HOUDINI: Lifelong Learning as Program Synthesis

Abstract

We present a neurosymbolic framework for the lifelong learning of algorithmic tasks that mix perception and procedural reasoning. Reusing high-level concepts across domains and learning complex procedures are key challenges in lifelong learning. We show that a program synthesis approach that combines gradient descent with combinatorial search over programs can be a more effective response to these challenges than purely neural methods. Our framework, called HOUDINI, represents neural networks as strongly typed, differentiable functional programs that use symbolic higher-order combinators to compose a library of neural functions. Our learning algorithm consists of: (1) a symbolic program synthesizer that performs a type-directed search over parameterized programs, and decides on the library functions to reuse, and the architectures to combine them, while learning a sequence of tasks; and (2) a neural module that trains these programs using stochastic gradient descent. We evaluate HOUDINI on three benchmarks that combine perception with the algorithmic tasks of counting, summing, and shortest-path computation. Our experiments show that HOUDINI transfers high-level concepts more effectively than traditional transfer learning and progressive neural networks, and that the typed representation of networks significantly accelerates the search.

Cited by in corpus (13)

• Compacting, Picking and Growing for Unforgetting Continual Learning
• Learning Interpretable Models with Causal Guarantees
• Imitation-Projected Programmatic Reinforcement Learning
• Synthesizing Action Sequences for Modifying Model Decisions
• Learning Differentiable Programs with Admissible Neural Heuristics
• Learning Neurosymbolic Generative Models via Program Synthesis
• Learning Task-General Representations with Generative Neuro-Symbolic Modeling
• Program Synthesis Guided Reinforcement Learning for Partially Observed Environments
• Formal Methods with a Touch of Magic
• Web Question Answering with Neurosymbolic Program Synthesis
• Learning a Deep Generative Model like a Program: the Free Category Prior
• Type-driven Neural Programming by Example
• Hybrid Memoised Wake-Sleep: Approximate Inference at the Discrete-Continuous Interface