Searcharxiv⌕ Search

arXiv subjects

Yuxuan Tian

Publications and source records attributed to Yuxuan Tian.

26 records · Page 2Linked to original sources

KeepKV: Achieving Periodic Lossless KV Cache Compression for Efficient LLM Inference

Efficient inference of large language models (LLMs) is hindered by an ever-growing key-value (KV) cache, making KV cache compression a critical research direction. Traditional methods selectively evict less important KV cache entries, which leads to information loss and hallucinations. Recently, merging-based strategies have been explored to retain more information by merging KV pairs that would be discarded; however, these existing approaches inevitably introduce inconsistencies in attention distributions before and after merging, causing degraded generation quality. To overcome this challenge, we propose KeepKV, a novel adaptive KV cache merging method designed to preserve performance under strict memory constraints, achieving single-step lossless compression and providing error bounds for multi-step compression. KeepKV introduces the Electoral Votes mechanism that records merging history and adaptively adjusts attention scores. Moreover, it further leverages a novel Zero Inference-Perturbation Merging method, compensating for attention loss resulting from cache merging. Extensive experiments on various benchmarks and LLM architectures demonstrate that KeepKV substantially reduces memory usage while successfully retaining essential context information, achieving over 2x inference throughput improvement and maintaining superior generation quality even with only 10% KV cache budgets.

cs.LG↗

DP-FedPGN: Finding Global Flat Minima for Differentially Private Federated Learning via Penalizing Gradient Norm

To prevent inference attacks in Federated Learning (FL) and reduce the leakage of sensitive information, Client-level Differentially Private Federated Learning (CL-DPFL) is widely used. However, current CL-DPFL methods usually result in sharper loss landscapes, which leads to a decrease in model generalization after differential privacy protection. By using Sharpness Aware Minimization (SAM), the current popular federated learning methods are to find a local flat minimum value to alleviate this problem. However, the local flatness may not reflect the global flatness in CL-DPFL. Therefore, to address this issue and seek global flat minima of models, we propose a new CL-DPFL algorithm, DP-FedPGN, in which we introduce a global gradient norm penalty to the local loss to find the global flat minimum. Moreover, by using our global gradient norm penalty, we not only find a flatter global minimum but also reduce the locally updated norm, which means that we further reduce the error of gradient clipping. From a theoretical perspective, we analyze how DP-FedPGN mitigates the performance degradation caused by DP. Meanwhile, the proposed DP-FedPGN algorithm eliminates the impact of data heterogeneity and achieves fast convergence. We also use Rényi DP to provide strict privacy guarantees and provide sensitivity analysis for local updates. Finally, we conduct effectiveness tests on both ResNet and Transformer models, and achieve significant improvements in six visual and natural language processing tasks compared to existing state-of-the-art algorithms. The code is available at https://github.com/junkangLiu0/DP-FedPGN

cs.LG↗

Detecting Flow Gaps in Data Streams

Data stream monitoring is a crucial task which has a wide range of applications. The majority of existing research in this area can be broadly classified into two types, monitoring value sum and monitoring value cardinality. In this paper, we define a third type, monitoring value variation, which can help us detect flow gaps in data streams. To realize this function, we propose GapFilter, leveraging the idea of Sketch for achieving speed and accuracy. To the best of our knowledge, this is the first work to detect flow gaps in data streams. Two key ideas of our work are the similarity absorption technique and the civilian-suspect mechanism. The similarity absorption technique helps in reducing memory usage and enhancing speed, while the civilian-suspect mechanism further boosts accuracy by organically integrating broad monitoring of overall flows with meticulous monitoring of suspicious flows.We have developed two versions of GapFilter. Speed-Oriented GapFilter (GapFilter-SO) emphasizes speed while maintaining satisfactory accuracy. Accuracy-Oriented GapFilter (GapFilter-AO) prioritizes accuracy while ensuring considerable speed. We provide a theoretical proof demonstrating that GapFilter secures high accuracy with minimal memory usage. Further, extensive experiments were conducted to assess the accuracy and speed of our algorithms. The results reveal that GapFilter-AO requires, on average, 1/32 of the memory to match the accuracy of the Straw-man solution. GapFilter-SO operates at a speed 3 times faster than the Straw-man solution. All associated source code has been open-sourced and is available on GitHub.

cs.DB↗

Hidden Sketch: A Space-Efficient Reversible Sketch for Tracking Frequent Items in Data Streams

