Reproduce “A practical introduction to the Rational Speech Act modeling framework” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Gregory Scontras, Michael Henry Tessler, Michael Franke
The Rational Speech Act (RSA) framework offers a computational approach to understanding pragmatic reasoning, formalizing how speakers and listeners infer meaning through recursive social cognition. It models diverse linguistic phenomena by integrating formal semantics with Bayesian inference and game theory, yielding quantitative predictions for human behavior.
Key insights
- Language understanding functions as a dynamic process of agents reasoning about each other's intentions and rationality, rather than merely a vacuum-sealed formal system.
- Probabilistic programming languages are crucial enablers, allowing researchers to define and implement complex stochastic generative processes that facilitate a 'thinking in code' approach.
- The framework effectively formalizes Gricean Maxims of Conversation, showing how principles like informativity and cost drive pragmatic inferences.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.