Reservoir computing experiment - a Liquid State Machine…
0 reactions
Reservoir computing experiment - a Liquid State Machine with simulated
biological constraints
Built a reservoir computing system (Liquid State Machine) as a learning experiment. Instead of a standard static reservoir, I added biological simulation layers on top to see how constraints affect behavior.
What it actually does (no BS):
\- LSM with 2000+ reservoir neurons, Numba JIT-accelerated
\- Hebbian + STDP plasticity (the reservoir rewires during runtime)
\- Neurogenesis/atrophy reservoir can grow or shrink neurons dynamically
\- A hormone system (3 floats: dopamine, cortisol, oxytocin) that modulates learning rate, reflex sensitivity, and noise injection
\- Pain : gaussian noise injected into reservoir state, degrades performance
\- Differential retina (screen capture → |frame(t) - frame(t-1)|) as input
\- Ridge regression readout layer, trained online
What it does NOT do:
\- It's NOT a general intelligence but you should integrate LLM in future (LSM as main brain and LLM as second brain)
\- The "personality" and "emotions" are parameter modulation, not emergent
Why I built it:
wanted to explore whether adding biological constraints (fatigue, pain,hormone cycles) to a reservoir computer creates interesting dynamics vs a vanilla LSM. It does the system genuinely behaves differently based on its "state." Whether that's useful is debatable.
14 Python modules, \~8000 lines, runs fully local (no APIs).
GitHub: https://github.com/JeevanJoshi2061/Project-Genesis-LSM.git
Curious if anyone has done similar work with constrained reservoir computing or bio-inspired dynamics.
#technology
biological constraints
Built a reservoir computing system (Liquid State Machine) as a learning experiment. Instead of a standard static reservoir, I added biological simulation layers on top to see how constraints affect behavior.
What it actually does (no BS):
\- LSM with 2000+ reservoir neurons, Numba JIT-accelerated
\- Hebbian + STDP plasticity (the reservoir rewires during runtime)
\- Neurogenesis/atrophy reservoir can grow or shrink neurons dynamically
\- A hormone system (3 floats: dopamine, cortisol, oxytocin) that modulates learning rate, reflex sensitivity, and noise injection
\- Pain : gaussian noise injected into reservoir state, degrades performance
\- Differential retina (screen capture → |frame(t) - frame(t-1)|) as input
\- Ridge regression readout layer, trained online
What it does NOT do:
\- It's NOT a general intelligence but you should integrate LLM in future (LSM as main brain and LLM as second brain)
\- The "personality" and "emotions" are parameter modulation, not emergent
Why I built it:
wanted to explore whether adding biological constraints (fatigue, pain,hormone cycles) to a reservoir computer creates interesting dynamics vs a vanilla LSM. It does the system genuinely behaves differently based on its "state." Whether that's useful is debatable.
14 Python modules, \~8000 lines, runs fully local (no APIs).
GitHub: https://github.com/JeevanJoshi2061/Project-Genesis-LSM.git
Curious if anyone has done similar work with constrained reservoir computing or bio-inspired dynamics.
#technology
| 4,000 mlx | |
| 8 | |
| 0 |