Research talk: System frontiers for dense retrieval

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=Nu2sy_zqOLQ


Duration: 8:36
260 views
0

Speakers:
Jason Li, Partner Group Engineering Manager, Microsoft
Knut Risvik, Distinguished Engineer, Microsoft

The Microsoft Bing search engine combines classic information retrieval and dense retrieval in multiple stages of the search funnel. Handling hundreds of billions of documents with constant updates creates massive system challenges to inference, search, and ranking. In this talk, we’ll present current and ongoing approaches and suggest research challenges to effectively combine dense and sparse retrieval at large scale.

Learn more about the 2021 Microsoft Research Summit: https://Aka.ms/researchsummit



Other Videos By Microsoft Research


2022-01-24Lightning talks: Gaming and Entertainment: Content creation at scale
2022-01-24Research talk: Evaluating human-like navigation in 3D video games
2022-01-24Fireside chat: Opportunities and challenges in human-oriented AI
2022-01-24Plenary: Statistical Imaginaries: An Ode to Responsible Data Science
2022-01-24Fireside chat: Smart network pipes unleashing new opportunities
2022-01-24Closing Remarks: Reinforcement Learning
2022-01-24Keynote: The Future: Converging the Cloud & Telecommunications Infrastructures
2022-01-24Research talk: Capturing the visual evolution of fashion in space and time
2022-01-24Research talk: Towards efficient generalization in continual RL using episodic memory
2022-01-24Closing Remarks: Research for Industry
2022-01-24Research talk: System frontiers for dense retrieval
2022-01-24Research talk: Attentive knowledge-aware graph neural networks for recommendation
2022-01-24Practical tips for productivity & wellbeing: Lessons from COVID-19 around time management
2022-01-24Tutorial, Research talk, and Q&A: ElectionGuard: Enabling voters to verify election integrity
2022-01-24Panel: Ecosystem resiliency for consumer goods
2022-01-24Panel: Monitoring for environmental equity: A hyperlocal air quality network in Chicago
2022-01-24Panel: Challenges and opportunities of causality
2022-01-24Panel: Computer vision in the next decade: Deeper or broader
2022-01-24Panel: Community-driven research: Emerging scholar spotlight
2022-01-24Panel: Perspectives on the new future of hybrid meetings
2022-01-24Panel: Experiments, models, inference and algorithms: Learning from experts who do it all


Tags:
Search & information retrieval
Microsoft Search
Productivity
search
recommendation
future of search
microsoft research summit