, (2) discover the corresponding causal relationships by agent curiosity-driven intervention policy, and (3) iteratively refine the Meta-Causal Graph through ongoing curiosity-driven exploration and agent experiences. Experiments on both synthetic tasks and a challenging robot arm manipulation task demonstrate that our method robustly captures shifts in causal dynamics and generalizes effectively to previously unseen contexts.

cs.LG↗

arXiv2025

Causal Sufficiency and Necessity Improves Chain-of-Thought Reasoning

Chain-of-Thought (CoT) prompting plays an indispensable role in endowing large language models (LLMs) with complex reasoning capabilities. However, CoT currently faces two fundamental challenges: (1) Sufficiency, which ensures that the generated intermediate inference steps comprehensively cover and substantiate the final conclusion; and (2) Necessity, which identifies the inference steps that are truly indispensable for the soundness of the resulting answer. We propose a causal framework that characterizes CoT reasoning through the dual lenses of sufficiency and necessity. Incorporating causal Probability of Sufficiency and Necessity allows us not only to determine which steps are logically sufficient or necessary to the prediction outcome, but also to quantify their actual influence on the final reasoning outcome under different intervention scenarios, thereby enabling the automated addition of missing steps and the pruning of redundant ones. Extensive experimental results on various mathematical and commonsense reasoning benchmarks confirm substantial improvements in reasoning efficiency and reduced token usage without sacrificing accuracy. Our work provides a promising direction for improving LLM reasoning performance and cost-effectiveness.

cs.CL↗

arXiv2025

Memory-Driven Self-Improvement for Decision Making with Large Language Models

Large language models (LLMs) have emerged as effective action policies for sequential decision-making (SDM) tasks due to their extensive prior knowledge. However, this broad yet general knowledge is often insufficient for specific decision-making tasks with limited task-related data, making it challenging to efficiently adapt LLMs to specific SDM tasks. To address this challenge, we propose a memory-driven self-improvement framework that combines LLM general prior knowledge with a compact memory of domain-specific experiences. Memory retains past interactions and associated Q-values, thereby capturing decision-relevant knowledge that facilitates accurate value estimation and informs the LLM prior refinement. The refined LLM prior, in turn, generates higher-reward trajectories that further enrich memory, forming a natural self-improvement framework where memory and LLM prior mutually reinforce each other. Experiments show that our memory-driven approach significantly outperforms both traditional RL and LLM-based baselines, e.g., improving performance by over 40\% on in-distribution tasks and over 75\% when generalized to unseen tasks in ALFWorld.

cs.LG↗

arXiv2025

Decentralized Dynamic Cooperation of Personalized Models for Federated Continual Learning

Federated continual learning (FCL) has garnered increasing attention for its ability to support distributed computation in environments with evolving data distributions. However, the emergence of new tasks introduces both temporal and cross-client shifts, making catastrophic forgetting a critical challenge. Most existing works aggregate knowledge from clients into a global model, which may not enhance client performance since irrelevant knowledge could introduce interference, especially in heterogeneous scenarios. Additionally, directly applying decentralized approaches to FCL suffers from ineffective group formation caused by task changes. To address these challenges, we propose a decentralized dynamic cooperation framework for FCL, where clients establish dynamic cooperative learning coalitions to balance the acquisition of new knowledge and the retention of prior learning, thereby obtaining personalized models. To maximize model performance, each client engages in selective cooperation, dynamically allying with others who offer meaningful performance gains. This results in non-overlapping, variable coalitions at each stage of the task. Moreover, we use coalitional affinity game to simulate coalition relationships between clients. By assessing both client gradient coherence and model similarity, we quantify the client benefits derived from cooperation. We also propose a merge-blocking algorithm and a dynamic cooperative evolution algorithm to achieve cooperative and dynamic equilibrium. Comprehensive experiments demonstrate the superiority of our method compared to various baselines. Code is available at: https://github.com/ydn3229/DCFCL.

cs.LG↗

arXiv2025

MF-LLM: Simulating Population Decision Dynamics via a Mean-Field Large Language Model Framework

Simulating collective decision-making involves more than aggregating individual behaviors; it emerges from dynamic interactions among individuals. While large language models (LLMs) offer strong potential for social simulation, achieving quantitative alignment with real-world data remains a key challenge. To bridge this gap, we propose the Mean-Field LLM (MF-LLM) framework, the first to incorporate mean field theory into LLM-based social simulation. MF-LLM models bidirectional interactions between individuals and the population through an iterative process, generating population signals to guide individual decisions, which in turn update the signals. This interplay produces coherent trajectories of collective behavior. To improve alignment with real-world data, we introduce IB-Tune, a novel fine-tuning method inspired by the Information Bottleneck principle, which retains population signals most predictive of future actions while filtering redundant history. Evaluated on a real-world social dataset, MF-LLM reduces KL divergence to human population distributions by 47\% compared to non-mean-field baselines, enabling accurate trend forecasting and effective intervention planning. Generalizing across 7 domains and 4 LLM backbones, MF-LLM provides a scalable, high-fidelity foundation for social simulation.

