Summary of Adapting to Teammates in a Cooperative Language Game, by Christopher Archibald and Spencer Brosnahan
Adapting to Teammates in a Cooperative Language Game
by Christopher Archibald, Spencer Brosnahan
First submitted to arxiv on: 26 Feb 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL)
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 novel approach to designing intelligent agents for playing Codenames, a game that requires language and coordination between teammates. The existing methods use a single internal language model, which often leads to good performance with some teammates but poor performance with others. To address this limitation, the authors propose an ensemble agent that adapts to individual teammates by selecting the best internal expert on each turn. This approach is evaluated using a novel single metric that captures the performance of a Codenames team in both solitaire and competitive games. The experimental results show that the ensemble agent outperforms the best internal expert with a teammate, without requiring any prior knowledge about the teammates or their compatibility. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates an intelligent agent to play the game Codenames. The agents usually use one language model to make choices, but this can lead to good performance with some teammates and bad performance with others. To fix this, the authors make a new kind of agent that chooses which internal expert to use based on the teammate they’re working with. This helps the agent do well with different teammates without needing prior knowledge about them. The paper also introduces a new way to measure how good an agent is at playing Codenames. |
Keywords
» Artificial intelligence » Language model
