arXiv subjects

Jiehan Zhou

Publications and source records attributed to Jiehan Zhou.

8 recordsLinked to original sources

arXiv2026

LLM-CoOpt: A Co-Design and Optimization Framework for Efficient LLM Inference on Heterogeneous Platforms

Major challenges in LLMs inference remain frequent memory bandwidth bottlenecks, computational redundancy, and inefficiencies in long-sequence processing. To address these issues, we propose LLM-CoOpt, a comprehensive algorithmhardware co-design framework aimed at improving both throughput and latency in LLM inference. LLM-CoOpt integrates three key strategies: (1) Key-Value Cache Optimization, termed Opt-KV, which improves memory access efficiency by optimizing both KV cache write and read paths, and introduces FP8 quantization to reduce memory footprint while maintaining accuracy; (2) Grouped-Query Attention for Computational Efficiency, termed Opt-GQA, which reduces the overall computational complexity by restructuring multi-head self-attention into grouped-query attention with shared key-value projections, enabling higher throughput and lower resource consumption; (3) Paged Attention for Long- Sequence Processing, termed Opt-Pa, which adopts a two-step strategy to first segment long sequences into manageable chunks and then apply lazy memory mapping and computation, significantly reducing memory pressure and improving performance on long-context inputs.Experiments on the LLaMa-13BGPTQ model demonstrate that LLM-CoOpt increases inference throughput by up to 13.43%, reduces latency by up to 16.79%, and maintains model accuracy. These results confirm that LLM-CoOpt provides a practical, high-performance optimization path for real-world inference of large-scale language models.

cs.DC↗

arXiv2026

Opt4GPTQ: Co-Optimizing Memory and Computation for 4-bit GPTQ Quantized LLM Inference on Heterogeneous Platforms

The increasing adoption of large language models (LLMs) on heterogeneous computing platforms poses significant challenges to achieving high inference efficiency. To address these efficiency bottlenecks across diverse platforms, this paper proposes Opt4GPTQ, a practical optimization method designed for 4-bit GPTQ quantized LLMs inference on heterogeneous AI accelerators. Built upon the vLLM serving system, Opt4GPTQ integrates three platform-level optimization strategies: Shared Memory Buffering Optimization (SMB-Opt), which caches frequently accessed data in shared memory and employs single-threaded writes; Vectorized Memory Loading Optimization (VML-Opt), which utilizes vectorized memory operations for efficient data loading; and Inline Assembly Optimization (ILA-Opt), which directly leverages hardwarenative vector half-precision addition and fused multiply-accumulate instructions. Experimental results show that Opt4GPTQ effectively improves performance across various models while maintaining original model accuracy, achieving throughput gains of up to 84.42%. This work highlights the critical role of platformlevel engineering in enabling efficient LLMs inference on emerging architectures and provides valuable methodologies for future heterogeneous platform adaptation.

cs.DC↗

arXiv2025

KV Cache Compression for Inference Efficiency in LLMs: A Review

Withtherapid advancement of large language models (LLMs), the context length for inference has been continuously increasing, leading to an exponential growth in the demand for Key-Value (KV) caching. This has resulted in a significant memory bottleneck, limiting the inference efficiency and scalability of the models. Therefore, optimizing the KV cache during inference is crucial for enhancing performance and efficiency. This review systematically examines current KV cache optimization techniques, including compression strategies such as selective token strategies, quantization, and attention compression. We evaluate the effectiveness, trade-offs, and application scenarios of these methods, providing a comprehensive analysis of their impact on memory usage and inference speed. We focus on identifying the limitations and challenges of existing methods, such as compatibility issues with different models and tasks. Additionally, this review highlights future research directions, including hybrid optimization techniques, adaptive dynamic strategies, and software-hardware co-design. These approaches aim to improve inference efficiency and promote the practical application of large language models.

cs.DC↗

arXiv2025

Opt-GPTQ: An Optimized GPTQ Combining Sparse Attention and Quantization Techniques

In the field of deep learning, traditional attention mechanisms face significant challenges related to high computational complexity and large memory consumption when processing long sequence data. To address these limitations, we propose Opt-GPTQ, an optimized Gradient-based Post Training Quantization (GPTQ) combining the Grouped Query Attention (GQA) mechanism with paging memory management, optimizing the traditional Multi-Head Attention (MHA) mechanism by grouping query heads and sharing key-value vectors. Optimized GQA (Opt-GQA) effectively reduces computational complexity, minimizes memory fragmentation, and enhances memory utilization for large-scale models. Opt-GPTQ is optimized for Data Center Units (DCUs) and integrated into the vLLM model to maximize hardware efficiency. It customizes GPU kernels to further enhance attention computation by reducing memory access latency and boosting parallel computing capabilities. Opt-GQA integrates Attention with Linear Biases (ALiBi) to reduce overhead and enhance long-sequence processing. Experimental results show that Opt-GPTQ significantly reduces computation time and memory usage while improving model performance.

