Towards Machine Intelligence
arXiv:1603.08262
Abstract
There exists a theory of a single general-purpose learning algorithm which could explain the principles of its operation. This theory assumes that the brain has some initial rough architecture, a small library of simple innate circuits which are prewired at birth and proposes that all significant mental algorithms can be learned. Given current understanding and observations, this paper reviews and lists the ingredients of such an algorithm from both architectural and functional perspectives.
arXiv comment10 pages, submitted to AGI-16. arXiv admin note: substantial text overlap with arXiv:1512.01926
References in corpus (8)
- Deep Learning in Neural Networks: An Overview
- Going Deeper with Convolutions
- Modeling Temporal Dependencies in High-Dimensional Sequences: Application to Polyphonic Music Generation and Transcription
- Why Neurons Have Thousands of Synapses, A Theory of Sequence Memory in Neocortex
- Deep Learning of Representations: Looking Forward
- Properties of Sparse Distributed Representations and their Application to Hierarchical Temporal Memory
- Learning Representations by Maximizing Compression
- Simple Algorithmic Principles of Discovery, Subjective Beauty, Selective Attention, Curiosity & Creativity