arXiv subjects

Jinyu Wang

Publications and source records attributed to Jinyu Wang.

At least 19 recordsLinked to original sources

arXiv2026

Adaptive KV Cache Reuse for Fast Long-Context LLM Serving

In long-context Large Language Model (LLM) inference, the Time-To-First-Token (TTFT) latency incurred by the prefill stage has become the foremost bottleneck limiting interactive performance and deployment cost. KV Cache reuse offers a direct path to reduce redundant prefill, yet traditional prefix caching applies only to strict-prefix scenarios; directly reusing KV Cache in non-prefix settings breaks the cross-chunk global attention relationships and causes significant degradation in generation quality. When reusable KV Cache is offloaded to GPU-external cache pools, I/O overheads across heterogeneous hardware tiers further emerge as a new TTFT bottleneck. Efficient non-prefix KV Cache reuse therefore requires both semantic-consistency recovery and compute-I/O co-optimization. This paper presents CacheTune, a frequency-guided and hardware-aware KV Cache reuse system for long-context LLM serving. CacheTune first identifies, offline, the KV pairs most critical to cross-attention recovery through frequency-domain analysis, and then selectively recomputes only these semantic-critical tokens online while reusing the remaining KVs. To turn this semantic selection into end-to-end latency reduction, CacheTune further combines sparse KV transfer, multi-stream asynchronous overlap, deferred positional-encoding recovery, and hardware-aware adaptive recomputation-ratio tuning to balance computation and data movement across heterogeneous cache pools. Evaluations on mainstream LLMs and long-context tasks show that CacheTune achieves 3.72x-4.86x TTFT speedup and 3.93x-6.21x higher throughput while maintaining generation quality close to full recompute. Even when caches are offloaded to I/O-bound SSD/HDD storage, CacheTune sustains 2.34x-2.36x TTFT speedup through adaptive recomputation.

cs.AR↗

arXiv2026

HST-HGN: Heterogeneous Spatial-Temporal Hypergraph Networks with Bidirectional State Space Models for Global Fatigue Assessment

It remains challenging to assess driver fatigue from untrimmed videos under constrained computational budgets, due to the difficulty of modeling long-range temporal dependencies in subtle facial expressions. Some existing approaches rely on computationally heavy architectures, whereas others employ traditional lightweight pairwise graph networks, despite their limited capacity to model high-order synergies and global temporal context. Therefore, we propose HST-HGN, a novel Heterogeneous Spatial-Temporal Hypergraph Network driven by Bidirectional State Space Models. Spatially, we introduce a hierarchical hypergraph network to fuse pose-disentangled geometric topologies with multi-modal texture patches dynamically. This formulation encapsulates high-order synergistic facial deformations, effectively overcoming the limitations of conventional methods. In temporal terms, a Bi-Mamba module with linear complexity is applied to perform bidirectional sequence modeling. This explicit temporal-evolution filtering enables the network to distinguish highly ambiguous transient actions, such as yawning versus speaking, while encompassing their complete physiological lifecycles. Extensive evaluations across diverse fatigue benchmarks demonstrate that HST-HGN achieves state-of-the-art performance. In particular, our method strikes a balance between discriminative power and computational efficiency, making it well-suited for real-time in-cabin edge deployment.

cs.CV↗

arXiv2026

MedFeat: Model-Aware and Explainability-Driven Feature Engineering with LLMs for Clinical Tabular Prediction

In clinical tabular prediction, classical machine learning models with feature engineering often outperform neural methods. LLMs are increasingly used to automate this process, acting as domain experts that propose diverse feature transformations to boost downstream performance. However, existing LLM-based methods decouple feature generation from the downstream model: the LLM receives no signal about which features currently drive predictions or where the model's representational capacity falls short, so proposals are neither targeted to promising regions of the feature space nor tailored to the learner's inductive bias. This shortcoming is amplified in healthcare data, which simultaneously exhibits class imbalance, heterogeneous feature spaces, and strict interpretability requirements. In this paper, we propose MedFeat, the first feature engineering framework inspired by the workflow of machine learning practitioners, leveraging model-awareness and feature importance signals to iteratively guide feature discovery for clinical tabular learning. We evaluate MedFeat on a broad range of challenging real-world clinical tasks and show that it statistically significantly outperforms state-of-the-art baselines, with an average improvement of more than 10% over the baseline across models with distinct inductive biases.

