Searcharxiv⌕ Search

arXiv subjects

Junhao Shen

Publications and source records attributed to Junhao Shen.

At least 19 recordsLinked to original sources

MMDM: Measuring Massive Multitask Decision Making

System-1 Decision Models directly make structured probabilistic decisions over query-specified options given a state in a single forward pass, without autoregressively generating reasoning trajectories or textual responses. Recent evaluations cover diverse tasks but often rely on existing benchmarks, where prior exposure to test material can obscure generalization. Moreover, real-world decision making often involves long contexts, large option spaces, and complex logical dependencies, challenges that existing benchmarks do not adequately capture together. Thus, we introduce MMDM (Massive Multitask Decision Making), a benchmark of 5,448 high-quality questions across 16 task categories, combining newly constructed tasks with source-grounded recompositions and rewrites for single-select, multi-select, and soft-label decision making in fresh contexts. It covers realistic scenarios such as agent state tracking, tool and skill selection, preference judgments, and rubric scoring, with option spaces ranging from 2 to 1,024 and inputs extending to 64,817 tokens. With MMDM, we conduct a comprehensive quantitative evaluation of 28 decision models, showing that many perform well on common knowledge yet struggle with interdependent options, constrained decisions, and multi-select tasks. For example, Perplexity Decider-V1.2 reaches 78.75% Hard Single-select Accuracy on world knowledge, but only 53.95% on option logic and 47.95% on constraint planning, while the best Hard Multi-select Accuracy across all models is 36.81%. These results highlight joint reasoning, constraint-aware selection, and complete multi-select recovery as key priorities for decision model development, with MMDM providing a testbed for measuring this progress.

cs.LG↗

Measure Convergence under Weak Domination for Unitarily Invariant Norms on Finite von Neumann Algebras

The classical Lebesgue dominated convergence theorem (DCT) is a cornerstone in commutative measure theory, while its extension to the noncommutative setting has remained a longstanding challenge due to the barriers posed by the noncommutative structure. Let \(α\) be a unitarily invariant norm defined on a diffuse tracial von Neumann algebra \((\mathcal{M}, τ)\), let $L^α(\mathcal M, τ)$ be the completion of $\mathcal M$ with respect to the norm $α$(which includes all $L^p(\mathcal M, τ)$ as special cases). In this paper, by replacing ``almost everywhere convergence" with Nelson's convergence in measure, we establish a noncommutative analog of the dominated convergence theorem for $L^α(\mathcal M, τ)$ under a { significantly weaker dominating condition} than those previously considered. For instance, in a \(II_1\) factor \(\mathcal{M}\), if \(P\) is a projection with \(τ(P)=\frac{1}{n}\), then \(n^2 P\) ``dominates" every operator of norm \(\leq n\). As an application, we obtain a complete characterization of center-preserving $*$-automorphism of finite von Neumann algebras, which unifies and extends the earlier works of Kadison and Haagerup on the structure of automorphisms.

math.OA↗

Work Agents: Systems, Capabilities, and Optimization

As foundation models evolve from information generation to digital work execution, agents are increasingly capable of completing complex tasks across tools, applications, and workflows. However, existing research lacks a unified perspective on how agents reliably accomplish delegated work. To address this gap, we present a work-centric, system-level perspective that views work agents as systems combining foundation models and runtime harnesses to execute tasks in stateful workspaces and deliver verifiable outcomes. We introduce two complementary dimensions, work persistence and work openness, to characterize working conditions, and classify digital work into five forms according to task scope and organization. Building on this framework, we identify five core capability families required for reliable work execution and examine approaches to model optimization, harness optimization, and their joint optimization. We further analyze existing evaluation methods and metrics from the perspective of work outcomes. Finally, we discuss future research directions, including sustained work delegation, adaptation to open environments, and the interaction between work execution and capability improvement. This paper provides a unified framework for understanding, optimizing, and evaluating work agents, supporting the development of systems capable of delivering reliable, verifiable, and practically valuable work outcomes.

cs.LG↗

Dynamic Skill Lifecycle Management for Agentic Reinforcement Learning

Large language model agents increasingly rely on external skills to solve complex tasks, where skills act as modular units that extend their capabilities beyond what parametric memory alone supports. Existing skill-based RL methods assume external skills either accumulate as persistent guidance or internalized into the policy, eventually leading to zero-skill inference. We argue this assumption is overly restrictive, since with limited parametric capacity and uneven marginal contribution across skills, the optimal active skill set is non-monotonic, task- and stage-dependent. In this work, we propose SLIM, a framework of dynamic Skill LIfecycle Management for agentic reinforcement learning (RL), which dynamically updates the active external skill set as the policy evolves.Specifically, SLIM estimates each active skill's marginal external contribution through leave-one-skill-out validation, then applies three lifecycle operations: retaining high-value skills, retiring skills whose contribution becomes negligible after sufficient exposure, and expanding the skill bank when persistent failures reveal missing capability coverage. Experiments show that SLIM outperforms the best baselines by an average of 7.1\% points across ALFWorld and SearchQA. Results further indicate that policy learning and external skill retention are not mutually exclusive: some skills are absorbed into the policy, while others continue to provide external value, supporting SLIM as a more general paradigm for skill-based agentic RL.

cs.LG↗

Progressive Agent Skill Generation via Reinforcement Learning

Recent large language model agents often use external skills as modular procedural units that condition inference and improve complex task solving. Thus, automatically generating high-quality skills from documents or experience has become an important problem. Existing skill generation methods largely rely on heuristics or pipeline-style consolidation, which must be specially designed for different evidence sources. In contrast, learning-based approaches offer a more unified way to model skill generation across heterogeneous sources. However, learning-based skill generation remains challenging because skills lack a natural supervision signal based on relevance or correctness; their value can largely be determined only by whether they improve the behavior of the agent on downstream tasks. To address this challenge, we propose Skill-$α$, a reinforcement learning method that learns a unified policy for progressive skill generation. Specifically, we construct each skill by repeatedly applying the learned policy to successive source evidence and introduce a novel rollback reward that evaluates each edit by comparing downstream execution under the original and edited skills on an anchored query. Extensive experiments show that Skill-$α$ generates more effective skills than methods based on heuristics or pipelines in both document-to-skill and experience-to-skill settings. Under the main GPT-4o worker, Skill-$α$ improves average downstream success rates over the strongest skill-generation baseline by 3.1 points on CL-Bench and 6.7 points on tau2-bench. Further ablations and analysis validate the importance of rollback reward and progressive generation.

cs.LG↗

Achieving Olympiad-Level Geometry Large Language Model Agent via Complexity Boosting Reinforcement Learning

Large language model (LLM) agents exhibit strong mathematical problem-solving abilities and can even solve International Mathematical Olympiad (IMO) level problems with the assistance of formal proof systems. However, due to weak heuristics for auxiliary constructions, AI for geometry problem solving remains dominated by expert models such as AlphaGeometry 2, which rely heavily on large-scale data synthesis and search for both training and evaluation. In this work, we make the first attempt to build a medalist-level LLM agent for geometry and present InternGeometry. InternGeometry overcomes the heuristic limitations in geometry by iteratively proposing propositions and auxiliary constructions, verifying them with a symbolic engine, and reflecting on the engine's feedback to guide subsequent proposals. A dynamic memory mechanism enables InternGeometry to conduct more than two hundred interactions with the symbolic engine per problem. To further accelerate learning, we introduce Complexity-Boosting Reinforcement Learning (CBRL), which gradually increases the complexity of synthesized problems across training stages. Built on InternThinker-32B, InternGeometry solves 44 of 50 IMO geometry problems (2000-2024), exceeding the average gold medalist score (40.9), using only 13K training examples, just 0.004% of the data used by AlphaGeometry 2, demonstrating the potential of LLM agents on expert-level geometry tasks. InternGeometry can also propose novel auxiliary constructions for IMO problems that do not appear in human solutions.

cs.AI↗

Co-Evolution in Agentic Systems: Toward Self-Directed Evolution Beyond Human Design

