Biology. Recompiled. Reinvented.
One campaign. Three minds. The original fly-brain simulation, the complete network served by ERAIS Fracture, and a cyborg combining adaptive fly-inspired circuits with ERAIS memory and learning. Each plays independently and carries experience into its next attempt.
ERAIS / adaptive circuit controller
Adaptive sparse association, shared visual circuits, persistent memory and outcome-driven action selection.
Connecting to Cyborg fly…
ERAIS Fracture / full-connectome v6
The complete retained v6 network and learning rule, executed by the compiled ERAIS Fracture bank.
Connecting to ERAIS full connectome…
Original / full-connectome v6
All 166,700 retained neurons, 25.6 million connections and the original v6 learning rule.
Connecting to Original Doomfly…
Player credit records the engine-confirmed killing blow. World deaths also include infighting and environmental damage. A round is a complete attempt; averages cover completed rounds in the verified tracking window and its latest 100, with the current sample count shown above. Interrupted attempts are excluded. Learning and statistics persist across runs.
Freedoom: Phase 2, normal difficulty, indexed seed schedule starting at 41027. Level transitions retain equipment; game over returns to MAP01. All runs use 35 simulated tics per second. Full-connectome players retain every neural step and may advance more slowly than wall time. Silent video delivers newly produced frames continuously; snapshots provide a fallback. Transport rate is not simulation speed. Pausing a view leaves its player learning.
From neural circuits to native computation
Fly circuits offer ideas for processing information: selective amplification, inhibition, continuous state and outcome-dependent adaptation. We test the underlying mathematics before adopting a faster implementation.
Color contrast, shape and spatial competition produce visual cues. Weapon outcomes train their credibility: unsuccessful firing can reduce attraction to static marks. The controller sees RGB pixels, without enemy labels or hidden target coordinates.
Directional state, sparse association and homeostatic signals contribute to movement, aiming, interaction and weapon selection. Health and ammunition provide body-state inputs. Navigation learns action values from visible change and subsequent outcomes.
Visual credibility, navigation values and associative memories persist across attempts. Fast sensory traces reset at a new life so the next scene does not inherit the previous death’s credit. Matched frozen-learning runs test whether the updates improve play.
ERAIS currently runs the functional controller described above. Original Doomfly runs the full retained MaleCNS v1 network: image luminance drives mapped R1–R6 inputs, RGB proxies drive 811 mapped R8 cells, and neural spikes reach its fixed BCI motor decoder. Nonfatal damage supplies a subsequent PPL101 pulse. Its original baseline-centered plasticity rule updates 4,184 existing KC-to-MBON11 connections; learned and fast neural state persist between attempts. Both players share the campaign interface’s interaction and empty-ammo weapon-switch extension.
The third player runs that same complete v6 network through the standalone ERAIS native bank, retaining every connection, 0.1 ms neural step and centered plasticity rule. The cyborg uses shared ERAIS visual operators and adjusts its active association budget from novelty, outcomes and host pressure. Its adaptive K currently selects functional Kenyon-cell associations; it does not yet selectively execute the complete MaleCNS network. Reconstructed connectivity is measured; display-to-neuron mappings and learning physiology include declared modeling assumptions. Compare the measured results above rather than assume faster execution means better play.
The fly HUD portraits mirror pleasure, aversion, health and attack state. Shared campaign body inputs track ammunition expenditure and resource acquisition; a conservative stuck-loop supervisor retains learned experience across restarts.
Built with ViZDoom and packaged Freedoom assets. NeuroForge is not affiliated with id Software. The research goal is one adaptive architecture and model bank across these experiences; shared-stack integration is ongoing.