Profiles2023-2026
Zeyi Wen
The Hong Kong University of Science and Technology
10Papers
6Last author
14Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers10
- A Better Start: Sensitivity-Aware Warm-Up for Robust and Efficient Fine-TuningAuthor 2 of 4AAAI 2026
- Representation Drift Compensation: A Near-Zero Inference Cost Enhancement for LLM DecompositionAuthor 3 of 3ICML 2026
- SoLA: Leveraging Soft Activation Sparsity and Low-Rank Decomposition for Large Language Model CompressionAuthor 3 of 3AAAI 2025 · 14 citations
- Towards Efficient Low-Order Hybrid Optimizer for Language Model Fine-TuningAuthor 3 of 3AAAI 2025 · 6 citations
- EcoTune: Token-Efficient Multi-Fidelity Hyperparameter Optimization for Large Language Model InferenceAuthor 3 of 3EMNLP 2025 · 1 citation
- ScalaGBM: Memory Efficient GBDT Training for High-Dimensional Data on GPUAuthor 2 of 6KDD 2025
- SEAL: Structure and Element Aware Learning Improves Long Structured Document RetrievalAuthor 6 of 6EMNLP 2025
- Enhancing the Performance of Bandit-based Hyperparameter OptimizationAuthor 2 of 4ICDE 2024 · 3 citations
- Exploiting Careful Design of SVM Solution for Aspect-term Sentiment AnalysisAuthor 4 of 4EMNLP Findings 2024 · 1 citation
- Local Path Integration for AttributionAuthor 3 of 4AAAI 2023 · 16 citations
