arXiv subjects

Meng Yuan

Publications and source records attributed to Meng Yuan.

At least 19 recordsLinked to original sources

arXiv2026

Unpaired Modality-Agnostic Generative Recommendation

Generative Recommendation (GR) formulates recommendation as autoregressive generation over discrete semantic identifiers (IDs). Although recent multimodal GR methods improve semantic ID construction with visual and textual information, they typically require item-level paired observations, restricting tokenization to the intersection of modality availability. Moreover, incorporating unpaired observations is nontrivial because small representation shifts may cross quantization boundaries and produce incompatible identifier sequences. To address this challenge, we propose \textbf{Unpair}ed Modality-Agnostic \textbf{G}enerative \textbf{R}ecommendation (UnpairGR), which learns a unified semantic-ID space from paired, image-only, and text-only observations. UnpairGR confines modality-specific processing to lightweight input projections while sharing the subsequent Transformer and residual codebooks across all observation conditions. Paired observations establish a reliability-guided cross-modal consensus, whereas unimodal observations directly refine the same representations and codes. The learned tokenizer is then fixed to provide stationary targets for a single autoregressive recommender, without feature imputation, modality-specific codebooks, or fallback mappings. Extensive experiments on three benchmark datasets demonstrate that UnpairGR consistently improves recommendation performance under both fully observed and incomplete-observation settings.

cs.IR↗

arXiv2026

Requirement--Evidence Alignment for Compositional E-Commerce Queries

Compositional e-commerce queries express multiple requirements that must hold jointly, yet existing rerankers collapse these constraints into aggregate relevance and often promote topical near misses over feasible products. In this paper, we introduce REAlign, a novel requirement-evidence-aligned reranking framework that explicitly connects typed query requirements with visible evidence. REAlign distinguishes satisfied, violated, and unsupported conditions, constructs requirement-targeted contrasts that expose failure modes, and optimizes duplicate-free partial rankings through Requirement-Aware Group-Relative Policy Optimization. Its list utility preserves relevance while incorporating requirement satisfaction, evidence support, material violations, and output validity. Experiments on two fixed-pool e-commerce benchmarks show consistent improvements over strong supervised and policy-optimization baselines under matched training budgets, with fewer violations among top-ranked candidates and larger gains at shallow ranks. Controlled ablations confirm the complementary value of requirement modeling, evidence grounding, and decomposed optimization.

cs.IR↗

arXiv2026

Ramanujan-type identities for alternating Hurwitz zeta functions

Around 1910, in an unpublished manuscript, Ramanujan proposed the following identity for $\zeta(2n+1)$: \[ \begin{aligned} \alpha^{-n}\,&\left\{\dfrac{1}{2}\,\zeta(2n + 1) + \sum_{m = 1}^{\infty}\dfrac{m^{-2n - 1}}{e^{2\alpha m} - 1}\right\} \\ &\quad\quad\quad\quad\quad\quad\quad-(-\beta)^{-n}\,\left\{\dfrac{1}{2}\,\zeta(2n + 1) + \sum_{m = 1}^{\infty}\dfrac{m^{-2n - 1}}{e^{2\beta m} - 1}\right\}\\ &=2^{2n}\sum_{k = 0}^{n + 1}\dfrac{(-1)^{k-1}B_{2k}\,B_{2n - 2k + 2}}{(2k)!(2n - 2k + 2)!}\,\alpha^{n - k + 1}\beta^k, \end{aligned} \] where $\alpha$, $\beta$ are positive numbers satisfying $\alpha\beta=\pi^2,n\in\mathbb{N},$ $B_n$ denotes the $n$-th Bernoulli number and $\zeta(z)$ is the Riemann zeta function. As shown by Berndt in the viewpoint of general transformation of analytic Eisenstein series, it is a natural companion of Euler's famous formula for even zeta values. In this paper, we extend Ramanujan's identity to the alternating Hurwitz zeta function. Then we systematically investigate the properties of the alternating Hurwitz zeta function $\zeta_E(z,x)$, as well as the corresponding Ramanujan-type identities, under different modular symmetry conditions. We also establish infinite series expressions for products of the tangent and hyperbolic tangent functions, and express the Dirichlet lambda function $\lambda(z)$ together with linear combinations of infinite series as convolution sums of special sequences. Furthermore, we define alternating Hurwitz kernels of even and odd orders, and obtain Ramanujan-type identities involving the alternating digamma function $\widetilde{\psi}(x)$ and Euler polynomials $E_n(x)$, as well as transformation formulas between even-order and odd-order alternating Hurwitz kernels.

