The Gumbel-max trick is a method to draw a sample from a categorical distribution, given by
its unnormalized (log-) probabilities. Over the past years, the machine learning community …

Search is an important technique in program synthesis that allows for adaptive strategies
such as focusing on particular search directions based on execution results. Several prior …

We develop ancestral Gumbel-Top-k sampling: a generic and efficient method for sampling
without replacement from discrete-valued Bayesian networks, which includes multivariate …

In reasoning about sequential events it is natural to pose probabilistic queries such as
“when will event A occur next” or “what is the probability of A occurring before B”, with …

Beam search is the default decoding strategy for many sequence generation tasks in NLP.
The set of approximate K-best items returned by the algorithm is a useful summary of the …

Beam search is a go-to strategy for decoding neural sequence models. The algorithm can
naturally be viewed as a subset optimization problem, albeit one where the corresponding …

Generative deep learning has become pivotal in molecular design for drug discovery and
materials science. A widely used paradigm is to pretrain neural networks on string …

We consider the problem of automatically constructing computer programs from input-output
examples. We investigate how to augment probabilistic and neural program synthesis …

The constructive approach within Neural Combinatorial Optimization (NCO) treats a
combinatorial optimization problem as a finite Markov decision process, where solutions are …

Current methods for end-to-end constructive neural combinatorial optimization usually train
a policy using behavior cloning from expert solutions or policy gradient methods from …