An efficient probabilistic hardware architecture for diffusion-like models
This addresses the energy consumption problem for AI hardware developers, offering a scalable alternative to current inefficient probabilistic computers.
The authors tackled the inefficiency of existing probabilistic hardware for AI by proposing an all-transistor probabilistic computer that implements denoising models at the hardware level, achieving performance parity with GPUs on an image benchmark using about 10,000 times less energy.
The proliferation of probabilistic AI has promoted proposals for specialized stochastic computers. Despite promising efficiency gains, these proposals have failed to gain traction because they rely on fundamentally limited modeling techniques and exotic, unscalable hardware. In this work, we address these shortcomings by proposing an all-transistor probabilistic computer that implements powerful denoising models at the hardware level. A system-level analysis indicates that devices based on our architecture could achieve performance parity with GPUs on a simple image benchmark using approximately 10,000 times less energy.