math.NT↗

arXiv2026

SynGR: Unleashing the Potential of Cross-Modal Synergy for Generative Recommendation

Generative Recommendation (GR) has emerged as a promising paradigm by formulating item recommendation as a sequence-to-sequence generation task over item identifiers. Recent studies have incorporated multimodal signals to provide richer token-level evidence for generation. However, existing approaches largely rely on alignment-centric fusion and underexplore synergistic information across modalities. In practice, synergistic information plays a critical role in capturing emergent item properties that cannot be inferred from any single modality alone. Such properties encode intrinsic item semantics and guide user preferences, enabling models to move beyond surface-level feature matching. To address this limitation, we propose \textbf{SynGR}, a synergistic generative recommendation framework that explicitly encourages the exploitation of cross-modal dependencies during generation. By constraining overreliance on dominant modalities, SynGR enables the model to capture emergent item semantics beyond shared or modality-specific signals. Extensive experiments across three benchmark datasets demonstrate that SynGR achieves superior performance.

cs.IR↗

arXiv2026

Contouring Error Bounded Control for Biaxial Systems with Structural Flexibility and Input Delay

Precision contouring control is crucial in industrial machining processes, particularly for applications such as laser and water jet cutting, where contouring accuracy directly determines product quality. This paper presents a novel control strategy for biaxial machines featuring position-dependent flexibility and input delays, ensuring that the end-effector accurately traverses the desired contour within specified contouring error bounds and system constraints. To capture the rotation dynamics for systems with mechanical vibration, we introduce a high-fidelity model and explicitly consider the input delay with augmented system states. The controller design is based on the model predictive control scheme to enforce system states staying in robust control invariant sets defined by the reference model and switched linear time-invariant control-oriented models. The proposed algorithm is not restricted to a specific shape of the curve that is being traversed. The effectiveness of the proposed control algorithm is demonstrated in an experimental environment with discretizations and input delay. The results show that a bounded contouring error can be achieved by the proposed method in a performance degradation environment with a low commissioning effort.

eess.SY↗

arXiv2026

Uniform asymptotics for a multidimensional renewal risk model with random number of delayed claims and multivariate subexponentiality

In this paper we examine a multivariate risk model, with common renewal counting process, constant interest rate, and each claim vector is accompanied by a random number of delayed claim vectors. The interest is focused on the asymptotic behavior of the entrance probability of the discounted aggregate claims into some rare-sets, over a finite and an infinite time horizon. Our results study the the case where the main claims and the delayed claims have in some sense, asymptotic equivalent tails, but also the case where the delayed claims are negligible with comparisons with the main claims. More precisely, our estimations over finite time horizon are equipped with local uniformity, and are valid under the assumption of multivariate subexponential distributions for the claim distributions. On the case of infinite time horizon we need a mild restriction on the distribution class of multivariate subexponential distributions with positive lower Karamata index. The asymptotic relations reflect completely as all the sources of randomness, under the concrete rare-sets A, and the different dependence structures as well, without loosing elegance in spite of their generality. Further, we provide some more explicit formulas, together with relaxations of some assumptions, for the claim distributions from the multivariate regular variation. For the proof of the main results on infinite time case and for the construction of examples of multivariate distributions we need some closure properties of subexponential distributions with positive lower Karamata index. Especially, we present some necessary and sufficient conditions for the closure property with respect to convolution and some sufficient conditions for the closure property with respect to product convolution. Finally, we carry out some numerical studies to show the accuracy of our asymptotic estimations.

