Artificial intelligence and photonics are developing in a reciprocal relationship: AI helps design and interpret optical systems, while photonics offers a platform for fast, parallel, and potentially energy-efficient computation. This session explores how optical hardware and algorithms can be designed together to improve the way information is acquired, processed, and interpreted.
Physics-informed learning, differentiable optimization, and generative methods are opening new possibilities for inverse design while bringing fabrication and experimental constraints into the design process. This hardware–algorithm perspective also extends to advanced imaging, where optical encoding, data acquisition, and computational reconstruction can be optimized as one system. In turn, diffractive optics and integrated photonic processors can use the physical propagation and interference of light to perform computation relevant to AI inference. Progress in these areas is closely connected: better design and calibration methods enable more capable optical hardware, while the demands of imaging and computing reveal the performance and robustness challenges that those methods must address.
We welcome theoretical, computational, and experimental contributions on this interaction between AI and photonics, from device concepts and physical models to system demonstrations.