Summary of Sociolinguistically Informed Interpretability: a Case Study on Hinglish Emotion Classification, by Kushal Tatariya et al.
Sociolinguistically Informed Interpretability: A Case Study on Hinglish Emotion Classification
by Kushal Tatariya, Heather Lent, Johannes Bjerva, Miryam de Lhoneux
First submitted to arxiv on: 5 Feb 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The abstract proposes a study on emotion classification in NLP, focusing on whether pre-trained language models (PLMs) can learn associations between language choice and emotional expression across languages. The researchers examine the performance of three PLMs on a Hinglish emotion classification dataset, utilizing techniques like LIME and token-level language ID to analyze the models’ predictions. The results suggest that PLMs do learn these associations and that incorporating code-mixed data during pre-training can enhance task-specific learning when data is scarce. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Emotion classification in different languages can be tricky because emotions are expressed differently depending on the language used. Researchers studied whether machines, like language models, can pick up on this difference. They looked at three special machine learning models and how they performed on a dataset of Hindi-English code-mixed text that expresses emotions. The results show that these models can learn to recognize patterns between languages and emotional expressions. This is important because it means that if we have limited data in one language, but more data in another, the model might be able to use this extra information to make better predictions. |
Keywords
* Artificial intelligence * Classification * Machine learning * Nlp * Token
