Redefining domain adaptation
A perspective on domain adaptation for machine translation and automated speech recognition
The webinar was organised by Omniscien.
This webinar will demonstrate how Omniscien’s team gathers and synthesizes billions of sentences of training data that is used to teach and adapt AI systems. We will explore best practices and explain the limitations of traditional approaches. We will show how to use AI to create many millions of sentences of training data that can be used as part of domain adaptation. Essentially, this means using one form of AI to teach another AI form new knowledge.
Location
- Online
Links
Speakers
- Professor Philipp Koehn, Chief Scientist at Omniscien and Professor at Johns Hopkins University
- Dion Wiggins, Chief Technology Officer at Omniscien