arXiv subjects

Yusong Zhang

Publications and source records attributed to Yusong Zhang.

7 recordsLinked to original sources

arXiv2025

Lopsided HSS Iterative Method and Preconditioner for a class of Complex Symmetric Linear System

In this study, we propose the lopsided HSS (LHSS) iteration method for solving a class of complex symmetric indefinite systems of linear equations. This method employs an alternating iterative scheme, where each iteration entails solving two systems of equations with symmetric real coefficient matrices. This design is intended to reduce the high computational costs associated with complex arithmetic. Theoretical analysis shows that the upper bound of the convergence rate depends only on the maximum and minimum eigenvalues of the real symmetric matrices, as well as the iteration parameters. When the eigenvalues satisfy certain conditions, the method guarantees convergence for any positive iteration parameter. Building on this insight, we developed the preconditioned lopsided HSS iteration method (PLHSS). Theoretical results demonstrate that PLHSS exhibits superior convergence properties compared to the original method. Additionally, we derived the optimal parameters for the new approaches and corresponding optimal convergence rate. Furthermore, we derive the PLHSS preconditioner on the basis of the iterative method. The eigenvalues of the preconditioned matrix are well-clustered, and the eigenvectors are orthogonal with a specific inner product. Numerical experiments demonstrate the efficiency of the preconditioned GMRES and COCG methods. LHSS iteration methods and the relevant preconditioners show mesh size independent and parameter-insensitive convergence behavior for the test numerical examples.

math.NA↗

arXiv2025

RLMR: Reinforcement Learning with Mixed Rewards for Creative Writing

Large language models are extensively utilized in creative writing applications. Creative writing requires a balance between subjective writing quality (e.g., literariness and emotional expression) and objective constraint following (e.g., format requirements and word limits). Existing methods find it difficult to balance these two aspects: single reward strategies fail to improve both abilities simultaneously, while fixed-weight mixed-reward methods lack the ability to adapt to different writing scenarios. To address this problem, we propose Reinforcement Learning with Mixed Rewards (RLMR), utilizing a dynamically mixed reward system from a writing reward model evaluating subjective writing quality and a constraint verification model assessing objective constraint following. The constraint following reward weight is adjusted dynamically according to the writing quality within sampled groups, ensuring that samples violating constraints get negative advantage in GRPO and thus penalized during training, which is the key innovation of this proposed method. We conduct automated and manual evaluations across diverse model families from 8B to 72B parameters. Additionally, we construct a real-world writing benchmark named WriteEval for comprehensive evaluation. Results illustrate that our method achieves consistent improvements in both instruction following (IFEval from 83.36% to 86.65%) and writing quality (72.75% win rate in manual expert pairwise evaluations on WriteEval). To the best of our knowledge, RLMR is the first work to combine subjective preferences with objective verification in online RL training, providing an effective solution for multi-dimensional creative writing optimization.

cs.AI↗

arXiv2025

Multimodal LLM Integrated Semantic Communications for 6G Immersive Experiences

6G networks promise revolutionary immersive communication experiences including augmented reality (AR), virtual reality (VR), and holographic communications. These applications demand high-dimensional multimodal data transmission and intelligent data processing in real-time, which is extremely challenging over resource-limited wireless communication systems. Moreover, a joint understanding of the environment, context, and user intent is essential to deliver task-relevant content effectively. This article presents a novel multimodal large language model (MLLM) integrated semantic communications framework, termed MLLM-SC, which fully leverages reasoning and generative capabilities of pre-trained foundation models for context-aware and task-oriented wireless communication. The MLLM-SC framework adopts a device-edge collaborative architecture. At the edge, MLLM-empowered semantic guidance module analyzes multimodal inputs, user intents, and channel conditions to generate importance-aware attention maps prioritizing semantically critical information. An importance-aware semantic encoder and a resource-adaptive semantic decoder are jointly designed and optimized, which can utilize the semantic guidance for adaptive bandwidth allocation and high-quality content reconstruction or generation. Extensive case studies on visual question answering for AR/VR applications and diffusion-driven image generation validate the effectiveness of MLLM-SC.

cs.LG↗

arXiv2025

Rumor Detection on Social Media with Reinforcement Learning-based Key Propagation Graph Generator

