How Attention in 2017 Got Us ChatGPT

Details:

Streaming on DC2 YouTube here: https://youtube.com/live/zmt5kiOth_I?feature=share

This event will be in person. We will try to add an online streaming option if the space and hardware will accommodate. Check back here at 6:45pm ET on the evening of the event and if there's a link here, then that's an option. If there's not, then it's not. Same goes for recording. :)

6:00-6:30 Networking and refreshments - empanadas are back! We're ordering from Banditos Tacos & Tequila.

6:30-6:50 DC Data Science and DC Data Community Announcements

6:50-7:00 Opening Remarks & Introduction by Jennifer Ives, CEO/Cofounder, Watering Hole AI

7:00-8:30 Dr. John Kaufhold's talk "How Attention in 2017 got us ChatGPT"

8:30 Data Drinks and Networking at The Liberty Tavern located at 3195 Wilson Blvd (walking distance from Mason) and in the direction of the Clarendon metro.

This event is co-hosted by George Mason University and Arlington Economic Development.

Description:

The 2017 “Attention is All You Need” paper introduced the transformer architecture. In the past six years, interest in the transformer architecture has only grown, and now the capabilities of GPT4 are the subject of frenzied media attention in the popular press as well as research groups at OpenAI, Microsoft, HuggingFace, Argilla, Google, Facebook, and others. The emergence of the transformer architecture heralded the current era of large language models (LLMs). But how do they work? Should you be intimidated by their inner workings? Are they magic? Are they really artificial general intelligence (AGI)? What does each component do?

This talk will briefly walk through each component visually and make them understandable in historical perspective no matter your level of expertise in machine learning (ML). Dr. Kaufhold will make complex Transformer concepts accessible to a broad audience while also incorporating the more relevant themes of deep learning of interest to more advanced ML engineers — the themes that led to its dominance. This talk will serve as an intro for more talks on LLMs in the coming months.

Speaker Bio:

Dr. John Kaufhold is the founder of machine learning startup, Deep Learning Analytics, the fastest growing company by revenue three years in a row in Arlington, VA. He is a fellow of the Washington Academy of Sciences, an advisor and board member to Data Community DC, and advisor to ML startups PatentBots and RedShred. He holds the record for most attended Data Community DC event at his 2017 talk on deep learning. He began his machine learning career in speech recognition in 1993. He is an author of >40 peer-reviewed publications and named inventor on >20 patents. According to Google Scholar, Dr. Kaufhold's H-index is 25, with over 2,800 citations of his work.

Location:

Van Metre Hall Auditorium (134) at George Mason Square campus (Arlington, VA) is between the Clarendon and Virginia Square Metro stops. The street address is 3351 Fairfax Drive.