Skip to content

pgimenes/aries

Repository files navigation

ARIES: Autonomous Reasoning on Interactive thought graph EnvironmentS

As we approach the end of scaling laws in Large Language Model (LLM) training, test-time compute scaling has emerged as a transformative paradigm for complex reasoning tasks. Test-time compute scaling approaches can be generalized under the framework of topological reasoning, whereby intermediate solutions are arranged as graphs, on which transformations are performed to explore a solution space. However, prior works rely on pre-determined, task-specific transformation schedules which are subject to a hyperparameter set requiring extensive Bayesian search for high query efficiency. By viewing thought graph transformations as actions in a Markov Decision process, policy agents can be equipped to learn from feedback and tune effective action policies. In particular, LLMs can act as policy agents, collaborating with reasoning agents in a multi-agent architecture. While reasoning agents solve decomposed subproblems, LLM policy agents maintain visibility of the reasoning trace, dynamically adaptating the problem-solving strategy. Using off-the-shelf LLMs with no further training as policy agents can yield up to $3.3\times$ lower error values compared to static schedules in problems with low decomposition depth, as well as obviating any search requirement.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

3 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors