Research - Papers
Explore a selection of our published work on a variety of key research challenges in AI.
DiscoFuse: A Large-Scale Dataset for Discourse-based Sentence Fusion
Sentence fusion is the task of joining several independent sentences into a single coherent text. Current datasets for sentence fusion are small and insufficient for training modern neural models.…
Evaluating Text GANs as Language Models
Generative Adversarial Networks (GANs) are a promising approach for text generation that, unlike traditional language models (LM), does not suffer from the problem of “exposure bias”. However, A…
Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them
Word embeddings are widely used in NLP for a vast range of tasks. It was shown that word embeddings derived from text corpora reflect gender biases in society. This phenomenon is pervasive and…
Step-by-Step: Separating Planning from Realization in Neural Data-to-Text Generation
Data-to-text generation can be conceptually divided into two parts: ordering and structuring the information (planning), and generating fluent language describing the information (realization).…
Studying the Inductive Biases of RNNs with Synthetic Variations of Natural Languages
How do typological properties such as word order and morphological case marking affect the ability of neural sequence models to acquire the syntax of a language? Cross-linguistic comparisons of…
Value-based Search in Execution Space for Mapping Instructions to Programs
Training models to map natural language instructions to programs given target world supervision only requires searching for good programs at training time. Search is commonly done using beam search…
White-to-Black: Efficient Distillation of Black-Box Adversarial Attacks
We show that a neural network can learn to imitate the optimization process performed by white-box attack in a much more efficient manner. We train a black-box attack through this imitation process…
Neural network gradient-based learning of black-box function interfaces
Deep neural networks work well at approximating complicated functions when provided with data and trained by gradient descent methods. At the same time, there is a vast amount of existing functions…
Mapping Images to Scene Graphs with Permutation-Invariant Structured Prediction
Machine understanding of complex images is a key goal of artificial intelligence. One challenge underlying this task is that visual scenes contain multiple inter-related objects, and that global…
Memory Augmented Policy Optimization for Program Synthesis and Semantic Parsing
This paper presents Memory Augmented Policy Optimization (MAPO): a novel policy optimization formulation that incorporates a memory buffer of promising trajectories to reduce the variance of policy…