Papers by Andrew Mao
Cheater’s Bowl: Human vs. Computer Search Strategies for Open-Domain QA (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Open-domain and multi-hop QA is an important problem for both humans and computers. |
| Approach: | They propose a gamified interface where a human answers complex questions with access to traditional and modern search tools. |
| Outcome: | The proposed interface compares human queries to state-of-the-art QA models . human queries can improve the accuracy of existing systems, the authors argue . |
Improving the TENOR of Labeling: Re-evaluating Topic Models for Content Analysis (2024.eacl-long)
Copied to clipboard
Zongxia Li, Andrew Mao, Daniel Stephens, Pranav Goel, Emily Walpole, Alden Dima, Juan Fung, Jordan Boyd-Graber
| Challenge: | Existing evaluation metrics such as coherence and coherency are inadequate for neural topic models. |
| Approach: | They conduct the first evaluation of neural, supervised and classical topic models in an interactive task-based setting. |
| Outcome: | The proposed model performs better on cluster evaluation metrics and human evaluations than classical models on real-world tasks. |