math.PR↗

arXiv2026

Reinforcement Learning-Based Energy Management for Industrial Park with Heterogeneous Batteries under Demand Response

The integration of photovoltaic (PV) systems, stationary energy storage systems (ESSs), and electric vehicles (EVs) alongside demand response (DR) programmes in industrial parks presents opportunities to reduce costs and improve renewable energy utilisation. Coordinating these resources is challenging because office and production zones have distinct operational objectives, and battery ageing costs are often ignored. This paper proposes a DR-based energy management framework that jointly optimises grid interaction costs, thermal comfort, EV departure state-of-charge requirements, carbon emissions, and battery ageing. We model heterogeneous load characteristics using a dynamic energy distribution ratio and incorporate dispatch-level ageing models for both ESS and EV batteries. The problem is formulated as a Markov decision process (MDP) and solved with a deep deterministic policy gradient (DDPG) algorithm. High-fidelity simulations using data from a practical industrial park in China show the framework maintains indoor comfort while significantly reducing total operating costs, yielding savings of 44.58\% and 40.68\% compared with a rule-based DR strategy and a conventional time-of-use arbitrage approach, respectively.

eess.SY↗

arXiv2026

Long time smooth solutions of 2-D quadratic quasilinear wave equations in exterior domains with Neumann boundary conditions

For the 3-D quadratic quasilinear wave equations in exterior domains with Dirichlet or Neumann boundary conditions, the global existence or the maximal existence time of small data smooth solutions have been established in the past. However, so far it is still open for the corresponding 2-D Neumann boundary value problem. In this paper, we investigate the long time existence of small data solutions to 2-D quadratic quasilinear wave equations with homogeneous Neumann boundary values. Our main ingredients include: establishing some new pointwise spacetime decay estimates for the 2-D initial boundary value problem of the divergence form wave equations, and introducing a series of good unknowns to derive the required energy estimates. The obtained results can be directly applied to the initial boundary value problem of 2-D isentropic and irrotational compressible Euler equations for both the polytropic gases and the Chaplygin gases in exterior domains with impermeable conditions, the 2-D relativistic membrane equations and 2-D membrane equations with homogeneous Neumann boundary values.

math.AP↗

arXiv2026

Reinforcement Learning-based Home Energy Management with Heterogeneous Batteries and Stochastic EV Behaviour

The widespread adoption of photovoltaic (PV), electric vehicles (EVs), and stationary energy storage systems (ESS) in households increases system complexity while simultaneously offering new opportunities for energy regulation. However, effectively coordinating these resources under uncertainties remains challenging. This paper proposes a novel home energy management framework based on deep reinforcement learning (DRL) that can jointly minimise energy expenditure and battery degradation while guaranteeing occupant comfort and EV charging requirements. Distinct from existing studies, we explicitly account for the heterogeneous degradation characteristics of stationary and EV batteries in the optimisation, alongside stochastic user behaviour regarding arrival time, departure time, and driving distance. The energy scheduling problem is formulated as a constrained Markov decision process (CMDP) and solved using a Lagrangian soft actor-critic (SAC) algorithm. This approach enables the agent to learn optimal control policies that enforce physical constraints, including indoor temperature bounds and target EV state of charge upon departure, despite stochastic uncertainties. Numerical simulations over a one-year horizon demonstrate the effectiveness of the proposed framework in satisfying physical constraints while eliminating thermal oscillations and achieving significant economic benefits. Specifically, the method reduces the cumulative operating cost substantially compared to two standard rule-based baselines while simultaneously decreasing battery degradation costs by 8.44%.

eess.SY↗

arXiv2026

You Only Anonymize What Is Not Intent-Relevant: Suppressing Non-Intent Privacy Evidence

