Timezone: »

 
Poster
HOUDINI: Lifelong Learning as Program Synthesis
Lazar Valkov · Dipak Chaudhari · Akash Srivastava · Charles Sutton · Swarat Chaudhuri

Tue Dec 04 07:45 AM -- 09:45 AM (PST) @ Room 210 #97
in Tue Poster Session A »

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.

Keywords: Multitask and Transfer Learning · Program Induction

Author Information

Lazar Valkov (University of Edinburgh)
Dipak Chaudhari (Rice University)
Akash Srivastava (University of Edinburgh)
Charles Sutton (Google)
Swarat Chaudhuri (Rice University)

More from the Same Authors