EUVIMEDEUVIMED entdecken
Open Access

Hypermultiplexed integrated photonics–based optical tensor processor

Shaoyuan Ou iD, Kaiwen Xue, Lian Zhou, Chun-Ho Lee iD, Alexander Sludds, Ryan Hamerly iD, Ke Zhang iD, Hanke Feng, Yue Yu iD, Reshma Kopparapu, Eric Zhong iD, Cheng Wang iD und 3 weitere

Science Advances · 06.06.2025 · 11 (23)

DOI: 10.1126/sciadv.adu0228 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science Advances, erschienen am 06.06.2025. Verfasst von Shaoyuan Ou und 14 weiteren Autor:innen. Beteiligte Einrichtung: Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA 90089, USA. Die Arbeit wurde bisher 43-mal zitiert (Crossref) und nennt 70 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The escalating data volume and complexity resulting from the rapid expansion of artificial intelligence (AI), Internet of Things (IoT), and 5G/6G mobile networks is creating an urgent need for energy-efficient, scalable computing hardware. Here, we demonstrate a hypermultiplexed tensor optical processor that can perform trillions of operations per second using space-time-wavelength three-dimensional optical parallelism, enabling O(N 2 ) operations per clock cycle with O(N) modulator devices. The system is built with wafer-fabricated III/V micrometer-scale lasers and high-speed thin-film lithium niobate electro-optics for encoding at tens of femtojoules per symbol. Lasing threshold incorporates analog inline rectifier (ReLU) nonlinearity for low-latency activation. The system scalability is verified with machine learning models of 405,000 parameters. A combination of high clock rates, energy-efficient processing, and programmability unlocks the potential of light for low-energy AI accelerators for applications ranging from training of large AI models to real-time decision-making in edge deployment.

Autor:innen und Institutionen

  • Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA 90089, USA.
  • Department of Electrical Engineering and Computer Science, University of California, Berkeley, CA 94720, USA.
  • Research Laboratory of Electronics, MIT, Cambridge, MA 02139, USA.
  • PHI Laboratories, NTT Research Inc., 940 Stewart Drive, Sunnyvale, CA 94085, USA.
  • Department of Electrical Engineering & State Key Laboratory of Terahertz and Millimeter Waves, City University of Hong Kong, Kowloon, Hong Kong, China.
  • Opticore Inc., Berkeley, CA 94705, USA.

ORCID-Profile: Shaoyuan Ou, Chun-Ho Lee, Ryan Hamerly, Ke Zhang, Yue Yu, Eric Zhong, Cheng Wang, Dirk Englund, Mengjie Yu, Zaijun Chen

Zitieren

APA

Ou, S., Xue, K., Zhou, L., Lee, C. H., Sludds, A., Hamerly, R., Zhang, K., Feng, H., Yu, Y., Kopparapu, R., Zhong, E., Wang, C., Englund, D., Yu, M., & Chen, Z. (2025). Hypermultiplexed integrated photonics–based optical tensor processor. Science Advances, 11(23). https://doi.org/10.1126/sciadv.adu0228

Vancouver

Ou S, Xue K, Zhou L, Lee CH, Sludds A, Hamerly R, et al. Hypermultiplexed integrated photonics–based optical tensor processor. Science Advances. 2025;11(23). doi:10.1126/sciadv.adu0228

BibTeX
@article{Ou2025,
  author = {Ou, Shaoyuan and Xue, Kaiwen and Zhou, Lian and Lee, Chun-Ho and Sludds, Alexander and Hamerly, Ryan and Zhang, Ke and Feng, Hanke and Yu, Yue and Kopparapu, Reshma and Zhong, Eric and Wang, Cheng and Englund, Dirk and Yu, Mengjie and Chen, Zaijun},
  title = {Hypermultiplexed integrated photonics–based optical tensor processor},
  journal = {Science Advances},
  year = {2025},
  volume = {11},
  number = {23},
  doi = {10.1126/sciadv.adu0228}
}
Kostenlos starten →