Scholar
Haw-Shiuan Chang
Google Scholar ID: fdGXVd4AAAAJ
UMass Amherst
Citations & Impact
Citations
1,305
H-index
14
i10-index
16
Publications
20
Co-authors
10
list available
Publications
Resume (English only)
Academic Achievements
- Selected Publications:
- - REAL Sampling: Boosting Factuality and Diversity of Open-Ended Generation
- - Explaining and Improving Contrastive Decoding
- - To Copy, or not to Copy; That is a Critical Issue of the Output Softmax Layer in Neural Sequential Recommenders
Research Experience
- Current position: Postdoctoral Research Associate at UMass Amherst CIIR, Advisor: Hamed Zamani; Previously: Postdoctoral Scientist at Amazon AGI Foundations, Collaborators: Violet Peng, Mohit Bansal, and Tagyoung Chung; Before PhD: Worked with Professor Yu-Chiang Frank Wang and Dr. Kuan-Ta Chen at Academia Sinica, Taiwan.
Education
- PhD: University of Massachusetts Amherst, Advisor: Andrew McCallum; BS: National Yang Ming Chiao Tung University (NYCU), Taiwan, EECS Undergraduate Honors Program.
Background
- Research interests: Machine learning (ML), natural language processing (NLP), and information retrieval (IR). Currently focused on fundamentally narrowing the gap between large language models (LLMs) and human intelligence without relying on scaling laws such as increasing model size or training data. Specifically, (1) enhancing the generation factuality, diversity, and novelty of LLMs by encouraging the usage of more causal inferences and fewer guesses, (2) discovering LLMs’ limitations in their current architectures, decoding algorithms, evaluation, and training data, and (3) overcoming the limitations through techniques inspired by human behavior.
