arXiv subjects

Zhu Sun

Publications and source records attributed to Zhu Sun.

At least 37 records · Page 2Linked to original sources

arXiv2026

VulReaD: Knowledge-Graph-guided Software Vulnerability Reasoning and Detection

Software vulnerability detection (SVD) is a critical challenge in modern systems. Large language models (LLMs) offer natural-language explanations alongside predictions, but most work focuses on binary evaluation, and explanations often lack semantic consistency with Common Weakness Enumeration (CWE) categories. We propose VulReaD, a knowledge-graph-guided approach for vulnerability reasoning and detection that moves beyond binary classification toward CWE-level reasoning. VulReaD leverages a security knowledge graph (KG) as a semantic backbone and uses a strong teacher LLM to generate CWE-consistent contrastive reasoning supervision, enabling student model training without manual annotations. Students are fine-tuned with Odds Ratio Preference Optimization (ORPO) to encourage taxonomy-aligned reasoning while suppressing unsupported explanations. Across three real-world datasets, VulReaD improves binary F1 by 8-10% and multi-class classification by 30% Macro-F1 and 18% Micro-F1 compared to state-of-the-art baselines. Results show that LLMs outperform deep learning baselines in binary detection and that KG-guided reasoning enhances CWE coverage and interpretability.

cs.SE↗

arXiv2026

MMGRid: Navigating Temporal-aware and Cross-domain Generative Recommendation via Model Merging

Model merging (MM) offers an efficient mechanism for integrating multiple specialized models without access to original training data or costly retraining. While MM has demonstrated success in domains like computer vision, its role in recommender systems (RSs) remains largely unexplored. Recently, Generative Recommendation (GR) has emerged as a new paradigm in RSs, characterized by rapidly growing model scales and substantial computational costs, making MM particularly appealing for cost-sensitive deployment scenarios. In this work, we present the first systematic study of MM in GR through a contextual lens. We focus on a fundamental yet underexplored challenge in real-world: how to merge generative recommenders specialized to different real-world contexts, arising from temporal evolving user behaviors and heterogeneous application domains. To this end, we propose a unified framework MMGRid, a structured contextual grid of GR checkpoints that organizes models trained under diverse contexts induced by temporal evolution and domain diversity. All checkpoints are derived from a shared base LLM but fine-tuned on context-specific data, forming a realistic and controlled model space for systematically analyzing MM across GR paradigms and merging algorithms. Our investigation reveals several key insights. First, training GR models from LLMs can introduce parameter conflicts during merging due to token distribution shifts and objective disparities; such conflicts can be alleviated by disentangling task-aware and context-specific parameter changes via base model replacement. Second, incremental training across contexts induces recency bias, which can be effectively balanced through weighted contextual merging. Notably, we observe that optimal merging weights correlate with context-dependent interaction characteristics, offering practical guidance for weight selection in real-world deployments.

cs.IR↗

arXiv2026

Diagnostic-Guided Dynamic Profile Optimization for LLM-based User Simulators in Sequential Recommendation

Recent advances in large language models (LLMs) have enabled realistic user simulators for developing and evaluating recommender systems (RSs). However, existing LLM-based simulators for RSs face two major limitations: (1) static and single-step prompt-based inference that leads to inaccurate and incomplete user profile construction; (2) unrealistic and single-round recommendation-feedback interaction pattern that fails to capture real-world scenarios. To address these limitations, we propose DGDPO (Diagnostic-Guided Dynamic Profile Optimization), a novel framework that constructs user profile through a dynamic and iterative optimization process to enhance the simulation fidelity. Specifically, DGDPO incorporates two core modules within each optimization loop: firstly, a specialized LLM-based diagnostic module, calibrated through our novel training strategy, accurately identifies specific defects in the user profile. Subsequently, a generalized LLM-based treatment module analyzes the diagnosed defect and generates targeted suggestions to refine the profile. Furthermore, unlike existing LLM-based user simulators that are limited to single-round interactions, we are the first to integrate DGDPO with sequential recommenders, enabling a bidirectional evolution where user profiles and recommendation strategies adapt to each other over multi-round interactions. Extensive experiments conducted on three real-world datasets demonstrate the effectiveness of our proposed framework.

