Generative Recursive Reasoning
The paper introduces Generative Recursive Reasoning Models (GRAM), a new framework for neural reasoning systems. GRAM enhances traditional Recursive Reasoning Models by enabling probabilistic multi-trajectory computation, allowing for multiple hypotheses and alternative solutions. This approach improves performance on structured reasoning tasks and supports both conditional and unconditional generation capabilities.
- ▪GRAM models reasoning as a stochastic latent trajectory.
- ▪It allows for multiple hypotheses and inference-time scaling.
- ▪The framework shows improved performance over deterministic models on various reasoning tasks.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.19376 (cs) [Submitted on 19 May 2026] Title:Generative Recursive Reasoning Authors:Junyeob Baek, Mingyu Jo, Minsu Kim, Mengye Ren, Yoshua Bengio, Sungjin Ahn View a PDF of the paper titled Generative Recursive Reasoning, by Junyeob Baek and 5 other authors View PDF HTML (experimental) Abstract:How should future neural reasoning systems implement extended computation? Recursive Reasoning Models (RRMs) offer a promising alternative to autoregressive sequence extension by performing iterative latent-state refinement with shared transition functions. Yet existing RRMs are largely deterministic, following a single latent trajectory and converging to a single prediction.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
