Emerging integrated photonic neural networks: a device-to-system path toward next-generation AI
Peer-Reviewed Publication
Updates every hour. Last Updated: 30-Oct-2026 14:17 ET (30-Oct-2026 18:17 GMT/UTC)
With artificial intelligence (AI) workloads driving rising demands for computing power, bandwidth, and energy efficiency, integrated photonic neural networks (PNNs) offer a promising route beyond electronic bottlenecks. This review surveys emerging PNN technologies from devices and computing architectures to integration, packaging, and systems, highlighting optoelectronic co-design, key challenges, and their potential as next-generation AI accelerators.
Researchers developed DexAnyTwist, a learning framework that enables dexterous robotic hands to twist many everyday objects, from bottles and valves to screwdrivers and light bulbs. By identifying different modes and assigning expert policies to them, the system improved simulation performance and transferred zero-shot to real-world objects.
Researchers have updated the classic nonsolvent induced phase separation (NIPS) theory. By optimizing a fluid observation cell, the team decoupled the formation of macrovoids and cellular pores and revealed their distinct formation mechanisms. A new thermodynamic parameter was proposed for accurate pore prediction. The developed ultrathin membrane delivered outstanding performance in vanadium flow batteries, outperforming commercial counterparts.
Superconducting diodes, which enable dissipationless one-way current flow, have long been limited by low operating temperatures and scalability. Now, writing in National Science Review, a team from Nanjing University reports scalable high-temperature superconducting diodes based on intrinsic Josephson junctions. The devices operate up to 86 K, show tunable nonreciprocity, and exhibit zero-field memory effects, providing a practical route toward energy-efficient superconducting electronics.
Artificial intelligence is fundamentally transforming chemical engineering across all sub‑disciplines, from reaction engineering and catalyst design to process control, materials discovery, and plant operations. This review highlights how machine learning and digital twins are enabling real‑time process optimization, predictive safety monitoring, and sustainable manufacturing. Foundation models and autonomous laboratories are accelerating molecular and materials discovery, while large language models extract structured knowledge from scientific literature. The convergence of AI with chemical engineering is driving a paradigm shift toward smarter, greener, and more efficient production systems.