cs.IR↗

arXiv2026

A Comparative Study of Traditional Machine Learning, Deep Learning, and Large Language Models for Mental Health Forecasting using Smartphone Sensing Data

Smartphone sensing offers an unobtrusive and scalable way to track daily behaviors linked to mental health, capturing changes in sleep, mobility, and phone use that often precede symptoms of stress, anxiety, or depression. While most prior studies focus on detection that responds to existing conditions, forecasting mental health enables proactive support through Just-in-Time Adaptive Interventions. In this paper, we present the first comprehensive benchmarking study comparing traditional machine learning (ML), deep learning (DL), and large language model (LLM) approaches for mental health forecasting using the College Experience Sensing (CES) dataset, the most extensive longitudinal dataset of college student mental health to date. We systematically evaluate models across temporal windows, feature granularities, personalization strategies, and class imbalance handling. Our results show that DL models, particularly Transformer (Macro-F1 = 0.58), achieve the best overall performance, while LLMs show strength in contextual reasoning but weaker temporal modeling. Personalization substantially improves forecasts of severe mental health states. By revealing how different modeling approaches interpret phone sensing behavioral data over time, this work lays the groundwork for next-generation, adaptive, and human-centered mental health technologies that can advance both research and real-world well-being.

cs.LG↗

arXiv2025

A Reinforcement Learning Method to Factual and Counterfactual Explanations for Session-based Recommendation

Session-based Recommendation (SR) systems have recently achieved considerable success, yet their complex, "black box" nature often obscures why certain recommendations are made. Existing explanation methods struggle to pinpoint truly influential factors, as they frequently depend on static user profiles or fail to grasp the intricate dynamics within user sessions. In response, we introduce FCESR (Factual and Counterfactual Explanations for Session-based Recommendation), a novel framework designed to illuminate SR model predictions by emphasizing both the sufficiency (factual) and necessity (counterfactual) of recommended items. By recasting explanation generation as a combinatorial optimization challenge and leveraging reinforcement learning, our method uncovers the minimal yet critical sequence of items influencing recommendations. Moreover, recognizing the intrinsic value of robust explanations, we innovatively utilize these factual and counterfactual insights within a contrastive learning paradigm, employing them as high-quality positive and negative samples to fine-tune and significantly enhance SR accuracy. Extensive qualitative and quantitative evaluations across diverse datasets and multiple SR architectures confirm that our framework not only boosts recommendation accuracy but also markedly elevates the quality and interpretability of explanations, thereby paving the way for more transparent and trustworthy recommendation systems.

cs.SI↗

arXiv2025

HyMoERec: Hybrid Mixture-of-Experts for Sequential Recommendation

We propose HyMoERec, a novel sequential recommendation framework that addresses the limitations of uniform Position-wise Feed-Forward Networks in existing models. Current approaches treat all user interactions and items equally, overlooking the heterogeneity in user behavior patterns and diversity in item complexity. HyMoERec initially introduces a hybrid mixture-of-experts architecture that combines shared and specialized expert branches with an adaptive expert fusion mechanism for the sequential recommendation task. This design captures diverse reasoning for varied users and items while ensuring stable training. Experiments on MovieLens-1M and Beauty datasets demonstrate that HyMoERec consistently outperforms state-of-the-art baselines.

cs.IR↗

arXiv2025

Language Model Evolutionary Algorithms for Recommender Systems: Benchmarks and Algorithm Comparisons

