

Four Key Tools for Robust Enterprise NLP with Yunyao Li - #537
Nov 18, 2021
In a lively discussion, Yunyao Li, a Senior Research Manager at IBM Research, tackles the intricacies of natural language processing in enterprise settings. She shares insights on productizing NLP, balancing customer needs with research rigor. Yunyao dives into the complexities of document discovery and the synergy of deep learning techniques. Highlighting the importance of human involvement, she discusses innovative data augmentation strategies to create high-quality datasets. Her unique perspective reveals how IBM empowers users to enhance AI transparency and accuracy.
AI Snips
Chapters
Transcript
Episode notes
System T for Enterprise Search
- Yunyao Li worked on System T, a declarative system for NLP, to improve enterprise search.
- The system aimed to understand each document deeply, unlike internet search which prioritizes quantity.
IBM's Focus on Empowering Others
- IBM Research focuses on building tools that empower others, unlike Google's model/idea-sharing approach.
- This focus stems from IBM's enterprise-facing nature, emphasizing practical application over theoretical research.
Four Key Challenges in Enterprise NLP
- Enterprise NLP faces four key challenges: complexity, small data, customization, and explainability.
- These challenges arise from the nature of enterprise data and the need for tailored solutions.