Banner for Large Language Models with Professor Veronika Rockova ( Pre-LPF event )

Large Language Models with Professor Veronika Rockova (Pre-LPF event)

by

Educational

Thu, Feb 15, 2024

5:30 PM – 6 PM CST (GMT-6)

Mindworks

224 S Michigan Ave, Chicago, IL 60604, United States

36
Registered

Registration

Details

Booth Behavioral Sciences and Booth AI Group are teaming up to host an LPF for the community AND a pre-LPF event with Veronika Rockova. 

Come to Mindworks on February 15 from 5:30 to 6:00 to hear about the latest research in the field and network with like-minded students, staff, and faculty from across the Booth community! Full talk description below and you can read more about professor Rockova's work here.

This event is immediately before the AI Group X Behavioral Sciences LPF. We welcome you to stay and mix and mingle at the LPF with others in the Booth community, but you are not required to. Mindworks (this event location) is connected to The Exchange (LPF location), so you do not have to worry about moving between events.

NOTE: There is limited capacity for the pre-LPF flash talk. If you register, please plan on attending, or cancelling your reservation if you know you won’t be able to make it.

---------------------------------
 Who wrote The Federalist papers?  

Attributing authorship of the 12 disputed Federalist papers has been attempted by historians, language theorists as well as statisticians. Some of authors' textual fingerprints have been identified through simple summary statistics  of text (e.g. lengths of sentences, overuse of certain words etc). With the emergence of Large Language Models (LLMs) capable of capturing deep contextual dependencies, we have an opportunity to validate/refute previous authorship predictions by employing more sophisticated representations of text as data. We replicate the original statistical analysis by Mosteller and Wallace and implement more advance text analysis techniques, including Latent Dirichlet Allocation as well as some of the most recent Transformer-based techniques such as BERT. This talk will review and explain the machinery behind LLM's and demonstrate its potential to shed some new light on The Federalist attribution problem.
 
Dress Casual (jeans ok)

Where

Mindworks

224 S Michigan Ave, Chicago, IL 60604, United States

Hosted By

Booth Behavioral Sciences (FT) | Website | View More Events

Rishi Sriram
Co-hosted with: Booth AI Group (FT)

Contact the organizers