{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/ranking-and-selecting-multi-hop-knowledge","title":"Ranking and Selecting Multi-Hop Knowledge Paths to Better Predict Human Needs","arxiv_id":"1904.00676","date":"2019-04-01","proceeding":"NAACL 2019 6","authors":["Debjit Paul","Anette Frank"],"abstract":"To make machines better understand sentiments, research needs to move from\npolarity identification to understanding the reasons that underlie the\nexpression of sentiment. Categorizing the goals or needs of humans is one way\nto explain the expression of sentiment in text. Humans are good at\nunderstanding situations described in natural language and can easily connect\nthem to the character's psychological needs using commonsense knowledge. We\npresent a novel method to extract, rank, filter and select multi-hop relation\npaths from a commonsense knowledge resource to interpret the expression of\nsentiment in terms of their underlying human needs. We efficiently integrate\nthe acquired knowledge paths in a neural model that interfaces context\nrepresentations with knowledge using a gated attention mechanism. We assess the\nmodel's performance on a recently published dataset for categorizing human\nneeds. Selectively integrating knowledge paths boosts performance and\nestablishes a new state-of-the-art. Our model offers interpretability through\nthe learned attention map over commonsense knowledge paths. Human evaluation\nhighlights the relevance of the encoded knowledge.","url_abs":"http://arxiv.org/abs/1904.00676v1","url_pdf":"http://arxiv.org/pdf/1904.00676v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"ranking-and-selecting-multi-hop-knowledge","repo_url":"https://github.com/debjitpaul/Multi-Hop-Knowledge-Paths-Human-Needs","is_official":0,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"common-sense-reasoning","task_name":"Common Sense Reasoning"}],"methods":[{"method_slug":"interpretability","method_name":"Interpretability"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1904.00676","atlas_url":"https://app.syntology.ai/?focus=1904.00676","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}