The spread of rumors on social media, particularly during significant events like the US elections and the COVID-19 pandemic, poses a serious threat to social stability and public health. Current rumor detection methods primarily rely on propagation graphs to improve the model performance. However, the effectiveness of these methods is often compromised by noisy and irrelevant structures in the propagation process. To tackle this issue, techniques such as weight adjustment and data augmentation have been proposed. However, they depend heavily on rich original propagation structures, limiting their effectiveness in handling rumors that lack sufficient propagation information, especially in the early stages of dissemination. In this work, we introduce the Key Propagation Graph Generator (KPG), a novel reinforcement learning-based framework, that generates contextually coherent and informative propagation patterns for events with insufficient topology information and identifies significant substructures in events with redundant and noisy propagation structures. KPG comprises two key components: the Candidate Response Generator (CRG) and the Ending Node Selector (ENS). CRG learns latent variable distributions from refined propagation patterns to eliminate noise and generate new candidates for ENS, while ENS identifies the most influential substructures in propagation graphs and provides training data for CRG. Furthermore, we develop an end-to-end framework that utilizes rewards derived from a pre-trained graph neural network to guide the training process. The resulting key propagation graphs are then employed in downstream rumor detection tasks. Extensive experiments conducted on four datasets demonstrate that KPG outperforms current state-of-the-art methods.

cs.SI↗

arXiv2024

Human-LLM Collaborative Construction of a Cantonese Emotion Lexicon

Large Language Models (LLMs) have demonstrated remarkable capabilities in language understanding and generation. Advanced utilization of the knowledge embedded in LLMs for automated annotation has consistently been explored. This study proposed to develop an emotion lexicon for Cantonese, a low-resource language, through collaborative efforts between LLM and human annotators. By integrating emotion labels provided by LLM and human annotators, the study leveraged existing linguistic resources including lexicons in other languages and local forums to construct a Cantonese emotion lexicon enriched with colloquial expressions. The consistency of the proposed emotion lexicon in emotion extraction was assessed through modification and utilization of three distinct emotion text datasets. This study not only validates the efficacy of the constructed lexicon but also emphasizes that collaborative annotation between human and artificial intelligence can significantly enhance the quality of emotion labels, highlighting the potential of such partnerships in facilitating natural language processing tasks for low-resource languages.

cs.HC↗

arXiv2024

CompassDB: Pioneering High-Performance Key-Value Store with Perfect Hash

Modern mainstream persistent key-value storage engines utilize Log-Structured Merge tree (LSM-tree) based designs, optimizing read/write performance by leveraging sequential disk I/O. However, the advent of SSDs, with their significant improvements in bandwidth and IOPS, shifts the bottleneck from I/O to CPU. The high compaction cost and large read/write amplification associated with LSM trees have become critical bottlenecks. In this paper, we introduce CompassDB, which utilizes a Two-tier Perfect Hash Table (TPH) design to significantly decrease read/write amplification and compaction costs. CompassDB utilizes a perfect hash algorithm for its in-memory index, resulting in an average index cost of about 6 bytes per key-value pair. This compact index reduces the lookup time complexity from $O(log N)$ to $O(1)$ and decreases the overall cost. Consequently, it allows for the storage of more key-value pairs for reads or provides additional memory for the memtable for writes. This results in substantial improvements in both throughput and latency. Our evaluation using the YCSB benchmark tool shows that CompassDB increases throughput by 2.5x to 4x compared to RocksDB, and by 5x to 17x compared to PebblesDB across six typical workloads. Additionally, CompassDB significantly reduces average and 99th percentile read/write latency, achieving a 50% to 85% reduction in comparison to RocksDB.

cs.DB↗

arXiv2015

Energy Efficient Resource Allocation for Control Data Separation Architecture based H-CRAN with Heterogeneous Fronthaul

Control data separation architecture (CDSA) is a more efficient architecture to overcome the overhead issue than the conventional cellular networks, especially for the huge bursty traffic like Internet of Things, and over-the-top (OTT) content service. In this paper, we study the optimization issue of network energy efficiency of the CDSA-based heterogeneous cloud radio access networks (H-CRAN) networks, which has heterogeneous fronthaul between control base station (CBS) and data base stations (DBSs). We first present a modified power consumption model for the CDSA-based H-CRAN, and then formulate the optimization problem with constraint of overall capacity of wireless fronthaul. We work out the resource assignment and power allocation by the convex relaxation approach Using fractional programming method and Lagrangian dual decomposition method, we derive the close-form optimal solution and verify it by comprehensive system-level simulation. The simulation results show that our proposed algorithm has 8% EE gain compared to the static algorithm, and the CDSA-based H-CRAN networks can achieve up to 16% EE gain compared to the conventional network even under strict fronthaul capacity limit.

cs.IT↗