Abstract
People use search engines to look up information on the Internet, using search queries related to their information needs. This disclosure describes the use of machine learning techniques, including supervised learning and reinforcement learning to train a search agent to search deeper for better, more accurate, better supported answers by interacting with the search engine. The interaction mimics strategies utilized by human experts to carry out accurate web search. The search agent can be modular, and to provide answers to a user query, performs operations such as formulation of new queries in a sequence, analysis of intermediate results, and selection of results based on a chosen success metric that can take into account factors such as accuracy, diversity, presence of justification, etc.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Ciaramita, Massimiliano; Huebscher, Michelle Chen; Boerschinger, Benjamin; and Kilcher, Yannic, "Machine-learning Based Automatic Formulation of Query Sequences to Improve Search", Technical Disclosure Commons, (June 01, 2021)
https://www.tdcommons.org/dpubs_series/4340