Kirill P. Kalinin, Jannes Gladrow, Jiaqi Chu, James H. Clegg, Daniel Cletheroe, Douglas J. Kelly, Babak Rahmani, Grace Brennan, Burcu Canakci, Fabian Falck, Michael Hansen, Jim Kleewein, Heiner Kremer, Greg O’Shea, Lucinda Pickup, Saravan Rajmohan, Ant Rowstron, Victor Ruhle, Lee Braine, Shrirang Khedekar, Natalia G. Berloff, Christos Gkantsidis, Francesca Parmigiani, Hitesh Ballani
{"title":"用于人工智能推理和组合优化的模拟光学计算机","authors":"Kirill P. Kalinin, Jannes Gladrow, Jiaqi Chu, James H. Clegg, Daniel Cletheroe, Douglas J. Kelly, Babak Rahmani, Grace Brennan, Burcu Canakci, Fabian Falck, Michael Hansen, Jim Kleewein, Heiner Kremer, Greg O’Shea, Lucinda Pickup, Saravan Rajmohan, Ant Rowstron, Victor Ruhle, Lee Braine, Shrirang Khedekar, Natalia G. Berloff, Christos Gkantsidis, Francesca Parmigiani, Hitesh Ballani","doi":"10.1038/s41586-025-09430-z","DOIUrl":null,"url":null,"abstract":"Artificial intelligence (AI) and combinatorial optimization drive applications across science and industry, but their increasing energy demands challenge the sustainability of digital computing. Most unconventional computing systems1–7 target either AI or optimization workloads and rely on frequent, energy-intensive digital conversions, limiting efficiency. These systems also face application-hardware mismatches, whether handling memory-bottlenecked neural models, mapping real-world optimization problems or contending with inherent analog noise. Here we introduce an analog optical computer (AOC) that combines analog electronics and three-dimensional optics to accelerate AI inference and combinatorial optimization in a single platform. This dual-domain capability is enabled by a rapid fixed-point search, which avoids digital conversions and enhances noise robustness. With this fixed-point abstraction, the AOC implements emerging compute-bound neural models with recursive reasoning potential and realizes an advanced gradient-descent approach for expressive optimization. We demonstrate the benefits of co-designing the hardware and abstraction, echoing the co-evolution of digital accelerators and deep learning models, through four case studies: image classification, nonlinear regression, medical image reconstruction and financial transaction settlement. Built with scalable, consumer-grade technologies, the AOC paves a promising path for faster and sustainable computing. Its native support for iterative, compute-intensive models offers a scalable analog platform for fostering future innovation in AI and optimization. An analog optical computer that combines analog electronics, three-dimensional optics, and an iterative architecture accelerates artificial intelligence inference and combinatorial optimization in a single platform, paving a promising path for faster and sustainable computing.","PeriodicalId":18787,"journal":{"name":"Nature","volume":"645 8080","pages":"354-361"},"PeriodicalIF":48.5000,"publicationDate":"2025-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.comhttps://www.nature.com/articles/s41586-025-09430-z.pdf","citationCount":"0","resultStr":"{\"title\":\"Analog optical computer for AI inference and combinatorial optimization\",\"authors\":\"Kirill P. Kalinin, Jannes Gladrow, Jiaqi Chu, James H. Clegg, Daniel Cletheroe, Douglas J. Kelly, Babak Rahmani, Grace Brennan, Burcu Canakci, Fabian Falck, Michael Hansen, Jim Kleewein, Heiner Kremer, Greg O’Shea, Lucinda Pickup, Saravan Rajmohan, Ant Rowstron, Victor Ruhle, Lee Braine, Shrirang Khedekar, Natalia G. Berloff, Christos Gkantsidis, Francesca Parmigiani, Hitesh Ballani\",\"doi\":\"10.1038/s41586-025-09430-z\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Artificial intelligence (AI) and combinatorial optimization drive applications across science and industry, but their increasing energy demands challenge the sustainability of digital computing. Most unconventional computing systems1–7 target either AI or optimization workloads and rely on frequent, energy-intensive digital conversions, limiting efficiency. These systems also face application-hardware mismatches, whether handling memory-bottlenecked neural models, mapping real-world optimization problems or contending with inherent analog noise. Here we introduce an analog optical computer (AOC) that combines analog electronics and three-dimensional optics to accelerate AI inference and combinatorial optimization in a single platform. This dual-domain capability is enabled by a rapid fixed-point search, which avoids digital conversions and enhances noise robustness. With this fixed-point abstraction, the AOC implements emerging compute-bound