Neural sampling from cognitive maps enables goal-directed imagination and planning | Nature Machine Intelligence
Lin et al. introduce a brain-inspired generative model that provides two key features of intelligence: planning and problem-solving. It uses cognitive maps, stochastic computing and compositional coding, and requires only local synaptic plasticity.
Download PDF Subjects Computational science Learning algorithms Network models A preprint version of the article is available at bioRxiv. Abstract Artificial intelligence systems are becoming more intelligent, but at a very high cost in terms of energy consumption and training requirements. By contrast, our brains only require 20 W of energy, they learn online and they can instantly adjust to changing contingencies. This begs the question what data structures, algorithms and learning methods enable brains to achieve that, and whether these can be ported into artificial devices. We are addressi
Explore this link on the map →related reading
- pdfopenreview.net
- LLM Daydreaming · Gwern.netgwern.net
- Continuous Thought Machinespub.sakana.ai
- A GPS for the mind | thesephist.comthesephist.com
- Language Models in Plato's Cave - by Sergey Levinesergeylevine.substack.com
- Building machines that learn and think like people | Behavioral and Brain Sciences | Cambridge Corecambridge.org
- What If We Had Bigger Brains? Imagining Minds beyond Ours-Stephen Wolfram Writingswritings.stephenwolfram.com
- Memory makes computation universal, remember?thinks.lol
- Human-like Neural Nets by Catapulting · Gwern.netgwern.net
- Natural Intelligencegreydanus.github.io
- How to Think | MIT Technology Reviewtechnologyreview.com
- World Models: Computing the Uncomputablenotboring.co