A5023226704- IoT and Edge/Fog Computing
- Software System Performance and Reliability
- Cloud Computing and Resource Management
- Advanced Manufacturing and Logistics Optimization
- Natural Language Processing Techniques
- Semantic Web and Ontologies
- Advanced Decision-Making Techniques
- Data Management and Algorithms
- Speech and dialogue systems
- E-commerce and Technology Innovations
- Belt Conveyor Systems Engineering
- Big Data Technologies and Applications
- Advanced Sensor and Control Systems
- Industrial Technology and Control Systems
- Technology and Data Analysis
- Hate Speech and Cyberbullying Detection
- Network Security and Intrusion Detection
- Transportation Planning and Optimization
- Evaluation Methods in Various Fields
- AI-based Problem Solving and Planning
- Fault Detection and Control Systems
- Intelligent Tutoring Systems and Adaptive Learning
- Civil and Geotechnical Engineering Research
- Vehicle Routing Optimization Methods
- Geomechanics and Mining Engineering
Guilin University of Electronic Technology
2024
Guilin University
2024
Hôpital privé du Confluent
2023
Guangxi University
2019-2023
Large language models (LLMs) have achieved significant progress across various domains, but their increasing scale results in high computational and memory costs. Recent studies revealed that LLMs exhibit sparsity, providing the potential to reduce model size through pruning techniques. However, existing methods typically follow a prune-then-finetune paradigm. Since pruned components still contain valuable information, direct removal often leads irreversible performance degradation, imposing...
Multimodal large language models (MLLMs) exhibit impressive capabilities but still face challenges in complex visual reasoning. While recent efforts attempt to enhance MLLMs' reasoning by incorporating OpenAI o1-like structured thinking through explicit search structures or teacher-guided distillation, they often struggle balance performance and efficiency. A critical limitation is their heavy reliance on extensive data spaces, resulting low-efficiency implicit insight extraction...
Event streaming is an increasingly critical infrastructure service used in many industries and there growing demand for cloud-native solutions. Confluent Cloud provides a massive scale event platform built on top of Apache Kafka with tens thousands clusters running 70+ regions across AWS, Google Cloud, Azure. This paper introduces Kora , the at core Cloud. We describe Kora's design that enables it to meet its goals, such as reliability, elasticity, cost efficiency. discuss abstractions which...
Online services are now commonly deployed via cloud computing based on Infrastructure as a Service (IaaS) to Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS). However, workload is not constant over time, so guaranteeing the quality of service (QoS) resource cost-effectiveness, which determined by on-demand requirements, challenging issue. In this article, authors propose neural network-based-method termed domain knowledge embedding regularization networks (DKRNN) for large-scale...
In-context Learning (ICL) enables large language models (LLMs) to tackle downstream tasks through sophisticated prompting and high-quality demonstrations. However, this traditional ICL paradigm shows limitations when facing complex mathematical reasoning tasks, primarily due its heavy dependence on example quality the necessity for human intervention in challenging scenarios. To address these limitations, paper presents HiAR-ICL, a \textbf{Hi}gh-level \textbf{A}utomated \textbf{R}easoning...
With the development and popularization of computer technology, computers have played an important role in improving production efficiency. In face opportunities challenges new era, China has established goal transitioning from "manufacturing" to "intelligent manufacturing", integrating intelligent technology into production, efficiency, gaining advantage international competition. this context, Yuchai, as a leading enterprise engine manufacturing industry, task developing competitiveness....
Existed many transportation information service system does not take into account the overall operation of road network and real-time intelligent guidance to drivers. This paper focuses on analysis important factors such as scratch degree travel time that affect driver's route choice behaviour, mathematical model relative evaluation are established. On this basis, combined with multi-agent method, selection behaviour based under is constructed, Repast S tool used carry out simulation...