Evolution of Large Language Models: The Good, The Bad and The OpportunitiesSpeaker: Tanmoy Chakraborty – New Delhi, India
Topic(s): Artificial Intelligence, Machine Learning, Computer Vision, Natural language processing
AbstractIn recent years, the field of natural language processing has seen a revolution with the development of large language models (LLMs) such as BERT, GPT-3, T5, and others. These models have been shown to be highly effective in a wide range of natural language processing tasks, from machine translation to chatbot conversations. However, the rise of large language models also raises concerns about their potential negative impacts. For example, some experts worry that these models could reinforce biases and stereotypes found in the data they are trained on, leading to unintended discrimination. In this talk, I will explore the evolution of large language models, including both the good and the bad. I will discuss their benefits, limitations, and potential ethical implications. I will also explore the opportunities that these models present for advancing natural language processing and expanding our understanding of language and communication. I will then present a series of our recent studies on various aspects of LLMs – in-context learning, cross-modal prompting, layer editing, effects of model hallucination, Transformer as a dynamical system, etc. Ultimately, the talk aims to provide a balanced perspective on the evolution of large language models and to spark a conversation about the opportunities and challenges presented by this rapidly evolving field.
About this LectureNumber of Slides: 50
Duration: 45 minutes
Languages Available: English
Request this Lecture
To request this particular lecture, please complete this online form.
Request a Tour
To request a tour with this speaker, please complete this online form.
All requests will be sent to ACM headquarters for review.