Summary of Kocosa: Korean Context-aware Sarcasm Detection Dataset, by Yumin Kim et al.
KoCoSa: Korean Context-aware Sarcasm Detection Dataset
by Yumin Kim, Heejae Suh, Mingi Kim, Dongyeon Won, Hwanhee Lee
First submitted to arxiv on: 22 Feb 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
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 paper introduces a new dataset, KoCoSa, for detecting sarcasm in Korean dialogue. The dataset consists of 12.8K daily dialogues with labels for the task. To generate the dataset, the authors propose an efficient pipeline that involves generating sarcastic dialogues from source dialogues using large language models, filtering out abnormal and toxic dialogues, and human annotation. A simple baseline model is also provided, which outperforms strong baselines like GPT-3.5 in the Korean sarcasm detection task. The results show that the task relies heavily on sufficient context. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper makes a new dataset for detecting sarcasm in Korean conversations. This dataset has lots of examples of how people talk sarcastically every day, and it also has labels to help computers understand what’s sarcastic. To make this dataset, the authors came up with an easy way to generate more examples from existing ones using big language models, remove bad or mean dialogues, and have humans check if something is sarcastic or not. They also made a simple computer program that can detect sarcasm better than some other strong programs. |
Keywords
» Artificial intelligence » Gpt