cs.DC↗

arXiv2024

TSViT: A Time Series Vision Transformer for Fault Diagnosis

Traditional fault diagnosis methods using Convolutional Neural Networks (CNNs) often struggle with capturing the temporal dynamics of vibration signals. To overcome this, the application of Transformer-based Vision Transformer (ViT) methods to fault diagnosis is gaining attraction. Nonetheless, these methods typically require extensive preprocessing, which increases computational complexity, potentially reducing the efficiency of the diagnosis process. Addressing this gap, this paper presents the Time Series Vision Transformer (TSViT), tailored for effective fault diagnosis. TSViT incorporates a convolutional layer to extract local features from vibration signals, alongside a transformer encoder to discern long-term temporal patterns. A thorough experimental comparison on three diverse datasets demonstrates TSViT's effectiveness and adaptability. Moreover, the paper delves into the influence of hyperparameter tuning on the model's performance, computational demand, and parameter count. Remarkably, TSViT achieves an unprecedented 100% average accuracy on two test sets and 99.99% on another, showcasing its exceptional diagnostic capabilities.

eess.SY↗

arXiv2022

Revisiting Digital Twins: Origins, Fundamentals and Practices

The Digital Twins (DT) has quickly become a hot topic since it was proposed. It not only appears in all kinds of commercial propaganda, but also is widely quoted by academic circles. However, there are misstatements and misuse of the term DT in business and academy. This paper revisits Digital Twins and defines it to be a more advanced system/product/service modelling and simulation environment that combines the most modern Information Communication Technology (ICTs) and engineering mechanisms digitization, and characterized by system/product/service life cycle management, physically geometric visualization, real-time sensing and measurement of system operating conditions, predictability of system performance/safety/lifespan, complete engineering mechanisms-based simulations. The idea of Digital Twins originates from modelling and simulation practices of engineering informatization, including Virtual Manufacturing (VM), Model Predictive Control (MPC), and Building Information Model (BIM). Based on the two-element VM model, we propose a three-element model to represent Digital Twins. Digital Twins does not have its own unique technical characteristics; the existing practices of Digital Twins are extensions of the engineering informatization embracing modern ICTs. These insights clarify the origin of Digital Twins and its technical essentials.

cs.CE↗

arXiv2021

Feature-context driven Federated Meta-Learning for Rare Disease Prediction

Millions of patients suffer from rare diseases around the world. However, the samples of rare diseases are much smaller than those of common diseases. In addition, due to the sensitivity of medical data, hospitals are usually reluctant to share patient information for data fusion citing privacy concerns. These challenges make it difficult for traditional AI models to extract rare disease features for the purpose of disease prediction. In this paper, we overcome this limitation by proposing a novel approach for rare disease prediction based on federated meta-learning. To improve the prediction accuracy of rare diseases, we design an attention-based meta-learning (ATML) approach which dynamically adjusts the attention to different tasks according to the measured training effect of base learners. Additionally, a dynamic-weight based fusion strategy is proposed to further improve the accuracy of federated learning, which dynamically selects clients based on the accuracy of each local model. Experiments show that with as few as five shots, our approach out-performs the original federated meta-learning algorithm in accuracy and speed. Compared with each hospital's local model, the proposed model's average prediction accuracy increased by 13.28%.

cs.LG↗

arXiv2021

A Survey on Federated Learning and its Applications for Accelerating Industrial Internet of Things

Federated learning (FL) brings collaborative intelligence into industries without centralized training data to accelerate the process of Industry 4.0 on the edge computing level. FL solves the dilemma in which enterprises wish to make the use of data intelligence with security concerns. To accelerate industrial Internet of things with the further leverage of FL, existing achievements on FL are developed from three aspects: 1) define terminologies and elaborate a general framework of FL for accommodating various scenarios; 2) discuss the state-of-the-art of FL on fundamental researches including data partitioning, privacy preservation, model optimization, local model transportation, personalization, motivation mechanism, platform & tools, and benchmark; 3) discuss the impacts of FL from the economic perspective. To attract more attention from industrial academia and practice, a FL-transformed manufacturing paradigm is presented, and future research directions of FL are given and possible immediate applications in Industry 4.0 domain are also proposed.

cs.DC↗