All Classes and Interfaces
Class
Description
Abstract base class for a named sub-network of spiking neurons (Nerons) and synapses.
A value updater that applies rational diminishing returns.
A named sub-network of spiking neurons (Nerons) and synapses.
Test harness that wires three
Bottles in a feedforward pipeline for use as a chess-position evaluator.A three-layer spiking network for evaluating chess positions, built on top of the connection-centric
Network
from bottled3.Main driver for the simple-model ChessBrain.
Drives a bottled3
ChessBrain with a fixed chess position and lets the BCM plasticity in the
Connections reshape the weights over many trials.A directed synaptic connection from one neuron to another.
A position evaluator.
A
Bottle with strictly feedforward, layered topology:Game loop and smoke-test main.
1-ply greedy search: for each legal move, make it, ask the evaluator for the
resulting score, and pick the move that is best for the side to move.
A
Bottle with three populations -- inputs, internals, outputs -- wired as:
input → internal (feedforward, sparse)
internal ↔ internal (Hopfield-style recurrence: two independent directional synapses per pair, density
configurable)
internal → output (feedforward, sparse)
Simple material count.
An intended move.
Container for neurons and connections, with a single-threaded driver loop.
Conductance-based spiking neuron with connection-driven plasticity.
A single conductance-based leaky integrate-and-fire neuron with a BCM-style sliding activity threshold.
A chess piece, with color baked in.
Population code over a 1-D scalar range using Gaussian tuning curves.
Immutable chess position.
Functions that return a new value based on the current "state" value and a new input.