cs.LG↗

arXiv2025

Rethinking Reward Models for Multi-Domain Test-Time Scaling

The reliability of large language models (LLMs) during test-time scaling is often assessed with \emph{external verifiers} or \emph{reward models} that distinguish correct reasoning from flawed logic. Prior work has studied both outcome reward models (ORMs), which assess only the final answer, and process reward models (PRMs), which score intermediate reasoning steps. Although PRMs are often viewed as advantageous due to their finer-grained supervision, much of the supporting evidence comes from math-adjacent settings, and their relative benefits across broader domains remain unclear. We present the first unified evaluation of four reward model variants, discriminative ORM and PRM (dORM, dPRM) and generative ORM and PRM (gORM, gPRM), across 14 diverse domains. Contrary to conventional wisdom, we find that (i) dORM performs on par with dPRM, (ii) gPRM is not competitive, and (iii) overall, gORM is the most robust, yielding significant and consistent gains across every tested domain. We attribute the worse performance of gPRM to the stepwise scoring process, which inherits label noise from LLM-based automatic labeling, leading to difficulties in evaluating long reasoning trajectories, including those involving self-correcting reasoning. Both our theoretical analysis and empirical observations indicate that stepwise aggregation compounds errors as reasoning length increases. These findings challenge the common assumption that fine-grained supervision is always better and support generative outcome verification for multi-domain deployment. Our \href{https://github.com/db-Lee/Multi-RM}{\underline{code}} is publicly available to facilitate future research in multi-domain settings.

cs.AI↗

arXiv2025

In-Context Compositional Q-Learning for Offline Reinforcement Learning

Accurate estimation of the Q-function is a central challenge in offline reinforcement learning. However, existing approaches often rely on a shared global Q-function, which is inadequate for capturing the compositional structure of tasks that consist of diverse subtasks. We propose In-context Compositional Q-Learning (ICQL), an offline RL framework that formulates Q-learning as a contextual inference problem and uses linear Transformers to adaptively infer local Q-functions from retrieved transitions without explicit subtask labels. Theoretically, we show that, under two assumptions -- linear approximability of the local Q-function and accurate inference of weights from retrieved context -- ICQL achieves a bounded approximation error for the Q-function and enables near-optimal policy extraction. Empirically, ICQL substantially improves performance in offline settings, achieving gains of up to 16.4% on kitchen tasks and up to 8.8% and 6.3% on MuJoCo and Adroit tasks, respectively. These results highlight the underexplored potential of in-context learning for robust and compositional value estimation and establish ICQL as a principled and effective framework for offline RL.

cs.LG↗

arXiv2025

The Illusion of Readiness in Health AI

Large language models have demonstrated remarkable performance in a wide range of medical benchmarks. Yet underneath the seemingly promising results lie salient growth areas, especially in cutting-edge frontiers such as multimodal reasoning. In this paper, we introduce a series of adversarial stress tests to systematically assess the robustness of flagship models and medical benchmarks. Our study reveals prevalent brittleness in the presence of simple adversarial transformations: leading systems can guess the right answer even with key inputs removed, yet may get confused by the slightest prompt alterations, while fabricating convincing yet flawed reasoning traces. Using clinician-guided rubrics, we demonstrate that popular medical benchmarks vary widely in what they truly measure. Our study reveals significant competency gaps of frontier AI in attaining real-world readiness for health applications. If we want AI to earn trust in healthcare, we must demand more than leaderboard wins and must hold AI systems accountable to ensure robustness, sound reasoning, and alignment with real medical demands.

cs.AI↗

arXiv2025

Sample-efficient LLM Optimization with Reset Replay

Recent advancements in LLM post-training, particularly through reinforcement learning and preference optimization, are key to boosting their reasoning capabilities. However, these methods often suffer from low sample efficiency and a susceptibility to primacy bias, a phenomenon where overfitting to initial experiences diminishes network plasticity and damages the learning process. To address these challenges, we introduce LLM optimization with Reset Replay (LoRR), a general and powerful plugin for enhancing sample efficiency in preference-based optimization. Its core mechanism enables high-replay training to maximize the utility of each data batch. To mitigate overfitting, LoRR orchestrates a periodic reset strategy that reuses the initial data and policy to maintain network plasticity, and further adopts a hybrid optimization objective to better exploit training data. Extensive experiments show that LoRR significantly boosts the performance of various preference optimization methods on both mathematical and general reasoning benchmarks. Notably, an iterative DPO framework augmented with LoRR achieves comparable performance on challenging math tasks, rivaling many complex or computationally expensive baselines. Our findings highlight that LoRR offers a practical and sample-efficient paradigm from limited offline data, unlocking greater performance with minimal changes to existing post-training workflows.