Anonymizing sensitive information in user text is essential for privacy, yet existing methods often apply uniform treatment across attributes, which can conflict with communicative intent and obscure necessary information. This is particularly problematic when personal attributes are integral to expressive or pragmatic goals. The central challenge lies in determining which attributes to protect, and to what extent, while preserving semantic and pragmatic functions. We propose IntentAnony, a utility-preserving anonymization approach that performs intent-conditioned exposure control. IntentAnony models pragmatic intent and constructs privacy inference evidence chains to capture how distributed cues support attribute inference. Conditioned on intent, it assigns each attribute an exposure budget and selectively suppresses non-intent inference pathways while preserving intent-relevant content, semantic structure, affective nuance, and interactional function. We evaluate IntentAnony using privacy inference success rates, text utility metrics, and human evaluation. The results show an approximately 30% improvement in the overall privacy--utility trade-off, with notably stronger usability of anonymized text compared to prior state-of-the-art methods. Our code is available at https://github.com/Nevaeh7/IntentAnony.

cs.CR↗

arXiv2025

DUET: Dual Model Co-Training for Entire Space CTR Prediction

The pre-ranking stage plays a pivotal role in large-scale recommender systems but faces an intrinsic trade-off between model expressiveness and computational efficiency. Owing to the massive candidate pool and strict latency constraints, industry systems often rely on lightweight two-tower architectures, which are computationally efficient yet limited in estimation capability. As a result, they struggle to capture the complex synergistic and suppressive relationships among candidate items, which are essential for producing contextually coherent and diverse recommendation lists. Moreover, this simplicity further amplifies the Sample Selection Bias (SSB) problem, as coarse-grained models trained on biased exposure data must generalize to a much larger candidate space with distinct distributions. To address these issues, we propose \textbf{DUET} (\textbf{DU}al Model Co-Training for \textbf{E}ntire Space C\textbf{T}R Prediction), a set-wise pre-ranking framework that achieves expressive modeling under tight computational budgets. Instead of scoring items independently, DUET performs set-level prediction over the entire candidate subset in a single forward pass, enabling information-aware interactions among candidates while amortizing the computational cost across the set. Moreover, a dual model co-training mechanism extends supervision to unexposed items via mutual pseudo-label refinement, effectively mitigating SSB. Validated through extensive offline experiments and online A/B testing, DUET consistently outperforms state-of-the-art baselines and achieves improvements across multiple core business metrics. At present, DUET has been fully deployed in Kuaishou and Kuaishou Lite Apps, serving the main traffic for hundreds of millions of users.

cs.IR↗

arXiv2025

Performance Consistency of Learning Methods for Information Retrieval Tasks

A range of approaches have been proposed for estimating the accuracy or robustness of the measured performance of IR methods. One is to use bootstrapping of test sets, which, as we confirm, provides an estimate of variation in performance. For IR methods that rely on a seed, such as those that involve machine learning, another approach is to use a random set of seeds to examine performance variation. Using three different IR tasks we have used such randomness to examine a range of traditional statistical learning models and transformer-based learning models. While the statistical models are stable, the transformer models show huge variation as seeds are changed. In 9 of 11 cases the F1-scores (in the range 0.0--1.0) had a standard deviation of over 0.075; while 7 of 11 precision values (also in the range 0.0--1.0) had a standard deviation of over 0.125. This is in a context where differences of less than 0.02 have been used as evidence of method improvement. Our findings highlight the vulnerability of transformer models to training instabilities and moreover raise questions about the reliability of previous results, thus underscoring the need for rigorous evaluation practices.

cs.IR↗

arXiv2025

Deep reinforcement learning-based joint real-time energy scheduling for green buildings with heterogeneous battery energy storage devices

