arXiv subjects

Jun Xu

Publications and source records attributed to Jun Xu.

At least 127 records · Page 7Linked to original sources

arXiv2025

PrLM: Learning Explicit Reasoning for Personalized RAG via Contrastive Reward Optimization

Personalized retrieval-augmented generation (RAG) aims to produce user-tailored responses by incorporating retrieved user profiles alongside the input query. Existing methods primarily focus on improving retrieval and rely on large language models (LLMs) to implicitly integrate the retrieved context with the query. However, such models are often sensitive to retrieval quality and may generate responses that are misaligned with user preferences. To address this limitation, we propose PrLM, a reinforcement learning framework that trains LLMs to explicitly reason over retrieved user profiles. Guided by a contrastively trained personalization reward model, PrLM effectively learns from user responses without requiring annotated reasoning paths. Experiments on three personalized text generation datasets show that PrLM outperforms existing methods and remains robust across varying numbers of retrieved profiles and different retrievers.

cs.IR↗

arXiv2025

Balancing Stylization and Truth via Disentangled Representation Steering

Generating stylized large language model (LLM) responses via representation editing is a promising way for fine-grained output control. However, there exists an inherent trade-off: imposing a distinctive style often degrades truthfulness. Existing representation editing methods, by naively injecting style signals, overlook this collateral impact and frequently contaminate the model's core truthfulness representations, resulting in reduced answer correctness. We term this phenomenon stylization-induced truthfulness collapse. We attribute this issue to latent coupling between style and truth directions in certain key attention heads, and propose StyliTruth, a mechanism that preserves stylization while keeping truthfulness intact. StyliTruth separates the style-relevant and truth-relevant subspaces in the model's representation space via an orthogonal deflation process. This decomposition enables independent control of style and truth in their own subspaces, minimizing interference. By designing adaptive, token-level steering vectors within each subspace, we dynamically and precisely control the generation process to maintain both stylistic fidelity and truthfulness. We validate our method on multiple styles and languages. Extensive experiments and analyses show that StyliTruth significantly reduces stylization-induced truthfulness collapse and outperforms existing inference-time intervention methods in balancing style adherence with truthfulness.

cs.CL↗

arXiv2025

A Survey of Controllable Learning: Methods and Applications in Information Retrieval

Controllability has become a crucial aspect of trustworthy machine learning, enabling learners to meet predefined targets and adapt dynamically at test time without requiring retraining as the targets shift. We provide a formal definition of controllable learning (CL), and discuss its applications in information retrieval (IR) where information needs are often complex and dynamic. The survey categorizes CL according to what is controllable (e.g., multiple objectives, user portrait, scenario adaptation), who controls (users or platforms), how control is implemented (e.g., rule-based method, Pareto optimization, hypernetwork and others), and where to implement control (e.g., pre-processing, in-processing, post-processing methods). Then, we identify challenges faced by CL across training, evaluation, task setting, and deployment in online environments. Additionally, we outline promising directions for CL in theoretical analysis, efficient computation, empowering large language models, application scenarios and evaluation frameworks.

cs.LG↗

arXiv2025

Benefit from Rich: Tackling Search Interaction Sparsity in Search Enhanced Recommendation

In modern online platforms, search and recommendation (S&R) often coexist, offering opportunities for performance improvement through search-enhanced approaches. Existing studies show that incorporating search signals boosts recommendation performance. However, the effectiveness of these methods relies heavily on rich search interactions. They primarily benefit a small subset of users with abundant search behavior, while offering limited improvements for the majority of users who exhibit only sparse search activity. To address the problem of sparse search data in search-enhanced recommendation, we face two key challenges: (1) how to learn useful search features for users with sparse search interactions, and (2) how to design effective training objectives under sparse conditions. Our idea is to leverage the features of users with rich search interactions to enhance those of users with sparse search interactions. Based on this idea, we propose GSERec, a method that utilizes message passing on the User-Code Graphs to alleviate data sparsity in Search-Enhanced Recommendation. Specifically, we utilize Large Language Models (LLMs) with vector quantization to generate discrete codes, which connect similar users and thereby construct the graph. Through message passing on this graph, embeddings of users with rich search data are propagated to enhance the embeddings of users with sparse interactions. To further ensure that the message passing captures meaningful information from truly similar users, we introduce a contrastive loss to better model user similarities. The enhanced user representations are then integrated into downstream search-enhanced recommendation models. Experiments on three real-world datasets show that GSERec consistently outperforms baselines, especially for users with sparse search behaviors.