cs.LG↗

arXiv2025

Formula-R1: Incentivizing LLM Reasoning over Complex Tables with Numerical Computation via Formula-Driven Reinforcement Learning

Tables are a fundamental medium for organizing and analyzing data, making table reasoning a critical capability for intelligent systems. Although large language models (LLMs) exhibit strong general reasoning abilities, they still struggle with accurate numerical reasoning over tabular data, particularly in complex table settings beyond simple relational lookup. Spreadsheet formulas provide a powerful and expressive interface for executable symbolic operations, enabling rich reasoning patterns that remain largely underexplored by existing LLMs. In this paper, we introduce Formula-R1, a model trained via Formula Tuning (Fortune), a formula-driven reinforcement learning (RL) framework for table reasoning. Formula Tuning trains LLMs to generate executable spreadsheet formulas for question answering over general tabular data, using execution success and answer correctness as reward signals, thereby reducing reliance on supervised formula annotations. We demonstrate the effectiveness of Formula Tuning through extensive experiments on seven table reasoning benchmarks. It substantially improves LLM performance on table reasoning, particularly for tasks involving complex tables and multi-step numerical computation. Moreover, Formula-R1 consistently outperforms prior methods under controlled comparison settings. Beyond empirical gains, our extensive analyses provide insights into the role of RL in formula-driven table reasoning, highlighting the broader potential of formula-driven RL to enhance reasoning capabilities in LLMs.

cs.AI↗

arXiv2025

KVmix: Gradient-Based Layer Importance-Aware Mixed-Precision Quantization for KV Cache

The high memory demands of the Key-Value (KV) Cache during the inference of Large Language Models (LLMs) severely restrict their deployment in resource-constrained platforms. Quantization can effectively alleviate the memory pressure caused by KV Cache. However, existing methods either rely on static one-size-fits-all precision allocation or fail to dynamically prioritize critical KV in long-context tasks, forcing memory-accuracy-throughput tradeoffs. In this work, we propose a novel mixed-precision quantization method for KV Cache named KVmix. KVmix leverages gradient-based importance analysis to evaluate how individual Key and Value projection matrices affect the model loss, enabling layer-specific bit-width allocation for mix-precision quantization. It dynamically prioritizes higher precision for important layers while aggressively quantizing less influential ones, achieving a tunable balance between accuracy and efficiency. KVmix also introduces a dynamic long-context optimization strategy that adaptively keeps full-precision KV pairs for recent pivotal tokens and compresses older ones, achieving high-quality sequence generation with low memory usage. Additionally, KVmix provides efficient low-bit quantization and CUDA kernels to optimize computational overhead. On LLMs such as Llama and Mistral, KVmix achieves near-lossless inference performance with extremely low quantization configuration (Key 2.19bit Value 2.38bit), while delivering a remarkable 4.9x memory compression and a 5.3x speedup in inference throughput.

cs.LG↗

arXiv2025

OMGM: Orchestrate Multiple Granularities and Modalities for Efficient Multimodal Retrieval

Vision-language retrieval-augmented generation (RAG) has become an effective approach for tackling Knowledge-Based Visual Question Answering (KB-VQA), which requires external knowledge beyond the visual content presented in images. The effectiveness of Vision-language RAG systems hinges on multimodal retrieval, which is inherently challenging due to the diverse modalities and knowledge granularities in both queries and knowledge bases. Existing methods have not fully tapped into the potential interplay between these elements. We propose a multimodal RAG system featuring a coarse-to-fine, multi-step retrieval that harmonizes multiple granularities and modalities to enhance efficacy. Our system begins with a broad initial search aligning knowledge granularity for cross-modal retrieval, followed by a multimodal fusion reranking to capture the nuanced multimodal information for top entity selection. A text reranker then filters out the most relevant fine-grained section for augmented generation. Extensive experiments on the InfoSeek and Encyclopedic-VQA benchmarks show our method achieves state-of-the-art retrieval performance and highly competitive answering results, underscoring its effectiveness in advancing KB-VQA systems.

