Compressibility and Multifractal Properties of Self-Compressing Neural Networks
M.S. thesis (SDSU, 2026). Networks that compress their own internal states train faster and more accurately, and their representations turn out to be multifractal.
Working on systems that actually grow
My formal background is in math, but most of my projects are research engineering balancing machine learning and neuroscience: local learning rules, memory, world models, and dynamical systems. This site is a mix of a public notebook and a showcase for projects and writing.
Threads
- Learning without backpropHebbian rules, predictive coding, forward-only and evolutionary learning, benchmarked fairly.1 project
- Memory & the hippocampusGrid cells, attractor networks, and associative memory that learns online.2 projects
- Self-supervised world modelsJoint-embedding predictive architectures, from images to agents that learn skills.2 projects
- Dynamical systems & timeKoopman operators, HiPPO memories, and forecasting chaotic systems.2 projects
- Generation & pattern formationAudio codecs, self-organizing codebooks, diffusion, and growth models from biology.6 projects
- Mathematical foundationsCompression, fractal geometry and information geometry of learning systems.3 projects
Recent writing
- seedling
- seedling
Selected projects
Fire: artificial agency
Experiments in agents that learn on their own: streaming deep RL, unsupervised skill discovery (METRA, CSF), cognitive map learners, and TD-JEPA.
Hippocampus models
A continuous-attractor grid-cell simulator and Vector-HaSH associative memory, extended with online Hebbian learning and metaplasticity tuned by evolutionary search.
Alternatives to backpropagation
Hebbian learning, predictive coding, forward-forward, PEPITA, reservoirs and evolutionary methods, benchmarked against backprop on sample efficiency and wall-clock time.
JEPA & Saccade JEPA
I-JEPA built from the paper before the official release, trained on ImageNet with one GPU, plus Saccade JEPA: an original variant that learns by predicting small eye movements.