AALOK
Built at LOQM · IIT Bombay
India's first free-space PELM

AALOK

Analog Adaptive Learning via Optical Kernels

A single free-space photonic extreme learning machine that learns images, audio, and tabular data in one fixed optical pipeline — light does the feature map, only the readout is trained.

96.56% MNIST 95.67% spoken digits 100.0% Mushroom 0.0699 Abalone NRMSE
embedding mask · Fourier carrier
The optical pipeline

One apparatus. Every modality.

Data becomes a phase pattern on a spatial light modulator, free-space propagation does the high-dimensional mixing, an iris keeps only the informative first diffracted order, and a camera reads out 4096 intensity features. Nothing in the optics is trained.

Laser
532 nm
coherent CW
Phase SLM
Holoeye · 1920×1200
8 µm · φ = x + W
4f + iris
Fourier mixing
first order only
CMOS camera
|u|² intensity
4096 features
Σ
Ridge readout
trained layer
the only weights
Hardware interface · encoder

Build a run, get the control code.

Choose a modality and embedding. The mask on the right is computed live from the paper's exact equations — the same fixed phase pattern the SLM would display — and the script updates to drive the SLM, capture the frame, bin to 4096 features, and train the readout.

12
repo default 2000 · lowered here so the carrier stays legible
0.10
W · embedding mask
φ = x + W · encoded
0
π embedding phase · φ wraps to 2π
aalok_mnist_fourier.py

            
Measured performance

Four modalities, one fixed reservoir.

Test performance with the Fourier embedding after independent ridge-λ sweeps. To our knowledge this is the first free-space PELM spanning image, audio-derived, and tabular tasks in a single physical pipeline, and the first with spectrogram-based spoken-digit classification.

ModalityTaskBest λTest
MNIST10-class handwritten digits7.28×10⁻⁴96.56%
FSDDSpoken digits · log-Mel spectrogram1.89×10⁻³95.67%
MushroomBinary tabular · edible vs poisonous1.0×10⁻⁵100.00%
AbaloneTabular regression · shell-ring age4.89×10⁻³0.0704 NRMSE