Jipeng Sun will present his General Exam "Scaling Nano-Optics for the Future of Machine Perception" on Wednesday, May 6, 2026 at 10:00 AM in Friend 108. Committee Members: Felix Heide (advisor), Parastoo Abtahi, Kyle Jamieson Abstract: While billions of vision sensors power modern machine perception, the physics of traditional refractive optics limits the form factor, efficiency, and multi-modality of these edge devices. End-to-end optimized nano-optics offer a powerful, compact alternative, but conventional design methods severely oversimplify diffractive modeling due to computational constraints. This failure to capture complex light interactions across varying depths, wavelengths, and angles creates a massive sim-to-real gap, traditionally requiring heavy, power-hungry digital reconstruction to correct. To resolve this gap at the physical layer, this talk introduces a distributed nano-optics optimization framework capable of differentiably simulating real-world optical conditions. This scales nano-optics optimization across unprecedented parameters, including lens counts, polarization states, spectral bands, and macroscopic physical apertures. I will present the new capabilities unlocked by this large-area, fabrication-aware approach—including energy-efficient nano-optical eye tracking for smart glasses, broadband imaging that beats traditional spectral limits, and thin collaborative on-sensor array cameras. Ultimately, this work shifts the computational burden to the speed-of-light physics of the optical front-end, providing a scalable foundation for the next generation of machine eyes. Reading List: https://docs.google.com/document/d/15dZ8wys3EevOj1ZZAZb_aKgiRglExNqO5tsPYHiR... Everyone is invited to attend the talk, and those faculty wishing to remain for the oral exam following are welcome to do so.