AI Powered Smart Search Assistant
The students on the AI ProQuest team will deliver an end-to-end system/engine that can reliably suggest and assist users in forming “better” search queries of medical databases like Medline (PubMed) resulting in quick, precise search results.
ProQuest is a content aggregator and research and learning hub for students, librarians, instructors, and researchers. This project will focus on creating a Smart Search Assistant that leverages the power of AI to help users search better. The team will use publicly available and ProQuest-proprietary datasets to help achieve this goal. An example of a feature could be smart suggestions:
If a user tries to search for “heart attack” in a medical database, can the AI leverage context to suggest the word “myocardial infarction” (preferred term for heart attack)? Now think about a case like “blockage of blood flow to the heart”.
Some exciting challenges in this project include the following:
1. Tailoring suggestions based on user, database and other contextual features.
2. Leveraging high-quality taxonomies to assist in suggestive search queries.
3. Curating custom datasets for various features like auto-complete.
4. Addressing scalability concerns for large-scale deployment.
Approaches may include (but will not be limited to):
- Language Models
- Attention Mechanisms
- Recurrent Neural Networks
- Conditional Random Fields (CRF)
- Long-Short Term Memory (LSTM)
- Combinatory Categorical Grammars (CCG)
- Ensemble Techniques
The student team will deliver an end-to-end system/engine that can reliably suggest and assist users in forming “better” search queries.