cs.IR↗

arXiv2025

Bridging Search and Recommendation through Latent Cross Reasoning

Search and recommendation (S&R) are fundamental components of modern online platforms, yet effectively leveraging search behaviors to improve recommendation remains a challenging problem. User search histories often contain noisy or irrelevant signals that can even degrade recommendation performance, while existing approaches typically encode S&R histories either jointly or separately without explicitly identifying which search behaviors are truly useful. Inspired by the human decision-making process, where one first identifies recommendation intent and then reasons about relevant evidence, we design a latent cross reasoning framework that first encodes user S&R histories to capture global interests and then iteratively reasons over search behaviors to extract signals beneficial for recommendation. Contrastive learning is employed to align latent reasoning states with target items, and reinforcement learning is further introduced to directly optimize ranking performance. Extensive experiments on public benchmarks demonstrate consistent improvements over strong baselines, validating the importance of reasoning in enhancing search-aware recommendation.

cs.IR↗

arXiv2025

Think Before Recommend: Unleashing the Latent Reasoning Power for Sequential Recommendation

Sequential Recommendation (SeqRec) aims to predict the next item by capturing sequential patterns from users' historical interactions, playing a crucial role in many real-world recommender systems. However, existing approaches predominantly adopt a direct forward computation paradigm, where the final hidden state of the sequence encoder serves as the user representation. We argue that this inference paradigm, due to its limited computational depth, struggles to model the complex evolving nature of user preferences and lacks a nuanced understanding of long-tail items, leading to suboptimal performance. To address this issue, we propose \textbf{ReaRec}, the first inference-time computing framework for recommender systems, which enhances user representations through implicit multi-step reasoning. Specifically, ReaRec autoregressively feeds the sequence's last hidden state into the sequential recommender while incorporating special reasoning position embeddings to decouple the original item encoding space from the multi-step reasoning space. Moreover, we introduce two lightweight reasoning-based learning methods, Ensemble Reasoning Learning (ERL) and Progressive Reasoning Learning (PRL), to further effectively exploit ReaRec's reasoning potential. Extensive experiments on five public real-world datasets and different SeqRec architectures demonstrate the generality and effectiveness of our proposed ReaRec. Remarkably, post-hoc analyses reveal that ReaRec significantly elevates the performance ceiling of multiple sequential recommendation backbones by approximately 30\%-50\%. Thus, we believe this work can open a new and promising avenue for future research in inference-time computing for sequential recommendation.

cs.IR↗

arXiv2025

Investigating $^{238}$U Deformation via Dilepton Production in Relativistic Heavy-Ion Collisions

Due to their weak coupling to the strongly interacting matter produced in relativistic heavy-ion collisions, dileptons serve as a sensitive probe of the initial geometry of the colliding nuclei. In this study, we investigate the influence of initial nuclear quadrupole deformation, characterized by the parameter $β_2$, on dilepton production in $U+U$ collisions at $\sqrt{s_{NN}}=196$ GeV. The analysis is varried out using a modified multiphase transport model in which partonic interactions are described by the Nambu-Jona-Lasinio model. We observe a clear linear dependence of dilepton yields on $β_2^2$ in both the low-mass region (LMR, $<1 GeV/c^2$) and intermediate-mass region (IMR, $1-3 GeV/c^2$) of the dilepton spectrum for the most central collisions. Also, dilepton production in the IMR region exhibits a stronger sensitivity to nuclear deformation than in the LMR, reflecting the dominance of earlier partonic processes in this mass range. These results suggest that precise measurements of dilepton yields in relativistic heavy-ion collisions can provide a viable means to determine the deformation parameter $β_2$ of $^{238}$U.

nucl-th↗

arXiv2025

Configurable multi-agent framework for scalable and realistic testing of llm-based agents

