Uncovering Motivations, Stances and Anomalies through Private-State Recognition
Abstract
We report on natural language processing research on the automatic identification of private states (e.g. opinions, beliefs, evaluations, judgments, feelings) in unstructured text. We focus on the development and evaluation of neural network-based methods for fine-grained opinion analysis methods for inferring private states that are not explicitly mentioned, and methods for determining private-state-aware semantic equivalence.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 2018
- Accession Number
- AD1052275
Entities
People
- Claire Cardi
- Rada Michalcea
Organizations
- Cornell University