Arbitrary Computation

Arbitrary computation explores methods for performing any computable function, regardless of complexity, across diverse computational platforms. Current research focuses on optimizing algorithms for distributed and encrypted computation, including advancements in stochastic gradient descent, attention mechanisms (generalized to higher-order correlations), and fully homomorphic encryption schemes. These efforts aim to improve efficiency and security in areas like machine learning and data analysis, while also investigating novel architectures such as neuromorphic computing for energy-efficient general-purpose computation. The ultimate goal is to enable powerful and secure computation across a wider range of hardware and applications.