neural models with recursive reasoning potential and realizes an advanced gradient-descent approach for expressive optimization. We demonstrate the benefits of co-designing the hardware and abstraction, echoing the co-evolution of digital accelerators and deep learning models, through four case studies: image classification, nonlinear regression, medical image reconstruction and financial transaction settlement. Built with scalable, consumer-grade technologies, the AOC paves a promising path for faster and sustainable computing. Its native support for iterative, compute-intensive models offers a scalable analog platform for fostering future innovation in AI and optimization. An analog optical computer that combines analog electronics, three-dimensional optics, and an iterative architecture accelerates artificial intelligence inference and combinatorial optimization in a single platform, paving a promising path for faster and sustainable computing.\",\"PeriodicalId\":18787,\"journal\":{\"name\":\"Nature\",\"volume\":\"645 8080\",\"pages\":\"354-361\"},\"PeriodicalIF\":48.5000,\"publicationDate\":\"2025-09-03\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.nature.comhttps://www.nature.com/articles/s41586-025-09430-z.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Nature\",\"FirstCategoryId\":\"103\",\"ListUrlMain\":\"https://www.nature.com/articles/s41586-025-09430-z\",\"RegionNum\":1,\"RegionCategory\":\"综合性期刊\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"MULTIDISCIPLINARY SCIENCES\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Nature","FirstCategoryId":"103","ListUrlMain":"https://www.nature.com/articles/s41586-025-09430-z","RegionNum":1,"RegionCategory":"综合性期刊","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MULTIDISCIPLINARY SCIENCES","Score":null,"Total":0}
Analog optical computer for AI inference and combinatorial optimization
Artificial intelligence (AI) and combinatorial optimization drive applications across science and industry, but their increasing energy demands challenge the sustainability of digital computing. Most unconventional computing systems1–7 target either AI or optimization workloads and rely on frequent, energy-intensive digital conversions, limiting efficiency. These systems also face application-hardware mismatches, whether handling memory-bottlenecked neural models, mapping real-world optimization problems or contending with inherent analog noise. Here we introduce an analog optical computer (AOC) that combines analog electronics and three-dimensional optics to accelerate AI inference and combinatorial optimization in a single platform. This dual-domain capability is enabled by a rapid fixed-point search, which avoids digital conversions and enhances noise robustness. With this fixed-point abstraction, the AOC implements emerging compute-bound neural models with recursive reasoning potential and realizes an advanced gradient-descent approach for expressive optimization. We demonstrate the benefits of co-designing the hardware and abstraction, echoing the co-evolution of digital accelerators and deep learning models, through four case studies: image classification, nonlinear regression, medical image reconstruction and financial transaction settlement. Built with scalable, consumer-grade technologies, the AOC paves a promising path for faster and sustainable computing. Its native support for iterative, compute-intensive models offers a scalable analog platform for fostering future innovation in AI and optimization. An analog optical computer that combines analog electronics, three-dimensional optics, and an iterative architecture accelerates artificial intelligence inference and combinatorial optimization in a single platform, paving a promising path for faster and sustainable computing.
期刊介绍:
Nature is a prestigious international journal that publishes peer-reviewed research in various scientific and technological fields. The selection of articles is based on criteria such as originality, importance, interdisciplinary relevance, timeliness, accessibility, elegance, and surprising conclusions. In addition to showcasing significant scientific advances, Nature delivers rapid, authoritative, insightful news, and interpretation of current and upcoming trends impacting science, scientists, and the broader public. The journal serves a dual purpose: firstly, to promptly share noteworthy scientific advances and foster discussions among scientists, and secondly, to ensure the swift dissemination of scientific results globally, emphasizing their significance for knowledge, culture, and daily life.