Large-language-model (LLM) agents exhibit complex, context-sensitive behaviour that quickly renders static benchmarks and ad-hoc manual testing obsolete. We present Neo, a configurable, multi-agent framework that automates realistic, multi-turn evaluation of LLM-based systems. Neo couples a Question Generation Agent and an Evaluation Agent through a shared context-hub, allowing domain prompts, scenario controls and dynamic feedback to be composed modularly. Test inputs are sampled from a probabilistic state model spanning dialogue flow, user intent and emotional tone, enabling diverse, human-like conversations that adapt after every turn. Applied to a production-grade Seller Financial Assistant chatbot, Neo (i) uncovered edge-case failures across five attack categories with a 3.3% break rate close to the 5.8% achieved by expert human red-teamers, and (ii) delivered 10-12X higher throughput, generating 180 coherent test questions in around 45 mins versus 16h of human effort. Beyond security probing, Neo's stochastic policies balanced topic coverage and conversational depth, yielding broader behavioural exploration than manually crafted scripts. Neo therefore lays a foundation for scalable, self-evolving LLM QA: its agent interfaces, state controller and feedback loops are model-agnostic and extensible to richer factual-grounding and policy-compliance checks. We release the framework to facilitate reproducible, high-fidelity testing of emerging agentic systems.

cs.AI↗

arXiv2025

Reinforcement learning in pursuit-evasion differential game: safety, stability and robustness

Safety and stability are two critical concerns in pursuit-evasion (PE) problems in an obstacle-rich environment. Most existing works combine control barrier functions (CBFs) and reinforcement learning (RL) to provide an efficient and safe solution. However, they do not consider the presence of disturbances, such as wind gust and actuator fault, which may exist in many practical applications. This paper integrates CBFs and a sliding mode control (SMC) term into RL to simultaneously address safety, stability, and robustness to disturbances. However, this integration is significantly challenging due to the strong coupling between the CBF and SMC terms. Inspired by Stackelberg game, we handle the coupling issue by proposing a hierarchical design scheme where SMC and safe control terms interact with each other in a leader-follower manner. Specifically, the CBF controller, acting as the leader, enforces safety independently of the SMC design; while the SMC term, as the follower, is designed based on the CBF controller. We then formulate the PE problem as a zero-sum game and propose a safe robust RL framework to learn the min-max strategy online. A sufficient condition is provided under which the proposed algorithm remains effective even when constraints are conflicting. Simulation results demonstrate the effectiveness of the proposed safe robust RL framework.

eess.SY↗

arXiv2025

MoRE: A Mixture of Reflectors Framework for Large Language Model-Based Sequential Recommendation

Large language models (LLMs) have emerged as a cutting-edge approach in sequential recommendation, leveraging historical interactions to model dynamic user preferences. Current methods mainly focus on learning processed recommendation data in the form of sequence-to-sequence text. While effective, they exhibit three key limitations: 1) failing to decouple intra-user explicit features (e.g., product titles) from implicit behavioral patterns (e.g., brand loyalty) within interaction histories; 2) underutilizing cross-user collaborative filtering (CF) signals; and 3) relying on inefficient reflection update strategies. To address this, We propose MoRE (Mixture of REflectors), which introduces three perspective-aware offline reflection processes to address these gaps. This decomposition directly resolves Challenges 1 (explicit/implicit ambiguity) and 2 (CF underutilization). Furthermore, MoRE's meta-reflector employs a self-improving strategy and a dynamic selection mechanism (Challenge 3) to adapt to evolving user preferences. First, two intra-user reflectors decouple explicit and implicit patterns from a user's interaction sequence, mimicking traditional recommender systems' ability to distinguish surface-level and latent preferences. A third cross-user reflector captures CF signals by analyzing user similarity patterns from multiple users' interactions. To optimize reflection quality, MoRE's meta-reflector employs a offline self-improving strategy that evaluates reflection impacts through comparisons of presence/absence and iterative refinement of old/new versions, with a online contextual bandit mechanism dynamically selecting the optimal perspective for recommendation for each user. Code: https://github.com/E-qin/MoRE-Rec.

cs.IR↗

arXiv2025

MurreNet: Modeling Holistic Multimodal Interactions Between Histopathology and Genomic Profiles for Survival Prediction