cs.MA↗

arXiv2025

Large Language Models are Demonstration Pre-Selectors for Themselves

In-context learning (ICL) with large language models (LLMs) delivers strong few-shot performance by choosing few-shot demonstrations from the entire training data. However, existing ICL methods, which rely on similarity or diversity scores to choose demonstrations, incur high computational costs due to repeatedly retrieval from large-scale datasets for each query. To this end, we propose FEEDER (FEw yet Essential Demonstration prE-selectoR), a novel pre-selection framework that identifies a representative subset of demonstrations containing the most representative examples in the training data, tailored to specific LLMs. To construct this subset, we introduce the "sufficiency" and "necessity" metrics in the pre-selection stage and design a tree-based algorithm to identify representative examples efficiently. Once pre-selected, this representative subset can effectively replace the full training data, improving efficiency while maintaining comparable performance in ICL. Additionally, our pre-selected subset also benefits fine-tuning LLMs, where we introduce a bi-level optimization method that enhances training efficiency without sacrificing performance. Experiments with LLMs ranging from 300M to 8B parameters show that FEEDER can reduce training data size by over 20% while maintaining performance and seamlessly integrating with various downstream demonstration selection strategies in ICL.

cs.CL↗

arXiv2025

Fine-Grained Interpretation of Political Opinions in Large Language Models

Studies of LLMs' political opinions mainly rely on evaluations of their open-ended responses. Recent work indicates that there is a misalignment between LLMs' responses and their internal intentions. This motivates us to probe LLMs' internal mechanisms and help uncover their internal political states. Additionally, we found that the analysis of LLMs' political opinions often relies on single-axis concepts, which can lead to concept confounds. In this work, we extend the single-axis to multi-dimensions and apply interpretable representation engineering techniques for more transparent LLM political concept learning. Specifically, we designed a four-dimensional political learning framework and constructed a corresponding dataset for fine-grained political concept vector learning. These vectors can be used to detect and intervene in LLM internals. Experiments are conducted on eight open-source LLMs with three representation engineering techniques. Results show these vectors can disentangle political concept confounds. Detection tasks validate the semantic meaning of the vectors and show good generalization and robustness in OOD settings. Intervention Experiments show these vectors can intervene in LLMs to generate responses with different political leanings.

cs.CL↗

arXiv2025

Learning Macroeconomic Policies through Dynamic Stackelberg Mean-Field Games

Macroeconomic outcomes emerge from individuals' decisions, making it essential to model how agents interact with macro policy via consumption, investment, and labor choices. We formulate this as a dynamic Stackelberg game: the government (leader) sets policies, and agents (followers) respond by optimizing their behavior over time. Unlike static models, this dynamic formulation captures temporal dependencies and strategic feedback critical to policy design. However, as the number of agents increases, explicitly simulating all agent-agent and agent-government interactions becomes computationally infeasible. To address this, we propose the Dynamic Stackelberg Mean Field Game (DSMFG) framework, which approximates these complex interactions via agent-population and government-population couplings. This approximation preserves individual-level feedback while ensuring scalability, enabling DSMFG to jointly model three core features of real-world policymaking: dynamic feedback, asymmetry, and large scale. We further introduce Stackelberg Mean Field Reinforcement Learning (SMFRL), a data-driven algorithm that learns the leader's optimal policies while maintaining personalized responses for individual agents. Empirically, we validate our approach in a large-scale simulated economy, where it scales to 1,000 agents (vs. 100 in prior work) and achieves a fourfold increase in GDP over classical economic methods and a nineteenfold improvement over the static 2022 U.S. federal income tax policy.

econ.TH↗

arXiv2025

Natural Language Reinforcement Learning

Artificial intelligence progresses towards the "Era of Experience," where agents are expected to learn from continuous, grounded interaction. We argue that traditional Reinforcement Learning (RL), which typically represents value as a scalar, can restrict agent's deep understanding of environments and hinders the active, deliberative learning crucial for navigating this new paradigm. To address the issue, we introduce Natural Language Reinforcement Learning (NLRL), a framework that extends RL principles into natural language counterparts. Central to NLRL is the Language Value Function (LVF), which redefines value as an interpretable linguistic narrative articulating the rationale behind an evaluation. NLRL further extends this concept to core RL components, including policy, the Bellman equation, and policy iteration. Leveraging recent advancements in Large Language Models (LLMs), NLRL can be practically implemented to achieve RL-like policy and value training through unsupervised environment interactions. Experiments over 4 multi-step agentic tasks demonstrate NLRL's effectiveness, efficiency, and its potential to foster deeper understanding and more active learning strategies.

cs.LG↗

arXiv2025

Estimating the Effects of Sample Training Orders for Large Language Models without Retraining

The order of training samples plays a crucial role in large language models (LLMs), significantly impacting both their external performance and internal learning dynamics. Traditional methods for investigating this effect generally require retraining the model with various sample orders, which is computationally infeasible for LLMs. In this work, we improve traditional methods by designing a retraining-free framework. By approximating Adam optimizer updates with first- and second-order Taylor expansions and utilizing random projection methods to store intermediate checkpoints, our framework can efficiently estimate model parameters for arbitrary training sample orders. Next, we apply our framework to two downstream research problems: (1) Training curriculum design for LLMs -- we base our retraining-free framework to propose a novel curriculum learning strategy that augments curriculum proposals with estimated model performances, enabling more informed sample scheduling. (2) LLMs' memorization and generalization effect analysis -- we use our retraining-free framework to estimate how the positions of training samples influence LLMs' capacity for memorization and generalization. We conduct extensive experiments to validate the effectiveness of our retraining-free framework in reproducing the true model performances, and further demonstrate its potential in optimizing LLM training curricula and analyzing the memorization and generalization effects of LLMs.

cs.LG↗