Cerenaut

agi

Research code & algorithms for artificial general intelligence.

sparse-unsupervised-capsules

Sparse unsupervised capsules

adaptive-memory-prediction-framework

Adaptive Memory Prediction Framework

Preprocess-NIST-SD19

Preprocessing for NIST Special Dataset 19 (uppercase single-character handwritten characters A..Z). Converts to same formatting as Yann Lecun MNIST (handwritten numerical digits 0..9), and produces a set of PNG images.

Preprocess-SVHN

Preprocess Google Street View House Number (SVHN) dataset, producing a set of PNG images.

pagi

ProjectAGI's Tensorflow framework for building a selective memory system based on convolutional, hierarchical sparse autoencoder-like components.

cortico-thalamic-circuit

aha

Artificial Hippocampus Algorithm

cerenaut-pt-core

Core components and utils for Cerenaut's PyTorch codebase.

Preprocess-CIFAR

Preprocess CIFAR dataset, creating a set of images.

bilateral-brain

run-framework

Scripts used to run experiments using AGIEF.

bilateral-motor-control

A bilateral neural network that mimics dual hemispheres in humans, applied to motor control.

cfsl

Continual few-shot learning

differentiable-plasticity

Implementations of the algorithms described in Differentiable plasticity: training plastic networks with gradient descent, a research paper from Uber AI Labs.

Preprocess-MNIST

Code to convert the native MNIST data format to PNG images.

rsm

Recurrent Sparse Memory

ARC

The Abstraction and Reasoning Corpus

bilateral

Moved to bilateral-brain

experiment-definitions

Definitions of AGIEF experiments. All experiments can be repeated from these.

pt-aha

AHA implemented in PyTorch. New features are being implemented here (not in the TF version).

WMAR

Model based RL, with world model and replay, for continual RL.