Investigating associative, switchable and negatable Winograd items on renewed French data sets
Download on ACL Anthology
Dataset
About me
Xiaoou Wang, PhD student in Wimmics (Inria Sophia Antipolis), I3S Laboratory (CNRS), founder of the website AIPrototypes, follow me on Google Scholar and Medium :)
NLP researcher particularly interested in:
1. Computational argumentation: namely Argument Mining and Argument Generation, including also Counter-argument Generation and their applications to the fight against online dis/misinformation
2. Advanced error type classification via Grammatical Error Correction (GEC, especially transformer-based Neural machine translation (NMT) approaches)
3. Advanced evaluation of students' written production
- [nlp/ml] Automated Essay Scoring (AES based on error analysis)
- [nlp/ml/dl] Similarity-based evaluation with Gold Standard
- [dl] Identification of argumentative elements without Gold Standard
5. In a broad sense, promoting/democratizing trustworthy, interpretable and socially beneficial Artificiel Intelligence
- Semi-automatic analysis of biomedical texts, popular science article written in French for DemoTal (project supported by the French Ministry of Culture).
- Better understand spell checkers to make informed decisions, idem.
- Argue instead of blocking: moderating online comments while promoting debate, idem.
- Beyond the simple positive-negative dichotomy, discover the aspect-based sentiment analysis (ABSA), idem.
- Extractive question answering from customer reviews, idem.
- Conducting POCs with limited data? From data quality to zero-shot learning, idem.