“Big Enough” Language Models: Domain-Specific Approaches to Speech AI

Presented by

Jason Rowley, Editor in Chief, Deepgram; Dan Miller, Lead Analyst, Opus Research; Derek Top, Research Director, Opus Research

About this talk

In the recent "The Inside Story: Factors Propelling Speech AI in 2023” webinar, we delved into the current state of speech AI, leveraging the insights gleaned from surveying 400 business leaders on the State of Voice Technology in 2023. In part two of this series, we will dive deeper into the world of language models, from their fundamental concepts to the latest advancements in the field. We will also shed light on the promises and perils of large language models, as well as discuss how domain-specific models can offer a tailored approach to address various challenges. In this webinar, you will learn: * Language model architecture and training paradigm  * Compromises between using LLMs versus DSLMs * Business impacts of DSLMs and real-world examples showcasing their benefits * What’s on the horizon for LLMs and DSLMs
Related topics:

More from this channel

Upcoming talks (0)
On-demand talks (56)
Subscribers (12043)
News, information, and research detailing Intelligent Assistants, Intelligent Authentication Conversational Commerce. Webcasts include in-depth analysis of market forecasts, real-world case studies, and proven business value.