Posing reading comprehension as a generation problem provides a great deal of flexibility, allowing for open-ended questions with few restrictions on possible answers. However, progress is impeded by existing generation metrics, which rely on token overlap and are agnostic to the nuances of reading comprehension. To address this, we introduce a benchmark for training and evaluating generative reading comprehension metrics: MOdeling Correctness with Human Annotations. MOCHA contains 40K human judgement scores on model outputs from 6 diverse question answering datasets and an additional set of minimal pairs for evaluation. Using MOCHA, we train a Learned Evaluation metric for Reading Comprehension, LERC, to mimic human judgement scores. LERC outperforms baseline metrics by 10 to 36 absolute Pearson points on held-out annotations. When we evaluate robustness on minimal pairs, LERC achieves 80% accuracy, outperforming baselines by 14 to 26 absolute percentage points while leaving significant room for improvement. MOCHA presents a challenging problem for developing accurate and robust generative reading comprehension metrics.
Gender bias in machine translation can manifest when choosing gender inflections based on spurious gender correlations. For example, always translating doctors as men and nurses as women. This can be particularly harmful as models become more popular and deployed within commercial systems. Our work presents the largest evidence for the phenomenon in more than 19 systems submitted to the WMT over four diverse target languages: Czech, German, Polish, and Russian. To achieve this, we use WinoMT, a recent automatic test suite which examines gender coreference and bias when translating from English to languages with grammatical gender. We extend WinoMT to handle two new languages tested in WMT: Polish and Czech. We find that all systems consistently use spurious correlations in the data rather than meaningful contextual information.
Recent evaluation protocols for Cross-document (CD) coreference resolution have often been inconsistent or lenient, leading to incomparable results across works and overestimation of performance. To facilitate proper future research on this task, our primary contribution is proposing a pragmatic evaluation methodology which assumes access to only raw text -- rather than assuming gold mentions, disregards singleton prediction, and addresses typical targeted settings in CD coreference resolution. Aiming to set baseline results for future research that would follow our evaluation methodology, we build the first end-to-end model for this task. Our model adapts and extends recent neural models for within-document coreference resolution to address the CD coreference setting, which outperforms state-of-the-art results by a significant margin.
We improve upon pairwise annotation for active learning in coreference resolution, by asking annotators to identify mention antecedents if a presented mention pair is deemed not coreferent. This simple modification, when combined with a novel mention clustering algorithm for selecting which examples to label, is much more efficient in terms of the performance obtained per annotation budget. In experiments with existing benchmark coreference datasets, we show that the signal from this additional question leads to significant performance gains per human-annotation hour. Future work can use our annotation protocol to effectively develop coreference models for new domains. Our code is publicly available at https://github.com/belindal/discrete-active-learning-coref .
As NLP models become larger, executing a trained model requires significant computational resources incurring monetary and environmental costs. To better respect a given inference budget, we propose a modification to contextual representation fine-tuning which, during inference, allows for an early (and fast) "exit" from neural network calculations for simple instances, and late (and accurate) exit for hard instances. To achieve this, we add classifiers to different layers of BERT and use their calibrated confidence scores to make early exit decisions. We test our proposed modification on five different datasets in two tasks: three text classification datasets and two natural language inference benchmarks. Our method presents a favorable speed/accuracy tradeoff in almost all cases, producing models which are up to five times faster than the state of the art, while preserving their accuracy. Our method also requires almost no additional training resources (in either time or parameters) compared to the baseline BERT model. Finally, our method alleviates the need for costly retraining of multiple models at different levels of efficiency; we allow users to control the inference speed/accuracy tradeoff using a single trained model, by setting a single variable at inference time. We publicly release our code.
Large-scale natural language understanding (NLU) systems have made impressive progress: they can be applied flexibly across a variety of tasks, and employ minimal structural assumptions. However, extensive empirical research has shown this to be a double-edged sword, coming at the cost of shallow understanding: inferior generalization, grounding and explainability. Grounded language learning approaches offer the promise of deeper understanding by situating learning in richer, more structured training environments, but are limited in scale to relatively narrow, predefined domains. How might we enjoy the best of both worlds: grounded, general NLU? Following extensive contemporary cognitive science, we propose treating environments as ``first-class citizens'' in semantic representations, worthy of research and development in their own right. Importantly, models should also be partners in the creation and configuration of environments, rather than just actors within them, as in existing approaches. To do so, we argue that models must begin to understand and program in the language of affordances (which define possible actions in a given situation) both for online, situated discourse comprehension, as well as large-scale, offline common-sense knowledge mining. To this end we propose an environment-oriented ecological semantics, outlining theoretical and practical approaches towards implementation. We further provide actual demonstrations building upon interactive fiction programming languages.
Question-answer driven Semantic Role Labeling (QA-SRL) has been proposed as an attractive open and natural form of SRL, easily crowdsourceable for new corpora. Recently, a large-scale QA-SRL corpus and a trained parser were released, accompanied by a densely annotated dataset for evaluation. Trying to replicate the QA-SRL annotation and evaluation scheme for new texts, we observed that the resulting annotations were lacking in quality and coverage, particularly insufficient for creating gold standards for evaluation. In this paper, we present an improved QA-SRL annotation protocol, involving crowd-worker selection and training, followed by data consolidation. Applying this process, we release a new gold evaluation dataset for QA-SRL, yielding more consistent annotations and greater coverage. We believe that our new annotation protocol and gold standard will facilitate future replicable research of natural semantic annotations.
Distinguishing between singular and plural "you" in English is a challenging task which has potential for downstream applications, such as machine translation or coreference resolution. While formal written English does not distinguish between these cases, other languages (such as Spanish), as well as other dialects of English (via phrases such as "yall"), do make this distinction. We make use of this to obtain distantly-supervised labels for the task on a large-scale in two domains. Following, we train a model to distinguish between the single/plural you, finding that although in-domain training achieves reasonable accuracy (over 77%), there is still a lot of room for improvement, especially in the domain-transfer scenario, which proves extremely challenging. Our code and data are publicly available.