AdaTest

Subscribers:
343,000
Published on ● Video Link: https://www.youtube.com/watch?v=9xQd8On3u3Q



Duration: 2:20
1,293 views
42


AdaTest is an open-source tool for finding and fixing bugs in natural language processing (NLP) systems. In this demo, learn how AdaTest capitalizes on large language models’ ability to generate tests and users’ knowledge of how a model being tested should behave. The result is an iterative testing-and-debugging loop that helps identify and address model failures more effectively than current user-driven or automated approaches.

Learn more about AdaTest:

Microsoft Research blog post: https://www.microsoft.com/en-us/research/blog/partnering-people-with-large-language-models-to-find-and-fix-bugs-in-nlp-systems/
Paper: https://www.microsoft.com/en-us/research/publication/adaptive-testing-and-debugging-of-nlp-models/




Other Videos By Microsoft Research


2022-06-30[VLP Tutorial @ CVPR 2022] Image-Text Pre-training Part I
2022-06-27MSR-IISc AI Seminar Series: GFlowNets and System 2 Deep Learning - Yoshua Bengio
2022-06-24Safer Roads, Safer India
2022-06-16MIT Technology Review’s Future Compute Conference 2022
2022-06-13Women in Data Science Community Panel Event: Women in Experimentation
2022-06-01Supercharge A/B testing w/automated causal inference |Community Workshop on Microsoft's Causal Tools
2022-06-01Update on Microsoft causal open-source libraries | Community Workshop on Microsoft's Causal Tools
2022-06-01Fireside chat with Stu Frost at Community Workshop on Microsoft's Causal Tools
2022-05-25AI Advanced Sustainability Research
2022-05-23MSR-IISc AI Seminar Series: Where on Earth is AI Headed? - Tom M. Mitchell
2022-05-23AdaTest
2022-05-19Materials research from the perspective of computing: data, model, and more
2022-05-03Microsoft New Future of Work Report 2022: AI-powered social platforms
2022-05-03Microsoft New Future of Work Report 2022: Team Collaboration
2022-05-03Microsoft New Future of Work Report 2022: Individual productivity and well-being
2022-05-03Microsoft New Future of Work Report 2022: Asynchronous collaboration
2022-05-03Microsoft New Future of Work Report 2022: Hybrid workspaces
2022-05-03Microsoft New Future of Work Report 2022: Organizational wellbeing
2022-05-02Designing an AI-driven Neighborhood Navigator with Black and Latinx NYC Residents
2022-04-25Overview of AI Research at the VUB Artificial Intelligence Lab
2022-04-25Deep Medicine: Leveraging AI for health in a relevant and impactful way in Africa