Tutorial 64: Spike Encoding Zoo + Auto-Optimizer

7 encoding schemes and automatic selection.

All 7 Encoders

from sc_neurocore.encoding import (
    rate_encode, latency_encode, delta_encode,
    phase_encode, burst_encode, rank_order_encode, sigma_delta_encode,
)
import numpy as np

values = np.array([0.2, 0.5, 0.8, 0.1])
T = 20

rate_spikes = rate_encode(values, T)           # Poisson rate coding
latency_spikes = latency_encode(values, T)     # Time-to-first-spike
phase_spikes = phase_encode(values, T)         # Phase within oscillation
burst_spikes = burst_encode(values, T)         # Burst length
rank_spikes = rank_order_encode(values, T)     # Firing order

signal = np.sin(np.linspace(0, 4*np.pi, 100))
delta_spikes = delta_encode(signal, threshold=0.2)
sd_spikes = sigma_delta_encode(signal, threshold=0.1)

Auto-Optimizer

from sc_neurocore.encoding import EncodingOptimizer

opt = EncodingOptimizer(T=32)
recs = opt.recommend(my_data)

for r in recs:
    print(f"{r.encoding}: score={r.score:.2f}, sparsity={r.sparsity:.2f}")
    print(f"  {r.reason}")

When to Use Each

Encoding	Best For	Energy
Rate	General purpose, robust	High (many spikes)
Latency	Classification, low-latency	Very low (1 spike)
Delta	Temporal change detection	Low (sparse)
Phase	Periodic signals, oscillations	Medium
Burst	Intensity-critical tasks	Medium
Rank-order	High-dimensional input	Very low
Sigma-delta	Continuous signals, ADC-like	Low