[Colloquium] Yangqiaoyu Zhou MS Presentation/Dec 6, 2022

[nitayack at cs.uchicago.edu]

This is an announcement of Yangqiaoyu Zhou's MS Presentation
===============================================
Candidate: Yangqiaoyu Zhou

Date: Tuesday, December 06, 2022

Time: 10 am CST

Remote Location:  https://uchicago.zoom.us/my/meetrosa?pwd=Vi92dkp5eXhKZEdqRGJDK0NWdVFsZz09

Location: JCL 298

M.S. Paper Title: FLamE: Few-Shot Learning from Natural Language Explanations

Abstract: 
Natural language explanations have the potential to provide rich information that in principle guide model reasoning. Yet, recent work by Lampinen et al. has shown limited utility of natural language explanations in improving classification.
     
To effectively learn from explanations, we present FLamE, a two-stage few-shot learning framework
that first generates explanations using GPT-3, and then fine-tunes a smaller model (e.g., RoBERTa) with generated explanations.
      
Our experiments on natural language inference demonstrate effectiveness over strong baselines,
increasing accuracy by 17.6% over GPT-3 Babbage and 5.7% over GPT-3 Davinci in e-SNLI. Despite improving classification performance, human evaluation reveals that the majority of generated explanations does not adequately justify classification decisions. Additional analyses point to the important role of label-specific cues 
(e.g., ``not know'' for the neutral label) in generated explanations.

Advisors: Chenhao Tan

Committee Members: Chenhao Tan, Allyson Ettinger, and Xuezhi Wang