Agentic systems are increasingly expected to improve after deployment, yet single-entity self-evolution is often bounded by a static learning context, such as fixed tasks and feedback. This survey focuses on co-evolution in agentic systems, a multi-component form of self-evolution in which multiple agents and their environment impose adaptive pressure on one another. To organize existing papers, we propose a progressive three-stage taxonomy that traces how the system gradually sheds human-engineered constraints. Agent--Agent Co-Evolution studies how agents adapt through dynamic peers, including adversarial, collaborative, and organizational adaptation. Agent--Environment Co-Evolution extends this loop to adaptive tasks, feedback, and interaction spaces that change with the agents. Meta Co-Evolution further explores the possibility of making the evolution mechanism itself evolvable. We also discuss open challenges in evaluating such systems, scaling them across multiple components, and keeping increasingly autonomous evolutionary processes safe and controllable. This survey provides a unified foundation for building robust and open-ended agentic systems that can improve beyond fixed human-designed paths.

cs.CL↗

Intern-S1-MO: Long-horizon Reasoning Agent for Olympiad?Level Mathematical Problem Solving

Large Reasoning Models (LRMs) have expanded the mathematical reasoning frontier through Chain-of-Thought (CoT) techniques and Reinforcement Learning with Verifiable Rewards (RLVR), capable of solving AIME-level problems. However, the performance of LRMs is heavily dependent on the extended reasoning context length. For solving ultra-hard problems like those in the International Mathematical Olympiad (IMO), the required reasoning complexity surpasses the space that an LRM can explore in a single round. Previous works attempt to extend the reasoning context of LRMs but remain prompt-based and built upon proprietary models, lacking systematic structures and training pipelines. Therefore, this paper introduces Intern-S1-MO, a long-horizon math agent that conducts multi-round hierarchical reasoning, composed of an LRM-based multi-agent system including reasoning, summary, and verification. By maintaining a compact memory in the form of lemmas, Intern-S1-MO can more freely explore the lemma-rich reasoning spaces in multiple reasoning stages, thereby breaking through the context constraints for IMO-level math problems. Furthermore, we propose OREAL-H, an RL framework for training the LRM using the online explored trajectories to simultaneously bootstrap the reasoning ability of LRM and elevate the overall performance of Intern-S1-MO. Experiments show that Intern-S1-MO can obtain 26 out of 35 points on the non-geometry problems of IMO2025, matching the performance of silver medalists. It also surpasses the current advanced LRMs on inference benchmarks such as HMMT2025, AIME2025, and CNMO2025. In addition, our agent officially participates in CMO2025 and achieves a score of 102/126 under the judgment of human experts, reaching the gold medal level.

cs.CL↗

Density of irreducible operators in the trace-class norm

In 1968, Paul Halmos initiated the research on density of the set of irreducible operators on a separable Hilbert space. Through the research, a long-standing unsolved problem inquires: is the set of irreducible operators dense in $B(H)$ with respect to the trace-class norm topology? Precisely, for each operator $T $ in $B(H)$ and every $\varepsilon >0$, is there a trace-class operator $K$ such that $T+K$ is irreducible and $\Vert K \Vert_1 < \varepsilon$? For $p>1$, to prove the $\Vert \cdot \Vert_p$-norm density of irreducible operators in $B(H)$, a type of Weyl-von Neumann theorem effects as a key technique. But the traditional method fails for the case $p=1$, where by $\Vert \cdot \Vert_p$-norm we denote the Schatten $p$-norm. In the current paper, for a large family of operators in $B(H)$, we give the above long-term problem an affirmative answer. The result is derived from a combination of techniques in both operator theory and operator algebras. Moreover, we discover that there is a strong connection between the problem and another related operator-theoretical problem related to type $\mathrm{II}_1$ von Neumann algebras.

math.OA↗

Products of irreducible operators in factors

Let $\mathcal M$ be a separable factor. An operator $T$ in $\mathcal{M}$ is said to be irreducible in $\mathcal{M}$ if the von Neumann algebra $W^*(T)$ generated by $T$ is an irreducible subfactor of $\mathcal{M}$, i.e., $W^*(T)'\cap\mathcal{M}=\mathbb{C}I$. In this paper, we show that every operator in a separable factor $\mathcal{M}$ is the product of two irreducible operators in $\mathcal{M}$, except the zero operator in factors of type $\mathrm{I}_{2n+1}$ for $n\geqslant 1$. This may be viewed as a multiplicative analogue of Radjavi's result which asserts that every operator on a separable Hilbert space is the sum of two irreducible operators.

