A5092138453- Speech Recognition and Synthesis
- Topic Modeling
- Text and Document Classification Technologies
- Cloud Computing and Resource Management
- Biomedical Text Mining and Ontologies
- Educational Technology and Pedagogy
- Energy Load and Power Forecasting
- Web Data Mining and Analysis
- Stock Market Forecasting Methods
- Artificial Intelligence in Healthcare and Education
- Natural Language Processing Techniques
- Educational Technology and Assessment
- Advanced Data Storage Technologies
- Machine Learning and ELM
- Machine Learning and Data Classification
Wuhan Ship Development & Design Institute
2023-2025
Yunnan University
2024
Capital University of Economics and Business
2023
Large Language Models (LLMs) have significantly advanced natural language processing with exceptional task generalization capabilities. Low-Rank Adaption (LoRA) offers a cost-effective fine-tuning solution, freezing the original model parameters and training only lightweight, low-rank adapter matrices. However, memory footprint of LoRA is largely dominated by parameters. To mitigate this, we propose LoRAM, memory-efficient scheme founded on intuition that many neurons in over-parameterized...
This article uses machine learning algorithms to predict the latency of an IO request without additional internal information on device, treating it as a black box model equipment performance prediction. Different from such CART in machines used amounts prediction, this paper mainly analyzes impact attributes prediction and implements compares multiple different workloads. Based GBDT algorithm attribute selection algorithm, metrics (attribute sets). It proposes that using "lastdelay" can...