IBM Neuro-Symbolic AI Summer School Day 2: Summarization (Pavan Kapanipathi)
IBM NeuroSymbolicAI Summer School 2022 Day 2: Neuro-symbolic Approaches for Summarization and NLG given by Pavan Kapanipathi:
- Hallucinations in Abstractive Summarization: intrinsic and extrinsic.
- Extrinsic hallucinations need external knowledge and reasoning.
- Top types of inference required: expert knowledge, proper nouns, temporal, and numerical.
- Ways to improve factual correctness: NeuroLogic, LinkBERT, Neural Unification.
Ximing Lu, Peter West, Rowan Zellers, Ronan Le Bras, Chandra Bhagavatula, and Yejin Choi. 2021. “NeuroLogic Decoding: (Un)supervised Neural Text Generation with Predicate Logic Constraints.” Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. https://doi.org/10.18653/v1/2021.naacl-main.339.
Michihiro Yasunaga, Jure Leskovec, and Percy Liang. 2022. “LInkBERT: Pretraining Language Models with Document Links.” In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 8003–16. Dublin, Ireland: Association for Computational Linguistics. https://aclanthology.org/2022.acl-long.551/
Gabriele Picco, .Hoang .Thanh Lam, Marco Luca Sbodio, and Vanessa Lopez. 2021. “Neural Unification for Logic Reasoning over Natural Language.” Findings of the Association for Computational Linguistics: EMNLP 2021. https://doi.org/10.18653/v1/2021.findings-emnlp.331.
Summer School site: https://ibm.github.io/neuro-symbolic-ai/events/ns-summerschool2022/
Originally posted on LinkedIn.