A brain that learns from live experience. System 1 by default: perception, memory, plasticity and action. Optional System 2 for self-observation and slow corrections.
Code for the paper: Putting An End to End-to-End: Gradient-Isolated Learning of Representations
NGC-Learn: Computational Neuroscience and NeuroAI in Python
A zero-dependency TypeScript memory kernel that gives any agent long-term memory. No training or backpropagation: each output reads a few tag-compatible neighbors, updated by a manual Hebbian rule. Links used often are reinforced; idle ones evaporate by half-life. Based on the paper Local Pheromone Network (arXiv:2606.30669).
A lightweight and flexible framework for Hebbian learning in PyTorch.
Flexible Inference for Predictive Coding Networks in JAX.
Implementation/simulation of the predictive forward-forward credit assignment algorithm for training neurobiologically-plausible recurrent neural network models.
Forward Pass Learning and Inference Library, for neural networks and general intelligence, Signal Propagation (sigprop)
Deep Spiking Reinforcement Learning
PyTorch implementation of the paper "Spatio-Temporal Decoupled Learning for Spiking Neural Networks"
Github page for SSDFA
We introduce Local recurrent Predictive coding model termed as Parallel temporal Neural Coding Network. Unlike classical RNNs, our model is pure local and doesn't require computing gradients backward in time; thus computationally more efficient compared to BPTT and can be used for online learning
A Computational Substrate for Self-Organizing Biologically-Plausible AI
Modular Forward-Forward Network with independent processing modules and central coordinator. CIFAR-10: 68.65%.
PyTorch implementation of the paper "Scaling Supervised Local Learning with Augmented Auxiliary Networks"
[TMLR] S-TLLR: STDP-inspired Temporal Local Learning Rule for Spiking Neural Networks
[IJCNN] TESS: A Scalable Temporally and Spatially Local Learning Rule for Spiking Neural Networks
[WACV] LLS: Local Learning Rule for Deep Neural Networks Inspired by Neural Activity Synchronization
Byte-level predictive-coding kernel that learns online from local prediction errors: no backpropagation, no attention matrix, no optimizer. Sparse fixed-fan-in synapses, slot-free distributed episodic memory, lesion-controlled reproducible experiments.
A predictive coding neural network to learn invariant representations from short video clips
To associate your repository with the local-learning topic, visit your repo's landing page and select "manage topics."