Intelligence Foundation Model: A New Perspective to Approach Artificial General Intelligence

TL;DR

This paper rethinks AGI by shifting from task-centric foundation models to an intelligence foundation model that learns underlying neural dynamics from diverse intelligent behaviors. It introduces two core components: a graph-based State Neural Network that captures neuron-like dynamics (neuron function, connectivity, and plasticity) and a Neuron Output Prediction objective that trains the model to predict neuronal outputs from temporal input dynamics. Training IFM relies on large-scale neuronal input–output data, collected via direct or indirect neuronal sampling, and is trainable with backpropagation and TBPTT, with a toy Pong demonstration illustrating the approach. The paper argues that grounding AI in biological-inspired dynamical principles offers a scalable path to generalization, while outlining substantial challenges and a roadmap combining biological and functional scaling.

Abstract

We propose a new perspective for approaching artificial general intelligence (AGI) through an intelligence foundation model (IFM). Unlike existing foundation models (FMs), which specialize in pattern learning within specific domains such as language, vision, or time series, IFM aims to acquire the underlying mechanisms of intelligence by learning directly from diverse intelligent behaviors. Vision, language, and other cognitive abilities are manifestations of intelligent behavior; learning from this broad range of behaviors enables the system to internalize the general principles of intelligence. Based on the fact that intelligent behaviors emerge from the collective dynamics of biological neural systems, IFM consists of two core components: a novel network architecture, termed the state neural network, which captures neuron-like dynamic processes, and a new learning objective, neuron output prediction, which trains the system to predict neuronal outputs from collective dynamics. The state neural network emulates the temporal dynamics of biological neurons, allowing the system to store, integrate, and process information over time, while the neuron output prediction objective provides a unified computational principle for learning these structural dynamics from intelligent behaviors. Together, these innovations establish a biologically grounded and computationally scalable foundation for building systems capable of generalization, reasoning, and adaptive learning across domains, representing a step toward truly AGI.

Figures (4)

• Figure 1: Comparison between IFM and existing FMs.
• Figure 2: Conceptual overview of IFM.
• Figure 3: State neural network.
• Figure 4: Pavlov’s classical conditioning experiment.