Photonic processor could enable ultrafast AI computations with extreme energy efficiency

The deep neural network models that power today’s machine-learning applications have grown so large and complex that they are pushing the limits of traditional electronic computing hardware.

Photonic hardware, which can perform machine-learning computations with light, offers a faster and more energy-efficient alternative. However, there are some types of computations that a photonic device can’t perform, requiring the use of off-chip electronics that hamper speed and efficiency.

Building on a decade of research, scientists from MIT and elsewhere have developed a new photonic chip that overcomes these roadblocks. They demonstrated a fully integrated photonic processor that can perform all the key computations of a deep neural network optically on the chip.

I hope so, the amount of electricity and water AI is gobbling up is too damn high!

Not an AI expert but I think this is a *massive* deal. I don’t know how these NOFUs work or how fast/small/cheap they are but if this kind of chip can be scaled up and retain much of its benefits compared to silicon, machine learning training and inference will become almost free in the future compared to the status quo. The vast majority of computational power used in ML will no longer require GPUs or similar, tanking a lot of investments made for the current AI boom. Training data will become the only limitation to ML performance – to overcome this we might gravitate more and more towards reinforcement learning methods. Getting a big reinforcement learning model to run on a photonic chip will probably be a major topic in ML going forward.