Cancer survival prediction requires integrating pathological Whole Slide Images (WSIs) and genomic profiles, a challenging task due to the inherent heterogeneity and the complexity of modeling both inter- and intra-modality interactions. Current methods often employ straightforward fusion strategies for multimodal feature integration, failing to comprehensively capture modality-specific and modality-common interactions, resulting in a limited understanding of multimodal correlations and suboptimal predictive performance. To mitigate these limitations, this paper presents a Multimodal Representation Decoupling Network (MurreNet) to advance cancer survival analysis. Specifically, we first propose a Multimodal Representation Decomposition (MRD) module to explicitly decompose paired input data into modality-specific and modality-shared representations, thereby reducing redundancy between modalities. Furthermore, the disentangled representations are further refined then updated through a novel training regularization strategy that imposes constraints on distributional similarity, difference, and representativeness of modality features. Finally, the augmented multimodal features are integrated into a joint representation via proposed Deep Holistic Orthogonal Fusion (DHOF) strategy. Extensive experiments conducted on six TCGA cancer cohorts demonstrate that our MurreNet achieves state-of-the-art (SOTA) performance in survival prediction.

eess.IV↗

arXiv2025

KAG-Thinker: Interactive Thinking and Deep Reasoning in LLMs via Knowledge-Augmented Generation

In this paper, we introduce KAG-Thinker, which upgrade KAG to a multi-turn interactive thinking and deep reasoning framework powered by a dedicated parameter-light large language model (LLM). Our approach constructs a structured thinking process for solving complex problems, enhancing the the logical coherence and contextual consistency of the reasoning process in question-answering (Q&A) tasks on domain-specific knowledge bases (KBs) within LLMs. Following the \textbf{Logical Form} guided retrieval and reasoning technology route of KAG, this framework first decomposes complex questions into independently solvable sub-problems (which are also referred to as logical forms) through \textbf{breadth decomposition}. Each such logical form is represented in two equivalent forms-natural language and logical function-and subsequently classified as either a Knowledge Retrieval or Reasoning Analysis task. Dependencies and parameter passing between these tasks are explicitly modeled via logical function interfaces. In the solving process, the Retrieval function performs retrieval tasks. It retrieves one-hop structured and unstructured information of specified knowledge unit. While the Math and Deduce functions are used to perform reasoning analysis tasks. Secondly, it is worth noting that, in the Knowledge Retrieval sub-problem tasks, LLMs and external knowledge sources are regarded as equivalent KBs. We use the \textbf{knowledge boundary} module to determine the optimal source using self-regulatory mechanisms such as confidence calibration and reflective reasoning, and use the \textbf{depth solving} module to enhance the comprehensiveness of knowledge acquisition...

cs.CL↗

arXiv2025

Spin polarization from nucleon-nucleon scatterings in intermediate-energy heavy-ion collisions

We propose a new mechanism of generating spin polarization in heavy-ion collisions dominated by nucleon degree of freedom. By incorporating the spin change in nucleon-nucleon scatterings based on the phase shift data together with the constraint of rigorous angular momentum conservation and Pauli blocking, we illustrate through a Boltzmann-Uehling-Uhlenbeck transport model that appreciable spin polarization (about $1 \sim 2\%$) can be generated in intermediate-energy heavy-ion collisions. This mechanism, together with the nuclear spin-orbit potential, may help to understand the spin polarization in few-GeV heavy-ion collisions dominated by nucleon degree of freedom.

nucl-th↗

arXiv2025

Similarity = Value? Consultation Value Assessment and Alignment for Personalized Search

Personalized search systems in e-commerce platforms increasingly involve user interactions with AI assistants, where users consult about products, usage scenarios, and more. Leveraging consultation to personalize search services is trending. Existing methods typically rely on semantic similarity to align historical consultations with current queries due to the absence of 'value' labels, but we observe that semantic similarity alone often fails to capture the true value of consultation for personalization. To address this, we propose a consultation value assessment framework that evaluates historical consultations from three novel perspectives: (1) Scenario Scope Value, (2) Posterior Action Value, and (3) Time Decay Value. Based on this, we introduce VAPS, a value-aware personalized search model that selectively incorporates high-value consultations through a consultation-user action interaction module and an explicit objective that aligns consultations with user actions. Experiments on both public and commercial datasets show that VAPS consistently outperforms baselines in both retrieval and ranking tasks.

cs.IR↗

arXiv2025

OIBench: Benchmarking Strong Reasoning Models with Olympiad in Informatics

As models become increasingly sophisticated, conventional algorithm benchmarks are increasingly saturated, underscoring the need for more challenging benchmarks to guide future improvements in algorithmic reasoning. This paper introduces OIBench, a high-quality, private, and challenging olympiad-level informatics dataset comprising 250 carefully curated original problems. We detail the construction methodology of the benchmark, ensuring a comprehensive assessment across various programming paradigms and complexities, and we demonstrate its contamination-resistant properties via experiments. We propose Time/Space Completion Curves for finer-grained efficiency analysis and enable direct human-model comparisons through high-level participant evaluations. Our experiments reveal that while open-source models lag behind closed-source counterparts, current SOTA models already outperform most human participants in both correctness and efficiency, while still being suboptimal compared to the canonical solutions. By releasing OIBench as a fully open-source resource (https://huggingface.co/datasets/AGI-Eval/OIBench), we hope this benchmark will contribute to advancing code reasoning capabilities for future LLMs.