math.OA↗

Irreducible operators in von Neumann algebras

Let $\mathcal{M}$ be a separable von Neumann algebra with center $\mathcal{Z}(\mathcal{M})$. An operator $T$ in $\mathcal{M}$ is called irreducible if the von Neumann algebra $W^*(T)$ generated by $T$ has trivial relative commutant, i.e., $W^*(T)'\cap\mathcal{M}=\mathcal{Z}(\mathcal{M})$. In this paper, we show that irreducible operators in $\mathcal{M}$ form a norm-dense $G_δ$ set, which is a generalization of Halmos' theorem. Moreover, we prove that every operator in $\mathcal{M}$ is the sum of two irreducible operators, which is an analogue of Radjavi's theorem.

math.OA↗

Reducible operators in non-$Γ$ type ${\rm II}_1$ factors

A famous question of Halmos asks whether every operator on a separable infinite-dimensional Hilbert space is a norm limit of reducible operators. In [30], Voiculescu gave this problem an affirmative answer by his remarkable non-commutative Weyl-von Neumann theorem. We investigate the existence or non-existence of an analogue of Voiculescu's result in factors of type ${\rm II}_1$. In the paper, we prove that, in the operator norm topology, the set of reducible operators is ${\it nowhere}$ dense in a non-$Γ$ factor $\mathcal M$ of type ${\rm II}_1$, where separable and non-separable cases of $\mathcal M$ are both considered. Main tools developed in the paper are a new characterization of Murray and von Neumann's Property $Γ$ for a factor of type ${\rm II}_1$ and a spectral gap property for a single operator in a non-$Γ$ factor of type ${\rm II}_1$.

math.OA↗

Semi-off-Policy Reinforcement Learning for Vision-Language Slow-Thinking Reasoning

Enhancing large vision-language models (LVLMs) with visual slow-thinking reasoning is crucial for solving complex multimodal tasks. However, since LVLMs are mainly trained with vision-language alignment, it is difficult to adopt on-policy reinforcement learning (RL) to develop the slow thinking ability because the rollout space is restricted by its initial abilities. Off-policy RL offers a way to go beyond the current policy, but directly distilling trajectories from external models may cause visual hallucinations due to mismatched visual perception abilities across models. To address these issues, this paper proposes SOPHIA, a simple and scalable Semi-Off-Policy RL for vision-language slow-tHInking reAsoning. SOPHIA builds a semi-off-policy behavior model by combining on-policy visual understanding from a trainable LVLM with off-policy slow-thinking reasoning from a language model, assigns outcome-based rewards to reasoning, and propagates visual rewards backward. Then LVLM learns slow-thinking reasoning ability from the obtained reasoning trajectories using propagated rewards via off-policy RL algorithms. Extensive experiments with InternVL2.5 and InternVL3.0 with 8B and 38B sizes show the effectiveness of SOPHIA. Notably, SOPHIA improves InternVL3.0-38B by 8.50% in average, reaching state-of-the-art performance among open-source LVLMs on multiple multimodal reasoning benchmarks, and even outperforms some closed-source models (e.g., GPT-4.1) on the challenging MathVision and OlympiadBench, achieving 49.08% and 49.95% pass@1 accuracy, respectively. Analysis shows SOPHIA outperforms supervised fine-tuning and direct on-policy RL methods, offering a better policy initialization for further on-policy training.

cs.LG↗

Voiculescu's Theorem in Properly Infinite Factors

In this paper, we investigate Voiculescu's theorem on approximate unitary equivalence in separable properly infinite factors. As applications, we establish the norm-denseness of the set of all reducible operators, prove a generalized Voiculescu's bicommutant theorem and a version of asymptotic bicommutant theorem, and obtain an interesting cohomological result. Additionally, we extend these results to multiplier algebras within separable type $\mathrm{III}$ factors. At last, a concept of the nuclear length is introduced.

math.OA↗

Learning Fair Decisions with Factor Models: Applications to Annuity Pricing

Fairness-aware statistical learning is essential for mitigating discrimination against protected attributes such as gender, race, and ethnicity in data-driven decision-making. This is particularly critical in high-stakes applications like insurance underwriting and annuity pricing, where biased business decisions can have significant financial and social consequences. Factor models are commonly used in these domains for risk assessment and pricing; however, their predictive outputs may inadvertently introduce or amplify bias. To address this, we propose a Fair Decision Model that incorporates fairness regularization to mitigate outcome disparities. Specifically, the model is designed to ensure that expected decision errors are balanced across demographic groups - a criterion we refer to as Decision Error Parity. We apply this framework to annuity pricing based on mortality modelling. An empirical analysis using Australian mortality data demonstrates that the Fair Decision Model can significantly reduce decision error disparity while also improving predictive accuracy compared to benchmark models, including both traditional and fair factor models.

