Projects

Current Work

• Single Trajectory Estimation of Sparse Linear Systems
• Learning Energy Functions for Contextual Spell-Checking
• Graph Powering for Word Representations

Expository Presentations & Technical Reports

• Neural Machine Translation
  Gene Li, Hrishikesh Khandeparkar.
  A survey of recent developments in NMT (seq2seq, transformer, unsupervised).
  Class Presentation for COS597E, Fall ‘18, Princeton University.
• Compressed Sensing Interpretations of Text Embeddings
  Gene Li.
  An overview of paper.
  Presentation for NLP Group Meeting, Summer ‘18, MIT.
• Identity Matters in Deep Learning
  Gene Li.
  An overview of paper. Longer report found here.
  Class Presentation for ELE538B, Spring ‘18, Princeton University.
• Acceleration via Depth Analysis
  Charlie Hou, Hrishikesh Khandeparkar, Gene Li.
  Some further exploration on paper.
  Class Project for ELE538B, Spring ‘18, Princeton University.
• Political Journalists Network
  Gene Li, Emmanuel Abbe.
  A modern dataset displaying community structure in Twitter connections of American journalists (cf this).
  Code and dataset is upcoming!
  Sophomore Independent Work, ELE298, Spring ‘16, Princeton University