Can Kernel Machines Be a Viable Alternative to Deep Neural Networks?

Join this lecture by Prof. Parthe Pandit, Assistant Professor at IIT Bombay, as he explores whether kernel machines can serve as viable alternatives to deep neural networks. Discover the renewed interest in kernel machines following the discovery of the Neural Tangent Kernel and its equivalence to wide neural networks. The talk presents two significant research results demonstrating the potential of kernel machines for modern large-scale applications: data-dependent supervised kernels and fast scalable training algorithms. Prof. Pandit, who holds the Thakur Family Chair at the Center for Machine Intelligence and Data Science at IIT Bombay, brings valuable insights from his experience as a Simons Postdoctoral Fellow at UC San Diego, his PhD from UCLA, and his recognition as an AI2050 Early Career Fellow by Schmidt Sciences in 2024.