Software Research Seminar: Talk 1

  • Newell-Simon Hall
  • Mauldin Auditorium 1305
  • Ph.D. Student
  • Ph.D. Program in Software Engineering, Institute for Software Research
  • Carnegie Mellon University

Deep Learning Architecture for Information Type Identification

Identifying semantic roles in privacy statements manually can be a time and effort extensive activity. The traditional methods to identify semantic roles rely heavily on the output of the different syntactic parsers. Previous research showed that errors in syntactic parsing are major sources of errors in the semantic role labeling (SRL) systems. Therefore, more recently the focus has been on SRL techniques based on word embeddings and neural networks which do not use any feature engineering.

In this talk, I will present early results from an exploratory study we conducted to identify information type roles using word embeddings and deep neural networks. I will also briefly describe word embeddings and recurrent neural networks.

For More Information, Please Contact: