← Back
Computation
Index: #54

AI is Still Narrow in its Reasoning and Planning

Current AI systems exhibit narrow reasoning and planning capabilities compared to human cognition. Broadening AI training methods to include holistic, brain-inspired architectures and cognitive frameworks can advance general intelligence (flagging that there is an AI safety risk here).

Foundational Capabilities (5)

Bayesian Cognitive Architectures  and Program Synthesis

Implement Bayesian cognitive models and probabilistic programming and program synthesis techniques to endow AI with more human-like reasoning, planning, and decision-making capabilities.
  • Ndea
    Company
  • ARC Prize
    Funding Program
  • Bayes3D: fast learning and inference in structured generative models of 3D objects and scenes
    Research and Reviews
  • DreamCoder: growing generalizable, interpretable knowledge with wake–sleep Bayesian program learning
    Research and Reviews
  • From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thought
    Research and Reviews
  • Generative Flow Networks
    Research and Reviews
  • Basis
    Research Org

    • Mimic Human Evolution in Silico

    Use evolutionary algorithms and generative models to simulate human evolution processes in computational systems, driving more robust and adaptable AI.
  • Autocurricula and the Emergence of Innovation from Social Interaction: A Manifesto for Multi-Agent Intelligence Research
    Research and Reviews
  • Open-Endedness is Essential for Artificial Superhuman Intelligence
    Research and Reviews
  • A social path to human-like artificial intelligence
    Technology Seed

    • Developmentally Realistic AI Training

    Train AI models on data similar to those developing humans or animals actually experience
  • Digital Twin Studies for Reverse Engineering the Origins of Visual Intelligence
    Research and Reviews
  • Self-supervised learning of video representations from a child's perspective
    Research and Reviews

    • Holistic Brain-Inspired Architectures

    Develop AI architectures that are inspired by the human brain's structure and functionality, enabling more flexible and general reasoning and planning.
  • E11 Bio: Scalable brain circuit mapping for transformative neuroscience
    FRO
  • LeCun’s “A Path Towards Autonomous Machine Intelligence” has an unsolved technical alignment problem, by Steve Byrnes
    Whitepapers and Essays

    • Improve Out of Context Reasoning in Large Language Models

    Improve the ability of large language models or systems based on them to synthesize disparate information outside their context windows to formulate new scientific hypotheses
  • Dwarkesh Patel on Creative Connectivity in AI
    Whitepapers and Essays