Summary of Iitk at Semeval-2024 Task 1: Contrastive Learning and Autoencoders For Semantic Textual Relatedness in Multilingual Texts, by Udvas Basak and Rajarshi Dutta and Shivam Pandey and Ashutosh Modi
IITK at SemEval-2024 Task 1: Contrastive Learning and Autoencoders for Semantic Textual Relatedness in Multilingual Texts
by Udvas Basak, Rajarshi Dutta, Shivam Pandey, Ashutosh Modi
First submitted to arxiv on: 6 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); 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 paper presents a system designed to tackle SemEval-2024 Task 1: Semantic Textual Relatedness, which involves detecting the degree of relatedness between sentence pairs for 14 languages. The authors participated in two subtasks: Track A (supervised) and Track B (unsupervised), focusing on contrastive learning with BERT and a similarity metric approach for the supervised track, while exploring autoencoders for the unsupervised track. Additionally, they created a bigram relatedness corpus using negative sampling, resulting in refined word embeddings. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper is about a new system that helps computers understand how similar two sentences are. It’s trying to solve this problem for 14 different languages. The team worked on two parts: one where they had help from labeled data (Track A) and another where they didn’t have any labels (Track B). They used special AI models called BERT and autoencoders to make predictions. They also created a big dataset of sentence pairs with their relatedness levels, which will help improve language understanding. |
Keywords
* Artificial intelligence * Bert * Language understanding * Supervised * Unsupervised