Green buildings (GBs) with renewable energy and building energy management systems (BEMS) enable efficient energy use and support sustainable development. Electric vehicles (EVs), as flexible storage resources, enhance system flexibility when integrated with stationary energy storage systems (ESS) for real-time scheduling. However, differing degradation and operational characteristics of ESS and EVs complicate scheduling strategies. This paper proposes a model-free deep reinforcement learning (DRL) method for joint real-time scheduling based on a combined battery system (CBS) integrating ESS and EVs. We develop accurate degradation models and cost estimates, prioritize EV travel demands, and enable collaborative ESS-EV operation under varying conditions. A prediction model optimizes energy interaction between CBS and BEMS. To address heterogeneous states, action coupling, and learning efficiency, the DRL algorithm incorporates double networks, a dueling mechanism, and prioritized experience replay. Experiments show a 37.94 percent to 40.01 percent reduction in operating costs compared to a mixed-integer linear programming (MILP) approach.

eess.SY↗

arXiv2025

FCKT: Fine-Grained Cross-Task Knowledge Transfer with Semantic Contrastive Learning for Targeted Sentiment Analysis

In this paper, we address the task of targeted sentiment analysis (TSA), which involves two sub-tasks, i.e., identifying specific aspects from reviews and determining their corresponding sentiments. Aspect extraction forms the foundation for sentiment prediction, highlighting the critical dependency between these two tasks for effective cross-task knowledge transfer. While most existing studies adopt a multi-task learning paradigm to align task-specific features in the latent space, they predominantly rely on coarse-grained knowledge transfer. Such approaches lack fine-grained control over aspect-sentiment relationships, often assuming uniform sentiment polarity within related aspects. This oversimplification neglects contextual cues that differentiate sentiments, leading to negative transfer. To overcome these limitations, we propose FCKT, a fine-grained cross-task knowledge transfer framework tailored for TSA. By explicitly incorporating aspect-level information into sentiment prediction, FCKT achieves fine-grained knowledge transfer, effectively mitigating negative transfer and enhancing task performance. Experiments on three datasets, including comparisons with various baselines and large language models (LLMs), demonstrate the effectiveness of FCKT. The source code is available on https://github.com/cwei01/FCKT.

cs.CL↗

arXiv2025

Benchmarking Large Multimodal Models for Ophthalmic Visual Question Answering with OphthalWeChat

Purpose: To develop a bilingual multimodal visual question answering (VQA) benchmark for evaluating VLMs in ophthalmology. Methods: Ophthalmic image posts and associated captions published between January 1, 2016, and December 31, 2024, were collected from WeChat Official Accounts. Based on these captions, bilingual question-answer (QA) pairs in Chinese and English were generated using GPT-4o-mini. QA pairs were categorized into six subsets by question type and language: binary (Binary_CN, Binary_EN), single-choice (Single-choice_CN, Single-choice_EN), and open-ended (Open-ended_CN, Open-ended_EN). The benchmark was used to evaluate the performance of three VLMs: GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B-Instruct. Results: The final OphthalWeChat dataset included 3,469 images and 30,120 QA pairs across 9 ophthalmic subspecialties, 548 conditions, 29 imaging modalities, and 68 modality combinations. Gemini 2.0 Flash achieved the highest overall accuracy (0.548), outperforming GPT-4o (0.522, P < 0.001) and Qwen2.5-VL-72B-Instruct (0.514, P < 0.001). It also led in both Chinese (0.546) and English subsets (0.550). Subset-specific performance showed Gemini 2.0 Flash excelled in Binary_CN (0.687), Single-choice_CN (0.666), and Single-choice_EN (0.646), while GPT-4o ranked highest in Binary_EN (0.717), Open-ended_CN (BLEU-1: 0.301; BERTScore: 0.382), and Open-ended_EN (BLEU-1: 0.183; BERTScore: 0.240). Conclusions: This study presents the first bilingual VQA benchmark for ophthalmology, distinguished by its real-world context and inclusion of multiple examinations per patient. The dataset reflects authentic clinical decision-making scenarios and enables quantitative evaluation of VLMs, supporting the development of accurate, specialized, and trustworthy AI systems for eye care.

cs.CV↗

arXiv2025

Lifelong reinforcement learning for health-aware fast charging of lithium-ion batteries

Fast charging of lithium-ion batteries remains a critical bottleneck for widespread adoption of electric vehicles and stationary energy storage systems, as improperly designed fast charging can accelerate battery degradation and shorten lifespan. In this work, we address this challenge by proposing a health-aware fast charging strategy that explicitly balances charging speed and battery longevity across the entire service life. The key innovation lies in establishing a mapping between anode overpotential and the state of health (SoH) of battery, which is then used to constrain the terminal charging voltage in a twin delayed deep deterministic policy gradient (TD3) framework. By incorporating this SoH-dependent voltage constraint, our designed deep learning method mitigates side reactions and effectively extends battery life. To validate the proposed approach, a high-fidelity single particle model with electrolyte is implemented in the widely adopted PyBaMM simulation platform, capturing degradation phenomena at realistic scales. Comparative life-cycle simulations against conventional CC-CV, its variants, and constant current-constant overpotential methods show that the TD3-based controller reduces overall degradation while maintaining competitively fast charge times. These results demonstrate the practical viability of deep reinforcement learning for advanced battery management systems and pave the way for future explorations of health-aware, performance-optimized charging strategies.

eess.SY↗

arXiv2025

Hyperbolic Diffusion Recommender Model

Diffusion models (DMs) have emerged as the new state-of-the-art family of deep generative models. To gain deeper insights into the limitations of diffusion models in recommender systems, we investigate the fundamental structural disparities between images and items. Consequently, items often exhibit distinct anisotropic and directional structures that are less prevalent in images. However, the traditional forward diffusion process continuously adds isotropic Gaussian noise, causing anisotropic signals to degrade into noise, which impairs the semantically meaningful representations in recommender systems. Inspired by the advancements in hyperbolic spaces, we propose a novel \textit{\textbf{H}yperbolic} \textit{\textbf{D}iffusion} \textit{\textbf{R}ecommender} \textit{\textbf{M}odel} (named HDRM). Unlike existing directional diffusion methods based on Euclidean space, the intrinsic non-Euclidean structure of hyperbolic space makes it particularly well-adapted for handling anisotropic diffusion processes. In particular, we begin by formulating concepts to characterize latent directed diffusion processes within a geometrically grounded hyperbolic space. Subsequently, we propose a novel hyperbolic latent diffusion process specifically tailored for users and items. Drawing upon the natural geometric attributes of hyperbolic spaces, we impose structural restrictions on the space to enhance hyperbolic diffusion propagation, thereby ensuring the preservation of the intrinsic topology of user-item graphs. Extensive experiments on three benchmark datasets demonstrate the effectiveness of HDRM.

cs.IR↗

arXiv2025

Bridging Social Psychology and LLM Reasoning: Conflict-Aware Meta-Review Generation via Cognitive Alignment

The rapid growth of scholarly submissions has overwhelmed traditional peer review systems, driving the need for intelligent automation to preserve scientific rigor. While large language models (LLMs) show promise in automating manuscript critiques, their ability to synthesize high-stakes meta-reviews, which require conflict-aware reasoning and consensus derivation, remains underdeveloped. Existing methods fail to effectively handle conflicting viewpoints within differing opinions, and often introduce additional cognitive biases, such as anchoring effects and conformity bias.To overcome these limitations, we propose the Cognitive Alignment Framework (CAF), a dual-process architecture that transforms LLMs into adaptive scientific arbitrators. By operationalizing Kahneman's dual-process theory, CAF introduces a three-step cognitive pipeline: review initialization, incremental integration, and cognitive alignment.Empirical validation shows that CAF outperforms existing LLM-based methods, with sentiment consistency gains reaching up to 19.47\% and content consistency improving by as much as 12.95\%.

cs.AI↗