In the evolutionary computing community, the remarkable language-handling capabilities and reasoning power of large language models (LLMs) have significantly enhanced the functionality of evolutionary algorithms (EAs), enabling them to tackle optimization problems involving structured language or program code. Although this field is still in its early stages, its impressive potential has led to the development of various LLM-based EAs. To effectively evaluate the performance and practical applicability of these LLM-based EAs, benchmarks with real-world relevance are essential. In this paper, we focus on LLM-based recommender systems (RSs) and introduce a benchmark problem set, named RSBench, specifically designed to assess the performance of LLM-based EAs in recommendation prompt optimization. RSBench emphasizes session-based recommendations, aiming to discover a set of Pareto optimal prompts that guide the recommendation process, providing accurate, diverse, and fair recommendations. We develop three LLM-based EAs based on established EA frameworks and experimentally evaluate their performance using RSBench. Our study offers valuable insights into the application of EAs in LLM-based RSs. Additionally, we explore key components that may influence the overall performance of the RS, providing meaningful guidance for future research on the development of LLM-based EAs in RSs. The source code of the proposed RSBench can be found at https://github.com/LiuJ-2023/RSBench/tree/main.

cs.NE↗

arXiv2025

Mitigating Strategy-Selection Bias in Reasoning for More Effective Test-Time Scaling

Test-time scaling (TTS) has been shown to improve the performance of large language models (LLMs) by sampling and aggregating diverse reasoning paths. However, existing research has overlooked a critical issue: selection bias of reasoning strategies during scaling. Specifically, when generating reasoning processes, LLMs tend to follow certain strategies (e.g., algebraic solutions for math problems) while neglecting other valid alternatives (e.g., geometric solutions), resulting in insufficient exploration of the solution space. To further understand the impact of this bias, we present a theoretical analysis that reveals when it undermines the effectiveness of test-time scaling. Motivated by this theoretical insight, we introduce TTS-Uniform, a framework designed to mitigate the selection bias of reasoning strategies. It (i) identifies potential strategies, (ii) uniformly allocates the sampling budget across them, and (iii) filters out unstable strategies prior to aggregation. Experimental results show that TTS-Uniform significantly enhances scaling effectiveness across multiple mainstream LLMs and benchmark datasets.

cs.AI↗

arXiv2025

Research on Conversational Recommender System Considering Consumer Types

Conversational Recommender Systems (CRS) provide personalized services through multi-turn interactions, yet most existing methods overlook users' heterogeneous decision-making styles and knowledge levels, which constrains both accuracy and efficiency. To address this gap, we propose CT-CRS (Consumer Type-Enhanced Conversational Recommender System), a framework that integrates consumer type modeling into dialogue recommendation. Based on consumer type theory, we define four user categories--dependent, efficient, cautious, and expert--derived from two dimensions: decision-making style (maximizers vs. satisficers) and knowledge level (high vs. low). CT-CRS employs interaction histories and fine-tunes the large language model to automatically infer user types in real time, avoiding reliance on static questionnaires. We incorporate user types into state representation and design a type-adaptive policy that dynamically adjusts recommendation granularity, diversity, and attribute query complexity. To further optimize the dialogue policy, we adopt Inverse Reinforcement Learning (IRL), enabling the agent to approximate expert-like strategies conditioned on consumer type. Experiments on LastFM, Amazon-Book, and Yelp show that CTCRS improves recommendation success rate and reduces interaction turns compared to strong baselines. Ablation studies confirm that both consumer type modeling and IRL contribute significantly to performance gains. These results demonstrate that CT-CRS offers a scalable and interpretable solution for enhancing CRS personalization through the integration of psychological modeling and advanced policy optimization.

cs.IR↗

arXiv2025

Routing Distilled Knowledge via Mixture of LoRA Experts for Large Language Model based Bundle Generation

