Papers by Jany Chan
Learning Latent Structures for Cross Action Phrase Relations in Wet Lab Protocols (2021.acl-long)
Copied to clipboard
| Challenge: | Lack of careful documentation has led to reproducibility crisis in biosciences . wet laboratory protocols are written in natural language and are unclear, composed of disconnected and distant parts and built upon implicit information that were referenced earlier or omitted entirely. |
| Approach: | They propose a model that extracts all action relationships across multiple sentences to generate a MSTG for a given protocol . they propose extending annotations in the WLP corpus for inter-sentence relations and implicit arguments to improve the model's temporal and causal relations. |
| Outcome: | The proposed model achieves an F1 score of 54.53% for temporal and causal relations in protocols from the corpus WLP-MSTG. |