stat.ME↗

Temporal Graph Neural Network-Powered Paper Recommendation on Dynamic Citation Networks

Due to the rapid growth of scientific publications, identifying all related reference articles in the literature has become increasingly challenging yet highly demanding. Existing methods primarily assess candidate publications from a static perspective, focusing on the content of articles and their structural information, such as citation relationships. There is a lack of research regarding how to account for the evolving impact among papers on their embeddings. Toward this goal, this paper introduces a temporal dimension to paper recommendation strategies. The core idea is to continuously update a paper's embedding when new citation relationships appear, enhancing its relevance for future recommendations. Whenever a citation relationship is added to the literature upon the publication of a paper, the embeddings of the two related papers are updated through a Temporal Graph Neural Network (TGN). A learnable memory update module based on a Recurrent Neural Network (RNN) is utilized to study the evolution of the embedding of a paper in order to predict its reference impact in a future timestamp. Such a TGN-based model learns a pattern of how people's views of the paper may evolve, aiming to guide paper recommendations more precisely. Extensive experiments on an open citation network dataset, including 313,278 articles from https://paperswithcode.com/about PaperWithCode, have demonstrated the effectiveness of the proposed approach.

cs.IR↗

OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems

Recent advancements have seen Large Language Models (LLMs) and Large Multimodal Models (LMMs) surpassing general human capabilities in various tasks, approaching the proficiency level of human experts across multiple domains. With traditional benchmarks becoming less challenging for these models, new rigorous challenges are essential to gauge their advanced abilities. In this work, we present OlympiadBench, an Olympiad-level bilingual multimodal scientific benchmark, featuring 8,476 problems from Olympiad-level mathematics and physics competitions, including the Chinese college entrance exam. Each problem is detailed with expert-level annotations for step-by-step reasoning. Evaluating top-tier models on OlympiadBench, we implement a comprehensive assessment methodology to accurately evaluate model responses. Notably, the best-performing model, GPT-4V, attains an average score of 17.97% on OlympiadBench, with a mere 10.74% in physics, highlighting the benchmark rigor and the intricacy of physical reasoning. Our analysis orienting GPT-4V points out prevalent issues with hallucinations, knowledge omissions, and logical fallacies. We hope that our challenging benchmark can serve as a valuable resource for helping future AGI research endeavors. The data and evaluation code are available at \url{https://github.com/OpenBMB/OlympiadBench}

cs.CL↗

Inductive Cognitive Diagnosis for Fast Student Learning in Web-Based Online Intelligent Education Systems

Cognitive diagnosis aims to gauge students' mastery levels based on their response logs. Serving as a pivotal module in web-based online intelligent education systems (WOIESs), it plays an upstream and fundamental role in downstream tasks like learning item recommendation and computerized adaptive testing. WOIESs are open learning environment where numerous new students constantly register and complete exercises. In WOIESs, efficient cognitive diagnosis is crucial to fast feedback and accelerating student learning. However, the existing cognitive diagnosis methods always employ intrinsically transductive student-specific embeddings, which become slow and costly due to retraining when dealing with new students who are unseen during training. To this end, this paper proposes an inductive cognitive diagnosis model (ICDM) for fast new students' mastery levels inference in WOIESs. Specifically, in ICDM, we propose a novel student-centered graph (SCG). Rather than inferring mastery levels through updating student-specific embedding, we derive the inductive mastery levels as the aggregated outcomes of students' neighbors in SCG. Namely, SCG enables to shift the task from finding the most suitable student-specific embedding that fits the response logs to finding the most suitable representations for different node types in SCG, and the latter is more efficient since it no longer requires retraining. To obtain this representation, ICDM consists of a construction-aggregation-generation-transformation process to learn the final representation of students, exercises and concepts. Extensive experiments across real-world datasets show that, compared with the existing cognitive diagnosis methods that are always transductive, ICDM is much more faster while maintains the competitive inference performance for new students.

cs.AI↗