About the talk
Empowering citizen developers in NLP & democratizing it for all makers in developer-friendly workflows
In this talk, Mady Mantha will demonstrate two cutting edge machine learning products to deliver far more accurate and performant natural language applications than some of the biggest language models today. These products take what’s proven to work in the latest research and apply it to production to enable all makers to use it reliably.
AI assistants that can understand context and handle complex tasks are transforming business, from internal operations, to banking, to healthcare. But, building enterprise-grade AI assistants requires tools and techniques that go beyond basic FAQ bots and are up to the challenge. In addition, making the latest NLP advances available in developer friendly workflows to democratize conversational AI has been instrumental in cutting cost and increasing revenue. Mady will talk about the two new products: A new neural network architecture called Dual Intent and Entity Transformer (DIET) outperforms the popular language model BERT, is 6X faster to train and advances the current state of the art in intent and entity classification. A new dialogue architecture (Transformer Embedding Policy- TED) that handles complex, multi-turn interactions with fewer training examples and data to make truly contextual assistants a reality. Mady will show how both DIET and TED are available in developer friendly workflows that free up valuable developer and product teams’ time and empower all makers to meaningfully contribute to conversational AI product development and drive measurable value.
Buy this talk
Buy this video
With ConferenceCast.tv, you get access to our library of the world's best conference talks.