Timezone: »

 
Poster
Learning Discrete Energy-based Models via Auxiliary-variable Local Exploration
Hanjun Dai · Rishabh Singh · Bo Dai · Charles Sutton · Dale Schuurmans

Tue Dec 08 09:00 PM -- 11:00 PM (PST) @ Poster Session 2 #654
in Poster Session 3 »

Discrete structures play an important role in applications like program language modeling and software engineering. Current approaches to predicting complex structures typically consider autoregressive models for their tractability, with some sacrifice in flexibility.
Energy-based models (EBMs) on the other hand offer a more flexible and thus more powerful approach to modeling such distributions, but require partition function estimation. In this paper we propose \modelshort, a new algorithm for learning conditional and unconditional EBMs for discrete structured data, where parameter gradients are estimated using a learned sampler that mimics local search. We show that the energy function and sampler can be trained efficiently via a new variational form of power iteration, achieving a better trade-off between flexibility and tractability. Experimentally, we show that learning local search leads to significant improvements in challenging application domains. Most notably, we present an energy model guided fuzzer for software testing that achieves comparable performance to well engineered fuzzing engines like libfuzzer.

Author Information

Hanjun Dai (Google Brain)
Rishabh Singh (Google Brain)
Bo Dai (Google Brain)
Charles Sutton (Google)
Dale Schuurmans (Google Brain & University of Alberta)

More from the Same Authors