Summary of Quito: Accelerating Long-context Reasoning Through Query-guided Context Compression, by Wenshan Wang et al.
QUITO: Accelerating Long-Context Reasoning through Query-Guided Context Compression
by Wenshan Wang, Yihang Wang, Yixing Fan, Huaming Liao, Jiafeng Guo
First submitted to arxiv on: 1 Aug 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 This paper introduces a novel method called Query-gUIded aTtention cOmpression (QUITO) that reduces the reasoning complexities and computation costs of large language models (LLMs). QUITO leverages attention over contexts to filter out useless information, utilizing a trigger token to calculate attention distributions. The authors propose three filtering methods to satisfy budget constraints for context length. Experimental results on NaturalQuestions and ASQA datasets demonstrate QUITO’s effectiveness, outperforming established baselines across various datasets and downstream LLMs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps big language models learn better by cutting down the unnecessary information they process. It introduces a new way to do this called QUITO, which looks at questions and contexts together to figure out what’s important. The authors tested QUITO on two big datasets and found that it works really well, beating existing methods in many cases. |
Keywords
» Artificial intelligence » Attention » Context length » Token