Large Language Models (LLMs) have shown potential in automatic bundle generation but suffer from prohibitive computational costs. Although knowledge distillation offers a pathway to more efficient student models, our preliminary study reveals that naively integrating diverse types of distilled knowledge from teacher LLMs into student LLMs leads to knowledge conflict, negatively impacting the performance of bundle generation. To address this, we propose RouteDK, a framework for routing distilled knowledge through a mixture of LoRA expert architecture. Specifically, we first distill knowledge from the teacher LLM for bundle generation in two complementary types: high-level knowledge (generalizable rules) and fine-grained knowledge (session-specific reasoning). We then train knowledge-specific LoRA experts for each type of knowledge together with a base LoRA expert. For effective integration, we propose a dynamic fusion module, featuring an input-aware router, where the router balances expert contributions by dynamically determining optimal weights based on input, thereby effectively mitigating knowledge conflicts. To further improve inference reliability, we design an inference-time enhancement module to reduce variance and mitigate suboptimal reasoning. Experiments on three public datasets show that our RouteDK achieves accuracy comparable to or even better than the teacher LLM, while maintaining strong computational efficiency. In addition, it outperforms state-of-the-art approaches for bundle generation.

cs.CL↗

arXiv2025

Space and Time Cost of Continuous Rotations in Surface Codes

While Clifford operations are relatively easy to implement in fault-tolerant quantum computers,continuous rotation gates remain a significant bottleneck in typical quantum algorithms. In this work, we ask the question: "What is the most efficient approach for implementing continuous rotations in a surface code architecture?" Several techniques have been developed to reduce the T-count or T-depth of rotations, such as Hamming weight phasing and catalyst towers. However, these methods often require additional a number of ancilla qubits, and thus the ultimate cost function one needs to optimise against should rather be the total runtime or the total space required for performing a rotation. We explicitly construct surface code layouts for catalyst towers in two practical application examples in the context of option pricing: (a) implementing a phase oracle circuit, which is a ubiquitous subroutine in many quantum algorithms, and (b) state preparation using a variational quantum circuit. Our analysis shows that, at small and medium code distances, catalyst towers not only reduce the runtime but can also decrease the total spacetime volume of rotations. However, at large code distances, conventional Clifford+T synthesis may prove more efficient. Additionally, we note that our conclusions are sensitive to specific application scenarios and the choices of various parameters. Nevertheless, catalyst towers may be particularly advantageous for early fault-tolerant quantum applications, where low and medium code distances are assumed and a spacetime tradeoff is needed to reduce the runtime of individual circuit runs, such as in scenarios involving high circuit repetition counts.

quant-ph↗

arXiv2025

Low Depth Phase Oracle Using a Parallel Piecewise Circuit

We explore the important task of applying a phase $\exp(i\,f(x))$ to a computational basis state $\left| x \right>$. The closely related task of rotating a target qubit by an angle depending on $f(x)$ is also studied. Such operations are key in many quantum subroutines, and frequently $f(x)$ can be well-approximated by a piecewise function; examples range from the application of diagonal Hamiltonian terms (such as the Coulomb interaction) in grid-based many-body simulation, to derivative pricing algorithms. Here we exploit a parallelisation of the piecewise approach so that all constituent elementary rotations are performed simultaneously, that is, we achieve a total rotation depth of one. Moreover, we explore the use of recursive catalyst `towers' to implement these elementary rotations efficiently. We find that strategies prioritising execution speed can achieve circuit depth as low as $O(\log{n}{+}\log{S})$ for a register of $n$ qubits and a piecewise approximation of $S$ sections (presuming prior preparation of enabling resource states), albeit total qubit count then scales with $S$. In the limit of multiple repetitions of the oracle, we find that catalyst tower approaches have an $O(S\cdot n)$ T-count.

