Kyle Shaffer

Creating an abridged version of a text involves shortening it while maintaining its linguistic qualities. In this paper, we examine this task from an NLP perspective for the first time. We present a new resource, AbLit, which is derived from abridged versions of English literature books. The dataset captures passage-level alignments between the original and abridged texts. We characterize the linguistic relations of these alignments, and create automated models to predict these relations as… Show more

Creating an abridged version of a text involves shortening it while maintaining its linguistic qualities. In this paper, we examine this task from an NLP perspective for the first time. We present a new resource, AbLit, which is derived from abridged versions of English literature books. The dataset captures passage-level alignments between the original and abridged texts. We characterize the linguistic relations of these alignments, and create automated models to predict these relations as well as to generate abridgements for new texts. Our findings establish abridgement as a challenging task, motivating future resources and research. The dataset is available at this http URL. Show less

Recent work in multilingual natural language processing has shown progress in various tasks such as natural language inference and joint multilingual translation. Despite success in learning across many languages, challenges arise where multilingual training regimes often boost performance on some languages at the expense of others. For multilingual named entity recognition (NER) we propose a simple technique that groups similar languages together by using embeddings from a pre-trained masked… Show more

Recent work in multilingual natural language processing has shown progress in various tasks such as natural language inference and joint multilingual translation. Despite success in learning across many languages, challenges arise where multilingual training regimes often boost performance on some languages at the expense of others. For multilingual named entity recognition (NER) we propose a simple technique that groups similar languages together by using embeddings from a pre-trained masked language model, and automatically discovering language clusters in this embedding space. Specifically, we fine-tune an XLM-Roberta model on a language identification task, and use embeddings from this model for clustering. We conduct experiments on 15 diverse languages in the WikiAnn dataset and show our technique largely outperforms three baselines: (1) training a multilingual model jointly on all available languages, (2) training one monolingual model per language, and (3) grouping languages by linguistic family. We also conduct analyses showing meaningful multilingual transfer for low-resource languages (Swahili and Yoruba), despite being automatically grouped with other seemingly disparate languages. Show less

In this paper we present a technique to train neural network models on small
amounts of data. Current methods for training neural networks on small amounts
of rich data typically rely on strategies such as fine-tuning a pre-trained neural
network or the use of domain-specific hand-engineered features. Here we take the
approach of treating network layers, or entire networks, as modules and combine
pre-trained modules with untrained modules, to learn the shift in… Show more

In this paper we present a technique to train neural network models on small
amounts of data. Current methods for training neural networks on small amounts
of rich data typically rely on strategies such as fine-tuning a pre-trained neural
network or the use of domain-specific hand-engineered features. Here we take the
approach of treating network layers, or entire networks, as modules and combine
pre-trained modules with untrained modules, to learn the shift in distributions
between data sets. The central impact of using a modular approach comes from
adding new representations to a network, as opposed to replacing representations
via fine-tuning. Using this technique, we are able surpass results using standard
fine-tuning transfer learning approaches, and we are also able to significantly
increase performance over such approaches when using smaller amounts of data. Show less

Students in a Massive Open Online Course (MOOC) interact with each other and the course staff through online discussion forums. While discussion forums play a central role in MOOCs, they also pose a challenge for instructors. The large number of student posts makes it difficult for instructors to know where to intervene to answer questions, resolve issues, and provide feedback.
In this work, we focus on automatically predicting speech acts in MOOC forum posts. Our speech act categories… Show more

Students in a Massive Open Online Course (MOOC) interact with each other and the course staff through online discussion forums. While discussion forums play a central role in MOOCs, they also pose a challenge for instructors. The large number of student posts makes it difficult for instructors to know where to intervene to answer questions, resolve issues, and provide feedback.
In this work, we focus on automatically predicting speech acts in MOOC forum posts. Our speech act categories describe the purpose or function of the post in the ongoing discussion. Specifically, we address three main research questions. First, we investigate whether crowdsourced workers can reliably label MOOC forum posts using our speech act definitions. Second, we investigate
whether our speech acts can help predict instructor interventions and assignment completion and
performance. Finally, we investigate which types of features (derived from the post content, author, and surrounding context) are most effective for predicting our different speech act categories. Show less

The Atherosclerosis Risk in Communities (ARIC) study conducts ongoing surveillance of hospitalized cardiovascular health events and death in 4 communities in the United States (NC, MI, MN and MD). Diagnostic criteria for heart failure (HF) has been manually abstracted from medical records since 2005, including the presence of symptoms consistent with HF decompensation (new onset or worsening shortness of breath, edema, paroxysmal nocturnal dyspnea, and orthopnea) during patients'… Show more

The Atherosclerosis Risk in Communities (ARIC) study conducts ongoing surveillance of hospitalized cardiovascular health events and death in 4 communities in the United States (NC, MI, MN and MD). Diagnostic criteria for heart failure (HF) has been manually abstracted from medical records since 2005, including the presence of symptoms consistent with HF decompensation (new onset or worsening shortness of breath, edema, paroxysmal nocturnal dyspnea, and orthopnea) during patients' hospitalizations. The manual chart abstraction process has high repeatability under a stringent quality control protocol, but is time consuming and costly. The goal of this study is to develop and test natural language processing (NLP) tools to extract information on complex symptoms from free-text electronic medical records. Show less