An efficient probabilistic hardware architecture for diffusion-like models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908704720289792 |
|---|---|
| author | Jelinčič, Andraž Lockwood, Owen Garlapati, Akhil Schillinger, Peter Chuang, Isaac Verdon, Guillaume McCourt, Trevor |
| author_facet | Jelinčič, Andraž Lockwood, Owen Garlapati, Akhil Schillinger, Peter Chuang, Isaac Verdon, Guillaume McCourt, Trevor |
| contents | The proliferation of probabilistic AI has prompted 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_23972 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | An efficient probabilistic hardware architecture for diffusion-like models Jelinčič, Andraž Lockwood, Owen Garlapati, Akhil Schillinger, Peter Chuang, Isaac Verdon, Guillaume McCourt, Trevor Machine Learning Artificial Intelligence The proliferation of probabilistic AI has prompted 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. |
| title | An efficient probabilistic hardware architecture for diffusion-like models |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2510.23972 |