quant-ph↗

arXiv2025

LLM-Driven Dual-Level Multi-Interest Modeling for Recommendation

Recently, much effort has been devoted to modeling users' multi-interests based on their behaviors or auxiliary signals. However, existing methods often rely on heuristic assumptions, e.g., co-occurring items indicate the same interest of users, failing to capture user multi-interests aligning with real-world scenarios. While large language models (LLMs) show significant potential for multi-interest analysis due to their extensive knowledge and powerful reasoning capabilities, two key challenges remain. First, the granularity of LLM-driven multi-interests is agnostic, possibly leading to overly fine or coarse interest grouping. Second, individual user analysis provides limited insights due to the data sparsity issue. In this paper, we propose an LLM-driven dual-level multi-interest modeling framework for more effective recommendation. At the user-individual level, we exploit LLMs to flexibly allocate items engaged by users into different semantic clusters, indicating their diverse and distinct interests. To alleviate the agnostic generation of LLMs, we adaptively assign these semantic clusters to users' collaborative multi-interests learned from global user-item interactions, allowing the granularity to be automatically adjusted according to the user's behaviors using an alignment module. To alleviate the limited insights derived from individual users' behaviors, at the user-crowd level, we propose aggregating user cliques into synthesized users with rich behaviors for more comprehensive LLM-driven multi-interest analysis. We formulate a max covering problem to ensure the compactness and representativeness of synthesized users' behaviors, and then conduct contrastive learning based on their LLM-driven multi-interests to disentangle item representations among different interests. Experiments on real-world datasets show the superiority of our approach against state-of-the-art methods.

cs.IR↗

arXiv2025

Context-Adaptive Graph Neural Networks for Next POI Recommendation

Next Point-of-Interest (POI) recommendation is a critical task in location-based services, aiming to predict users' next visits based on their check-in histories. While many existing methods leverage Graph Neural Networks (GNNs) to incorporate collaborative information and improve recommendation accuracy, most of them model each type of context using separate graphs, treating different factors in isolation. This limits their ability to model the co-influence of multiple contextual factors on user transitions during message propagation, resulting in suboptimal attention weights and recommendation performance. Furthermore, they often prioritize sequential components as the primary predictor, potentially undermining the semantic and structural information encoded in the POI embeddings learned by GNNs. To address these limitations, we propose a Context-Adaptive Graph Neural Networks (CAGNN) for next POI recommendation, which dynamically adjusts attention weights using edge-specific contextual factors and enables mutual enhancement between graph-based and sequential components. Specifically, CAGNN introduces (1) a context-adaptive attention mechanism that jointly incorporates different types of contextual factors into the attention computation during graph propagation, enabling the model to dynamically capture collaborative and context-dependent transition patterns; (2) a graph-sequential mutual enhancement module, which aligns the outputs of the graph- and sequential-based modules via the KL divergence, enabling mutual enhancement of both components. Experimental results on three real-world datasets demonstrate that CAGNN consistently outperforms state-of-the-art methods. Meanwhile, theoretical guarantees are provided that our context-adaptive attention mechanism improves the expressiveness of POI representations.

cs.IR↗

arXiv2025

LightKG: Efficient Knowledge-Aware Recommendations with Simplified GNN Architecture

Recently, Graph Neural Networks (GNNs) have become the dominant approach for Knowledge Graph-aware Recommender Systems (KGRSs) due to their proven effectiveness. Building upon GNN-based KGRSs, Self-Supervised Learning (SSL) has been incorporated to address the sparity issue, leading to longer training time. However, through extensive experiments, we reveal that: (1)compared to other KGRSs, the existing GNN-based KGRSs fail to keep their superior performance under sparse interactions even with SSL. (2) More complex models tend to perform worse in sparse interaction scenarios and complex mechanisms, like attention mechanism, can be detrimental as they often increase learning difficulty. Inspired by these findings, we propose LightKG, a simple yet powerful GNN-based KGRS to address sparsity issues. LightKG includes a simplified GNN layer that encodes directed relations as scalar pairs rather than dense embeddings and employs a linear aggregation framework, greatly reducing the complexity of GNNs. Additionally, LightKG incorporates an efficient contrastive layer to implement SSL. It directly minimizes the node similarity in original graph, avoiding the time-consuming subgraph generation and comparison required in previous SSL methods. Experiments on four benchmark datasets show that LightKG outperforms 12 competitive KGRSs in both sparse and dense scenarios while significantly reducing training time. Specifically, it surpasses the best baselines by an average of 5.8\% in recommendation accuracy and saves 84.3\% of training time compared to KGRSs with SSL. Our code is available at https://github.com/1371149/LightKG.

