Dec. 16, 2022
Alaa Alslaity & Rita Orji
1. Current State
1) Increasing research interest
2) Dominance of supervised learning techniques (e.g. SVM, Naive Bayes, etc.)
3) Focus on English text data
4) Accuracy as primary evaluation metric
2. Challenges
1) Understanding context: Interpreting figurative language, metaphors, subjective expressions, etc.
2) Handling multiple emotional expressions
3) Resolving web slang and ambiguity: slang, abbreviations, typos, etc.
3. Future Directions
1) Developing emotion detection for human-centered system design
(1) There is a need for human-centered systems that understand and reflect user emotions
2) Researching techniques applicable to diverse domains and languages
(1) Most research has focused on English text data -> There is a need for generalized techniques applicable across different languages and domains.
3) Exploring new approaches for context understanding and ambiguity resolution
(1) Potential directions include knowledge-based methods, multimodal techniques, transfer learning, etc.