cs.IR↗

arXiv2025

A Dual-Core Model for ENSO Diversity: Unifying Model Hierarchies for Realistic Simulations

Despite advances in climate modeling, simulating the El Ni\~no-Southern Oscillation (ENSO) remains challenging due to its spatiotemporal diversity and complexity. To address this, we build upon existing model hierarchies to develop a new unified modeling platform, which provides practical, scalable, and accurate tools for advancing ENSO research. Within this framework, we introduce a dual-core ENSO model (DCM) that integrates two widely used ENSO modeling approaches: a linear stochastic model confined to the equator and a nonlinear intermediate model extending off-equator. The stochastic model ensures computational efficiency and statistical accuracy. It captures essential ENSO characteristics and reproduces the observed non-Gaussian statistics. Meanwhile, the nonlinear model assimilates pseudo-observations from the stochastic model while resolving key air-sea interactions, such as feedback balances and spatial patterns of sea surface temperature anomalies (SSTA) during El Ni\~no peaks and improving western-central Pacific SSTA magnitudes and spatial accuracy. The DCM effectively captures the realistic dynamical and statistical features of the ENSO diversity and complexity. Notably, the computational efficiency of the DCM facilitates a rapid generation of extended ENSO datasets, overcoming observational limitations. The outcome facilitates the analysis of long-term variations, advancing our understanding of ENSO and many other climate phenomena.

physics.ao-ph↗

arXiv2025

PDA Construction via Union of Cartesian Product Cache Configurations for Coded Caching

Caching is an efficient technique to reduce peak traffic by storing popular content in local caches. Placement delivery array (PDA) proposed by Yan et al. is a combinatorial structure to design coded caching schemes with uncoded placement and one-shot linear delivery. By taking the $m$-fold Cartesian product of a small base PDA, Wang et al. constructed a big PDA while maintaining the memory ratio and transmission load unchanged, which achieves linear growth in both the number of users and coded caching gain. In order to achieve exponential growth in both the number of users and coded caching gain, in this paper we propose a PDA construction by taking the union operation of the cache configurations from the $m$-fold Cartesian product of a base PDA. The resulting PDA leads to a coded caching scheme with subpacketization increasing sub-exponentially with the number of users while keeping the load constant for fixed memory ratio. By applying the proposed construction to existing base PDAs, three new coded caching schemes are obtained, which cover some existing schemes as special cases and can achieve lower load with simultaneously lower subpacketization for some memory ratios.

cs.IT↗

arXiv2025

Improved Coded Caching Scheme for Multi-User Information Retrieval System

In this paper, we study the coded caching scheme for the $(L, K, M, N)$ multi-user information retrieval (MIR) system, which consists of a content library containing $N$ files, a base station (BS) with $L$ antennas that cannot access the library, and $K$ single-antenna users, each of which can cache at most $M$ files from the library. The users communicate with the others assisted by the BS to decode their required files. In this paper, we focus on designing a coded caching scheme with low communication latency measured by normalized delivery time (NDT), computational complexity, and subpacketizations. When $\frac{KM}{N}\geq L$ we first simply the precoding matrix in the downlink step to an identity matrix and use the multiple-antenna placement delivery array (MAPDA), which was originally proposed for the multiple-input single-output networks, to generate several new schemes for MIR system. Compared to the existing schemes, both the theoretical and numerical analyses show that our new schemes achieve much lower computational complexity and smaller subpacketizations with the same NDT.

cs.IT↗

arXiv2025

PIKE-RAG: sPecIalized KnowledgE and Rationale Augmented Generation

Despite notable advancements in Retrieval-Augmented Generation (RAG) systems that expand large language model (LLM) capabilities through external retrieval, these systems often struggle to meet the complex and diverse needs of real-world industrial applications. The reliance on retrieval alone proves insufficient for extracting deep, domain-specific knowledge performing in logical reasoning from specialized corpora. To address this, we introduce sPecIalized KnowledgE and Rationale Augmentation Generation (PIKE-RAG), focusing on extracting, understanding, and applying specialized knowledge, while constructing coherent rationale to incrementally steer LLMs toward accurate responses. Recognizing the diverse challenges of industrial tasks, we introduce a new paradigm that classifies tasks based on their complexity in knowledge extraction and application, allowing for a systematic evaluation of RAG systems' problem-solving capabilities. This strategic approach offers a roadmap for the phased development and enhancement of RAG systems, tailored to meet the evolving demands of industrial applications. Furthermore, we propose knowledge atomizing and knowledge-aware task decomposition to effectively extract multifaceted knowledge from the data chunks and iteratively construct the rationale based on original query and the accumulated knowledge, respectively, showcasing exceptional performance across various benchmarks.