cs.IR↗

arXiv2025

Reinforcement Speculative Decoding for Fast Ranking

Large Language Models (LLMs) have been widely adopted in ranking systems such as information retrieval (IR) systems and recommender systems (RSs). To alleviate the latency of auto-regressive decoding, some studies explore the single (first) token decoding for ranking approximation, but they suffer from severe degradation in tail positions. Although speculative decoding (SD) methods can be a remedy with verification at different positions, they face challenges in ranking systems due to their left-to-right decoding paradigm. Firstly, ranking systems require strict latency constraints, but verification rounds in SD methods remain agnostic; Secondly, SD methods usually discard listwise ranking knowledge about unaccepted items in previous rounds, hindering future multi-token prediction, especially when candidate tokens are the unaccepted items. In this paper, we propose a Reinforcement Speculative Decoding method for fast ranking inference of LLMs. To meet the ranking systems' latency requirement, we propose an up-to-down decoding paradigm that employs an agent to iteratively modify the ranking sequence under a constrained budget. Specifically, we design a ranking-tailored policy optimization, actively exploring optimal multi-round ranking modification policy verified by LLMs via reinforcement learning (RL). To better approximate the target LLM under the constrained budget, we trigger the agent fully utilizing the listwise ranking knowledge about all items verified by LLMs across different rounds in RL, enhancing the modification policy of the agent. More importantly, we demonstrate the theoretical robustness and advantages of our paradigm and implementation. Experiments on both IR and RS tasks show the effectiveness of our proposed method.

cs.AI↗

arXiv2025

Causal Deconfounding via Confounder Disentanglement for Dual-Target Cross-Domain Recommendation

In recent years, dual-target Cross-Domain Recommendation (CDR) has been proposed to capture comprehensive user preferences in order to ultimately enhance the recommendation accuracy in both data-richer and data-sparser domains simultaneously. However, in addition to users' true preferences, the user-item interactions might also be affected by confounders (e.g., free shipping, sales promotion). As a result, dual-target CDR has to meet two challenges: (1) how to effectively decouple observed confounders, including single-domain confounders and cross-domain confounders, and (2) how to preserve the positive effects of observed confounders on predicted interactions, while eliminating their negative effects on capturing comprehensive user preferences. To address the above two challenges, we propose a Causal Deconfounding framework via Confounder Disentanglement for dual-target Cross-Domain Recommendation, called CD2CDR. In CD2CDR, we first propose a confounder disentanglement module to effectively decouple observed single-domain and cross-domain confounders. We then propose a causal deconfounding module to preserve the positive effects of such observed confounders and eliminate their negative effects via backdoor adjustment, thereby enhancing the recommendation accuracy in each domain. Extensive experiments conducted on seven real-world datasets demonstrate that CD2CDR significantly outperforms the state-of-the-art methods.

cs.IR↗

arXiv2025

Enhancing New-item Fairness in Dynamic Recommender Systems

New-items play a crucial role in recommender systems (RSs) for delivering fresh and engaging user experiences. However, traditional methods struggle to effectively recommend new-items due to their short exposure time and limited interaction records, especially in dynamic recommender systems (DRSs) where new-items get continuously introduced and users' preferences evolve over time. This leads to significant unfairness towards new-items, which could accumulate over the successive model updates, ultimately compromising the stability of the entire system. Therefore, we propose FairAgent, a reinforcement learning (RL)-based new-item fairness enhancement framework specifically designed for DRSs. It leverages knowledge distillation to extract collaborative signals from traditional models, retaining strong recommendation capabilities for old-items. In addition, FairAgent introduces a novel reward mechanism for recommendation tailored to the characteristics of DRSs, which consists of three components: 1) a new-item exploration reward to promote the exposure of dynamically introduced new-items, 2) a fairness reward to adapt to users' personalized fairness requirements for new-items, and 3) an accuracy reward which leverages users' dynamic feedback to enhance recommendation accuracy. Extensive experiments on three public datasets and backbone models demonstrate the superior performance of FairAgent. The results present that FairAgent can effectively boost new-item exposure, achieve personalized new-item fairness, while maintaining high recommendation accuracy.

cs.IR↗