Conversational ML at Drift

BostonML is humbled to host a session showcasing some of the brilliant developments taking place at Drift.

Stay strong amidst the homestretch of the pandemic. Almost there!

Schedule:

6:45: Mingle in breakout rooms https://us02web.zoom.us/j/84009567552
7:00: Lights Camera Action
https://us02web.zoom.us/j/89849965270

Abstract:

In this talk, we describe a conversational AI system that diverges from established intent-based architures to achieve more human-like interactions. Authored bottom-up from data, Drift’s AI bots use an event-based dialogue manager to engage in coherent multi-turn conversations with flexible context switching. We will give an overview of how we select responses to natural language input, and drill down into the details of the ML models we use for NLU, and our human-in-the-loop process for continuous learning.

Speaker Bios:

Jeff Orkin is the Head of the Drift AI Lab (DAIL). Previously, Jeff founded Giant Otter Technologies, a conversational AI startup based on his Ph.D research in the Cognitive Machines group at the MIT Media Lab. Jeff spent the first decade of his career developing AI for video games, including FEAR and No One Lives Forever 2.

Chris Ward is a cat-loving machine learning enthusiast. He has a Masters in Computational Linguistics from Brandeis and has been working on conversational AI for 5 years.

Luke Van Seters is Tech Lead of the AI Bots team focused on scalability and reliability of the runtime along with improving the human-in-the-loop improvement process. Prior to AI, he helped develop Drift’s core chat product, integrations, and decision-tree bot.