Modern data stream applications demand memory-efficient solutions for accurately tracking frequent items, such as heavy hitters and heavy changers, under strict resource constraints. Traditional sketches face inherent accuracy-memory trade-offs: they either lose precision to reduce memory usage or inflate memory costs to enable high recording capacity. This paper introduces Hidden Sketch, a space-efficient reversible data structure for key and frequency encoding. Our design uniquely combines a Reversible Bloom Filter (RBF) and a Count-Min (CM) Sketch for invertible key and frequency storage, enabling precise reconstruction for both keys and their frequencies with minimal memory. Theoretical analysis establishes Hidden Sketch's space complexity and guaranteed reversibility, while extensive experiments demonstrate its substantial improvements in accuracy and space efficiency in frequent item tracking tasks. By eliminating the trade-off between reversibility and space efficiency, Hidden Sketch provides a scalable foundation for real-time stream analytics in resource-constrained environments.

cs.DB↗

ResidualSketch: Enhancing Layer Efficiency and Error Reduction in Hierarchical Heavy Hitter Detection with ResNet Innovations

In network management, swiftly and accurately identifying traffic anomalies, including Distributed Denial-of-Service (DDoS) attacks and unexpected network disruptions, is essential for network stability and security. Key to this process is the detection of Hierarchical Heavy Hitters (HHH), which significantly aids in the management of high-speed IP traffic. This study introduces ResidualSketch, a novel algorithm for HHH detection in hierarchical traffic analysis. ResidualSketch distinguishes itself by incorporating Residual Blocks and Residual Connections at crucial layers within the IP hierarchy, thus mitigating the Gradual Error Diffusion (GED) phenomenon in previous methods and reducing memory overhead while maintaining low update latency. Through comprehensive experiments on various datasets, we demonstrate that ResidualSketch outperforms existing state-of-the-art solutions in terms of accuracy and update speed across multiple layers of the network hierarchy. All related codes of ResidualSketch are open-source at GitHub.

cs.DS↗

FairKV: Balancing Per-Head KV Cache for Fast Multi-GPU Inference

KV cache techniques in Transformer models aim to reduce redundant computations at the expense of substantially increased memory usage, making KV cache compression an important and popular research topic. Recently, state-of-the-art KV cache compression methods implement imbalanced, per-head allocation algorithms that dynamically adjust the KV cache budget for each attention head, achieving excellent performance in single-GPU scenarios. However, we observe that such imbalanced compression leads to significant load imbalance when deploying multi-GPU inference, as some GPUs become overburdened while others remain underutilized. In this paper, we propose FairKV, a method designed to ensure fair memory usage among attention heads in systems employing imbalanced KV cache compression. The core technique of FairKV is Fair-Copying, which replicates a small subset of memory-intensive attention heads across GPUs using data parallelism to mitigate load imbalance. Our experiments on popular models, including LLaMA 70b and Mistral 24b model, demonstrate that FairKV increases throughput by 1.66x compared to standard tensor parallelism inference. Our code will be released as open source upon acceptance.

cs.DC↗

ProGroTrack: Deep Learning-Assisted Tracking of Intracellular Protein Growth Dynamics

Accurate tracking of cellular and subcellular structures, along with their dynamics, plays a pivotal role in understanding the underlying mechanisms of biological systems. This paper presents a novel approach, ProGroTrack, that combines the You Only Look Once (YOLO) and ByteTrack algorithms within the detection-based tracking (DBT) framework to track intracellular protein nanostructures. Focusing on iPAK4 protein fibers as a representative case study, we conducted a comprehensive evaluation of YOLOv5 and YOLOv8 models, revealing the superior performance of YOLOv5 on our dataset. Notably, YOLOv5x achieved an impressive mAP50 of 0.839 and F-score of 0.819. To further optimize detection capabilities, we incorporated semi-supervised learning for model improvement, resulting in enhanced performances in all metrics. Subsequently, we successfully applied our approach to track the growth behavior of iPAK4 protein fibers, revealing their two distinct growth phases consistent with a previously reported kinetic model. This research showcases the promising potential of our approach, extending beyond iPAK4 fibers. It also offers a significant advancement in precise tracking of dynamic processes in live cells, and fostering new avenues for biomedical research.

q-bio.QM↗

Fake News Identification using Machine Learning Algorithms Based on Graph Features

The spread of fake news has long been a social issue and the necessity of identifying it has become evident since its dangers are well recognized. In addition to causing uneasiness among the public, it has even more devastating consequences. For instance, it might lead to death during pandemics due to unverified medical instructions. This study aims to build a model for identifying fake news using graphs and machine learning algorithms. Instead of scanning the news content or user information, the research explicitly focuses on the spreading network, which shows the interconnection among people, and graph features such as the Eigenvector centrality, Jaccard Coefficient, and the shortest path. Fourteen features are extracted from graphs and tested in thirteen machine learning models. After analyzing these features and comparing the test result of machine learning models, the results reflect that propensity and centrality contribute highly to the classification. The best performing models reach 0.9913 and 0.9987 separately from datasets Twitter15 and Twitter16 using a modified tree classifier and Support Vector Classifier. This model can effectively predict fake news, prevent potential negative social impact caused by fake news, and provide a new perspective on graph feature selection for machine learning models.

cs.SI↗