Summary of In-context Learning with Retrieved Demonstrations For Language Models: a Survey, by Man Luo et al.
In-context Learning with Retrieved Demonstrations for Language Models: A Survey
by Man Luo, Xin Xu, Yue Liu, Panupong Pasupat, Mehran Kazemi
First submitted to arxiv on: 21 Jan 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Information Retrieval (cs.IR)
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 explores the concept of few-shot in-context learning (ICL) in language models, specifically pre-trained large language models. These models can adapt to new tasks with just a few demonstrations in the input context, but their performance is sensitive to the choice of these demonstrations. The authors propose retrieving demonstrations tailored to each input query, which improves efficiency and scalability while reducing biases inherent in manual example selection. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper talks about how computers can learn new things quickly by showing them just a little bit of information. This works especially well for large language models that are already good at understanding language. The problem is that the computer’s ability to learn depends on what examples it’s shown. Instead of always using the same examples, the authors suggest finding new examples that fit each specific task or question. This makes the learning process faster and more accurate, and it also helps reduce mistakes that can happen when people choose the examples. |
Keywords
» Artificial intelligence » Few shot
