Improving Conversational Interactions through Visually Situated Context
Center for Language & Speech Processing(CLSP), JHU via YouTube
Overview
Explore the advancements in conversational systems and the potential of visually situated context in a one-hour lecture by Dr. Larry Heck from Georgia Tech. Delve into the impact of Large Language Models (LLMs) like GPT-4, PaLM, and LLaMA on open-domain conversational interactions. Examine strategies for improving accuracy beyond pre-trained LLMs, focusing on context augmentation methods. Learn about leveraging various contextual elements, including personal, dialogue, geo, and time contexts, with special emphasis on the power of shared visual settings in grounding natural language conversations. Discover Dr. Heck's historical and recent work in this field, conducted at the AI Virtual Assistant (AVA) Lab at Georgia Tech. Gain insights from Dr. Heck's extensive experience in speech recognition, natural language processing, and AI, spanning roles at prestigious institutions and tech giants like SRI, Nuance, Yahoo!, Microsoft, Google, and Samsung.
Syllabus
Improving Conversational Interactions through Visually Situated Context - Larry Heck (GeorgiaTech)
Taught by
Center for Language & Speech Processing(CLSP), JHU