Summary of Toward In-context Teaching: Adapting Examples to Students’ Misconceptions, by Alexis Ross and Jacob Andreas
Toward In-Context Teaching: Adapting Examples to Students’ Misconceptions
by Alexis Ross, Jacob Andreas
First submitted to arxiv on: 7 May 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 introduces AdapT, a suite of computational models and evaluation methods to study how large language models can adapt as teachers to students of different types. The AdapT framework consists of simulated Bayesian student models for evaluating automated teaching methods, a platform for human student evaluations, and AToM (Adaptive Teaching model), which jointly infers past beliefs and optimizes future beliefs. In simulations across three learning domains, AToM outperforms LLM-based and standard Bayesian teaching models. Human experiments also show that both AToM and LLMs outperform non-adaptive random example selection. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper shows how computers can be used to teach students better by adapting to their needs. It creates a system called AdapT, which includes special student models and ways to test teaching methods with real people. The system also has a new way of teaching that does a good job of figuring out what students know and helping them learn more. In tests with simulated students, this approach did better than other methods. When tested with real students, it also worked well. |
