Hi, I'm Aidan!

Contact:

Email: aidan.w.san@gmail.com

Research:

Irony Detection Publication (Accepted Paper SemEval 2018)
  • 2 Stacked LSTMs (with dropout) for Sarcasm detection (Keras)
  • Uses Emoji Embeddings and Sentiment Scores
  • ACL Anthology
  • GitHub

Concept Modeling (CS 546)
  • Project to figure out concepts (ie. Addition, Arithmetic, Calculus) from MOOC courses and map their relationships
  • Concept Chunker - Uses IOB tagging to determine where multiword concept phrases start and end
  • 1 LSTM Hidden Layer (PyTorch)
  • Results on MOOC Lecture Dataset (Test): Accuracy: .961, Macro F1: .883
  • GitHub


Positions:

Facebook SWE Intern
  • Language and Translation Technology Team (Applied Machine Learning Org.)
  • Used Python to improve translation data quality
  • Attended weekly translation reading group

Chair of Special Interest Group for Natural Language Learning
  • Develop new course material about NLP topics
  • Logistic Regression, ANNs, Word Embeddings, Dependency Parsing

Course Assistant for Datastructures (CS225)