cs.AI↗

arXiv2025

Exploring the Escalation of Source Bias in User, Data, and Recommender System Feedback Loop

Recommender systems are essential for information access, allowing users to present their content for recommendation. With the rise of large language models (LLMs), AI-generated content (AIGC), primarily in the form of text, has become a central part of the content ecosystem. As AIGC becomes increasingly prevalent, it is important to understand how it affects the performance and dynamics of recommender systems. To this end, we construct an environment that incorporates AIGC to explore its short-term impact. The results from popular sequential recommendation models reveal that AIGC are ranked higher in the recommender system, reflecting the phenomenon of source bias. To further explore the long-term impact of AIGC, we introduce a feedback loop with realistic simulators. The results show that the model's preference for AIGC increases as the user clicks on AIGC rises and the model trains on simulated click data. This leads to two issues: In the short term, bias toward AIGC encourages LLM-based content creation, increasing AIGC content, and causing unfair traffic distribution. From a long-term perspective, our experiments also show that when AIGC dominates the content ecosystem after a feedback loop, it can lead to a decline in recommendation performance. To address these issues, we propose a debiasing method based on L1-loss optimization to maintain long-term content ecosystem balance. In a real-world environment with AIGC generated by mainstream LLMs, our method ensures a balance between AIGC and human-generated content in the ecosystem. The code and dataset are available at https://github.com/Yuqi-Zhou/Rec_SourceBias.

cs.IR↗

arXiv2025

Length-Induced Embedding Collapse in PLM-based Models

Text embeddings from PLM-based models enable a wide range of applications, yet their performance often degrades on longer texts. In this paper, we introduce a phenomenon we call Length Collapse, where embeddings of longer texts tend to cluster together. This clustering results in a distributional inconsistency between the embeddings of short and long texts. We further investigate how these differences contribute to the performance decline observed with longer texts across various downstream tasks. Through a rigorous theoretical analysis of the self-attention mechanism, which acts as a low-pass filter in PLM-based models, we demonstrate that as text length increases, the strength of low-pass filtering intensifies, causing embeddings to retain more low-frequency components. As a result, input token features become more similar, leading to clustering and ultimately the collapse of embeddings for longer texts. To address this issue, we propose a simple method, TempScale, which mitigates the Length Collapse phenomenon. By narrowing the gap in low-pass filtering rates between long and short texts, TempScale ensures more consistent embeddings across different text lengths. This approach leads to performance improvements of 0.94% on MTEB and 1.10% on LongEmbed, which focuses specifically on long-context retrieval, providing strong evidence for the validity of our analysis. The source code is available at https://github.com/Yuqi-Zhou/Length_Collapse.

cs.CL↗

arXiv2025

Legal Mathematical Reasoning with LLMs: Procedural Alignment through Two-Stage Reinforcement Learning

Legal mathematical reasoning is essential for applying large language models (LLMs) in high-stakes legal contexts, where outputs must be both mathematically accurate and procedurally compliant. However, existing legal LLMs lack structured numerical reasoning, and open-domain models, though capable of calculations, often overlook mandatory legal steps. To address this, we present LexNum, the first Chinese legal mathematical reasoning benchmark, covering three representative scenarios where each instance reflects legally grounded procedural flows. We further propose LexPam, a two-stage reinforcement learning framework for efficient legal reasoning training. Leveraging curriculum learning, we use a stronger teacher model to partition data into basic and challenging subsets. A lightweight 1.5B student model is then fine-tuned with Group Relative Policy Optimization, which avoids costly value networks and enables stable training from sparse, end-of-sequence rewards. The first stage improves accuracy and format; the second introduces a novel reward to guide procedural alignment via task-specific legal elements. Experiments show that existing models perform poorly on LexNum, while LexPam enhances both mathematical accuracy and legal coherence, and generalizes effectively across tasks and domains.

cs.CL↗