We present a series of programming assignments, adaptable to a range of experience levels from advanced undergraduate to PhD, to teach students design and implementation of modern NLP systems.
Conditional neural text generation models generate high-quality outputs, but often concentrate around a mode when what we really want is a diverse set of options.
Across different datasets (CNN/DM, XSum, MediaSum) and summary properties, such as abstractiveness and hallucination, we study what the model learns at different stages of its fine-tuning process.
Disfluencies is an under-studied topic in NLP, even though it is ubiquitous in human conversation.
Despite the prominence of neural abstractive summarization models, we know little about how they actually form summaries and how to understand where their decisions come from.
Compressive summarization systems typically rely on a crafted set of syntactic rules to determine what spans of possible summary sentences can be deleted, then learn a model of what to actually delete by optimizing for content selection (ROUGE).
An advantage of seq2seq abstractive summarization models is that they generate text in a free-form manner, but this flexibility makes it difficult to interpret model behavior.
In this work, we present a neural model for single-document summarization based on joint extraction and syntactic compression.
A hallmark of variational autoencoders (VAEs) for text processing is their combination of powerful encoder-decoder models, such as LSTMs, with simple latent distributions, typically multivariate Gaussians.
In this paper, we propose a novel deep architecture to utilize both structural and textual information of entities.
Recently, neural networks have achieved great success on sentiment classification due to their ability to alleviate feature engineering.