Mapping ongoing news headlines to event-related classes in a rich knowledge base can be an important component in a knowledge-based event analysis and forecasting solution. In this paper, we present a methodology for creating a benchmark dataset of news headlines mapped to event classes in Wikidata, and resources for the evaluation of methods that perform the mapping. We use the dataset to study two classes of unsupervised methods for this task: 1) adaptations of classic entity linking methods, and 2) methods that treat the problem as a zero-shot text classification problem. For the first approach, we evaluate off-the-shelf entity linking systems. For the second approach, we explore a) pre-trained natural language inference (NLI) models, and b) pre-trained large generative language models. We present the results of our evaluation, lessons learned, and directions for future work. The dataset and scripts for evaluation are made publicly available.

In this paper, we present the event detection models and systems we have developed for Multilingual Protest News Detection - Shared Task 1 at CASE 2021. The shared task has 4 subtasks which cover event detection at different granularity levels (from document level to token level) and across multiple languages (English, Hindi, Portuguese and Spanish). To handle data from multiple languages, we use a multilingual transformer-based language model (XLM-R) as the input text encoder. We apply a variety of techniques and build several transformer-based models that perform consistently well across all the subtasks and languages. Our systems achieve an average F_1 score of 81.2. Out of thirteen subtask-language tracks, our submissions rank 1st in nine and 2nd in four tracks.

Supervised models can achieve very high accuracy for fine-grained text classification. In practice, however, training data may be abundant for some types but scarce or even non-existent for others. We propose a hybrid architecture that uses as much labeled data as available for fine-tuning classification models, while also allowing for types with little (few-shot) or no (zero-shot) labeled data. In particular, we pair a supervised text classification model with a Natural Language Inference (NLI) reranking model. The NLI reranker uses a textual representation of target types that allows it to score the strength with which a type is implied by a text, without requiring training data for the types. Experiments show that the NLI model is very sensitive to the choice of textual representation, but can be effective for classifying unseen types. It can also improve classification accuracy for the known types of an already highly accurate supervised model.

Supervised learning models often perform poorly at low-shot tasks, i.e. tasks for which little labeled data is available for training. One prominent approach for improving low-shot learning is to use unsupervised pre-trained neural models. Another approach is to obtain richer supervision by collecting annotator rationales (explanations supporting label annotations). In this work, we combine these two approaches to improve low-shot text classification with two novel methods: a simple bag-of-words embedding approach; and a more complex context-aware method, based on the BERT model. In experiments with two English text classification datasets, we demonstrate substantial performance gains from combining pre-training with rationales. Furthermore, our investigation of a range of train-set sizes reveals that the simple bag-of-words approach is the clear top performer when there are only a few dozen training instances or less, while more complex models, such as BERT or CNN, require more training data to shine.

Electronic Health Records (EHRs) contain both structured content and unstructured (text) content about a patient’s medical history. In the unstructured text parts, there are common sections such as Assessment and Plan, Social History, and Medications. These sections help physicians find information easily and can be used by an information retrieval system to return specific information sought by a user. However, it is common that the exact format of sections in a particular EHR does not adhere to known patterns. Therefore, being able to predict sections and headers in EHRs automatically is beneficial to physicians. Prior approaches in EHR section prediction have only used text data from EHRs and have required significant manual annotation. We propose using sections from medical literature (e.g., textbooks, journals, web content) that contain content similar to that found in EHR sections. Our approach uses data from a different kind of source where labels are provided without the need of a time-consuming annotation effort. We use this data to train two models: an RNN and a BERT-based model. We apply the learned models along with source data via transfer learning to predict sections in EHRs. Our results show that medical literature can provide helpful supervision signal for this classification task.

Linking spans of natural language text to concepts in a structured source is an important task for many problems. It allows intelligent systems to leverage rich knowledge available in those sources (such as concept properties and relations) to enhance the semantics of the mentions of these concepts in text. In the medical domain, it is common to link text spans to medical concepts in large, curated knowledge repositories such as the Unified Medical Language System. Different approaches have different strengths: some are precision-oriented, some recall-oriented; some better at considering context but more prone to hallucination. The variety of techniques suggests that ensembling could outperform component technologies at this task. In this paper, we describe our process for building a Stacking ensemble using additional, auxiliary features for Entity Linking in the medical domain. We report experiments that show that naive ensembling does not always outperform component Entity Linking systems, that stacking usually outperforms naive ensembling, and that auxiliary features added to the stacker further improve its performance on three distinct datasets. Our best model produces state-of-the-art results on several medical datasets.