M
Minghan Li
Publications - 4
Citations - 8
Minghan Li is an academic researcher. The author has contributed to research in topics: Computer science & Question answering. The author has an hindex of 1, co-authored 4 publications receiving 3 citations.
Papers
More filters
Proceedings Article
Multi-Task Dense Retrieval via Model Uncertainty Fusion for Open-Domain Question Answering
TL;DR: Li et al. as mentioned in this paper proposed to train individual dense passage retrievers for different tasks and aggregate their predictions during test time, where they use uncertainty estimation as weights to indicate how probable a specific query belongs to each expert's expertise.
Proceedings Article
Simple and Effective Unsupervised Redundancy Elimination to Compress Dense Vectors for Passage Retrieval.
TL;DR: Pyserini et al. as discussed by the authors proposed a simple unsupervised compression pipeline that consists of principal component analysis (PCA), product quantization, and hybrid search to improve space efficiency.
Posted Content
Encoder Adaptation of Dense Passage Retrieval for Open-Domain Question Answering
Minghan Li,Jimmy Lin +1 more
TL;DR: The authors showed that the passage encoder has more influence on the lower bound of generalization while the question encoder seems to affect the upper bound in general, and that applying an out-of-distribution encoder usually hurts the retrieval accuracy while an OOD encoder sometimes even improves the accuracy.
Posted Content
Latte-Mix: Measuring Sentence Semantic Similarity with Latent Categorical Mixtures.
TL;DR: This paper theoretically proves that measuring the distance between the latent categorical mixtures, namely Latte-Mix, can better reflect the true sentence semantic similarity, and provides explanations for why models finetuned on labelled sentence pairs have better zero-shot performance.