cs.CL↗

arXiv2024

Probabilistic degenerate Bernstein polynomials

In recent years, both degenerate versions and probabilistic extensions of many special numbers and polynomials have been explored. For instance, degenerate Bernstein polynomials and probabilistic Bernstein polynomials were investigated earlier. Assume that Y is a random variable whose moment generating function exists in a neighborhood of the origin. The aim of this paper is to study probabilistic degenerate Bernstein polynomials associated with Y which are both probabilistic extension of the degenerate Bernstein polynomials and degenerate version of the probabilistic Bernstein polynomials associated with $Y$. We derive several explicit expressions and certain related identities for those polynomials. In addition, we treat the special cases of the Poisson random variable, the Bernoulli random variable and of the binomial random variable.

math.NT↗

arXiv2024

Unveiling Markov Heads in Pretrained Language Models for Offline Reinforcement Learning

Recently, incorporating knowledge from pretrained language models (PLMs) into decision transformers (DTs) has generated significant attention in offline reinforcement learning (RL). These PLMs perform well in RL tasks, raising an intriguing question: what kind of knowledge from PLMs has been transferred to RL to achieve such good results? This work first dives into this problem by analyzing each head quantitatively and points out Markov head, a crucial component that exists in the attention heads of PLMs. It leads to extreme attention on the last-input token and performs well only in short-term environments. Furthermore, we prove that this extreme attention cannot be changed by re-training embedding layer or fine-tuning. Inspired by our analysis, we propose a general method GPT2-DTMA, which equips a pretrained DT with Mixture of Attention (MoA), to accommodate diverse attention requirements during fine-tuning. Extensive experiments corroborate our theorems and demonstrate the effectiveness of GPT2-DTMA: it achieves comparable performance in short-term environments while significantly narrowing the performance gap in long-term environments.

cs.LG↗

arXiv2024

Neural Probabilistic Logic Learning for Knowledge Graph Reasoning

Knowledge graph (KG) reasoning is a task that aims to predict unknown facts based on known factual samples. Reasoning methods can be divided into two categories: rule-based methods and KG-embedding based methods. The former possesses precise reasoning capabilities but finds it challenging to reason efficiently over large-scale knowledge graphs. While gaining the ability to reason over large-scale knowledge graphs, the latter sacrifices reasoning accuracy. This paper aims to design a reasoning framework called Neural Probabilistic Logic Learning(NPLL) that achieves accurate reasoning on knowledge graphs. Our approach introduces a scoring module that effectively enhances the expressive power of embedding networks, striking a balance between model simplicity and reasoning capabilities. We improve the interpretability of the model by incorporating a Markov Logic Network based on variational inference. We empirically evaluate our approach on several benchmark datasets, and the experimental results validate that our method substantially enhances the accuracy and quality of the reasoning results.

cs.AI↗

arXiv2024

Mildly Constrained Evaluation Policy for Offline Reinforcement Learning

Offline reinforcement learning (RL) methodologies enforce constraints on the policy to adhere closely to the behavior policy, thereby stabilizing value learning and mitigating the selection of out-of-distribution (OOD) actions during test time. Conventional approaches apply identical constraints for both value learning and test time inference. However, our findings indicate that the constraints suitable for value estimation may in fact be excessively restrictive for action selection during test time. To address this issue, we propose a \textit{Mildly Constrained Evaluation Policy (MCEP)} for test time inference with a more constrained \textit{target policy} for value estimation. Since the \textit{target policy} has been adopted in various prior approaches, MCEP can be seamlessly integrated with them as a plug-in. We instantiate MCEP based on TD3BC (Fujimoto & Gu, 2021), AWAC (Nair et al., 2020) and DQL (Wang et al., 2023) algorithms. The empirical results on D4RL MuJoCo locomotion, high-dimensional humanoid and a set of 16 robotic manipulation tasks show that the MCEP brought significant performance improvement on classic offline RL methods and can further improve SOTA methods. The codes are open-sourced at \url{https://github.com/egg-west/MCEP.git}.

cs.LG↗