The growing demand for high-performance, energy-efficient computation has spurred interest in specialized solvers for combinatorial optimization in the form of Ising machines. These systems aim to find low-energy configurations of a highly non-convex energy landscape defined by a quadratic unconstrained binary optimization (QUBO) problem. Among these, continuous Ising machines are a particularly promising subclass in which the problem is embedded into a dynamical system that evolves continuously in time toward low-energy solutions. This physics-driven approach promises dramatic improvements in time-to-solution and energy efficiency over conventional digital algorithmic solvers but faces challenges in configurability, scalability and robustness.
This presentation provides a structured overview of continuous Ising machine technologies, operating principles, hardware implementations and key challenges. In the second part, we introduce GALENA, a CMOS analog in-memory-compute Ising machine.
19/6/2026 11:00 - 12:00
ESAT Aula L
