FlappyLearning
Program learning to play Flappy Bird by machine learning (Neuroevolution)
- 4k
- JavaScript
- MIT
- 3 y ago
Program learning to play Flappy Bird by machine learning (Neuroevolution)
Infrastructure for evolutionary AI systems. GPU-accelerated, PyTorch-based optimization with an ecosystem of algorithm libraries, benchmarks, and tools.
Official Codebase for "Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights" (ICML 2026 Spotlight)
Advanced evolutionary computation library built directly on top of PyTorch, created at NNAISENSE.
MetaDE is a GPU-accelerated evolutionary framework that optimizes Differential Evolution (DE) strategies via meta-level evolution. Supporting both JAX and PyTorch, it dynamically adapts mutation and crossover strategies for efficient large-scale black-box optimization.
Always sparse. Never dense. But never say never. A Sparse Training repository for the Adaptive Sparse Connectivity concept and its algorithmic instantiation, i.e. Sparse Evolutionary Training, to boost Deep Learning scalability on various aspects (e.g. memory and computational time efficiency, representation and generalization power).
A collection of Deep Neuroevolution resources or evolutionary algorithms applying in Deep Learning (constantly updating)
Artificial life. Particles driven by neural networks, fighting on a grid and evolving through genetic algorithms.
Evolutionary Computation: A Modern Perspective |<...>| This is an online book, which is free-access and actively-updated (1st Edition: from 2023 to 2027).
Paper: Challenges in High-dimensional Reinforcement Learning with Evolution Strategies
A Python implementation of the "CoSyne" algorithim, as described in this paper: https://pdfs.semanticscholar.org/966e/41903b4aff42601a188bd7b26d71ef120d11.pdf
GPU-accelerated Evolutionary Multiobjective Optimization Using Tensorized RVEA.
An implementation of "Conventional Neuroevolution" - evolving ANNs without backpropegation
Spiking Decision Transformer: Local Plasticity, Phase-Coding, and Dendritic Routing for Low-Power Sequence Control
Real-time artificial-life laboratory in memory-safe Rust: hundreds of thousands of organisms, each with its own heritable neural network, evolving through a shared energy-conserving chemical field. Watch, inspect, perturb, branch and record them live. Built on asupersync, FrankenTorch, FrankenNumPy and FrankenManim. Design-plan stage.
Artificial life simulator. Agents start with random neural networks and figure out how to survive through natural selection.
SONAPSE-EVO: Causal Neural Circuit Evolution, Autonomous Repair and Adaptive Intelligence
Tiny neural networks trained live and rendered as watchable videos. Night-Hunt: 32 mice share one evolving brain against an owl, learning sense by sense through neuroevolution. PyTorch + Pygame.