Linear Models of Computation and Program Learning
This work addresses the challenge of program learning for researchers in AI and programming languages, but it appears incremental as it builds on existing paradigms without presenting new experimental results.
The paper tackles the problem of making program learning more tractable by focusing on linear models of computation, such as probabilistic sampling and generalized animation, which allow linear combinations of execution runs. It connects these architectures to recent advances in higher-order probabilistic programming and explores links with partial inconsistency, non-monotonic inference, and vector semantics.
We consider two classes of computations which admit taking linear combinations of execution runs: probabilistic sampling and generalized animation. We argue that the task of program learning should be more tractable for these architectures than for conventional deterministic programs. We look at the recent advances in the "sampling the samplers" paradigm in higher-order probabilistic programming. We also discuss connections between partial inconsistency, non-monotonic inference, and vector semantics.