Aparna Elangovan, Jiayuan He, Yuan Li, Karin Verspoor. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Contrastive learning, which learns data representations by contrasting similar and dissimilar instances, has achieved great success in various domains including natural language processing (NLP). Recently, it has been demonstrated that incorporating class labels into contrastive learning, i.e., supervised contrastive learning (SCL), can further enhance the quality of the learned data representations. Although several works have shown empirically that incorporating SCL into classification models leads to better performance, the mechanism of how SCL works for classification is less studied. In this paper, we first investigate how SCL facilitates the classifier learning, where we show that the contrastive region, i.e., the data instances involved in each contrasting operation, has a crucial link to the mechanism of SCL. We reveal that the vanilla SCL is suboptimal since its behavior can be altered by variances in class distributions. Based on this finding, we propose aFocusedContrastiveLoss (FoCL) for classification. Compared with SCL, FoCL defines a finer contrastive region, focusing on the data instances surrounding decision boundaries. We conduct extensive experiments on three NLP tasks: text classification, named entity recognition, and relation extraction. Experimental results show consistent and significant improvements of FoCL over strong baselines on various benchmark datasets, especially in few-shot scenarios.

MOTIVATION: Protein-protein interactions (PPIs) are critical to normal cellular function and are related to many disease pathways. A range of protein functions are mediated and regulated by protein interactions through post-translational modifications (PTM). However, only 4% of PPIs are annotated with PTMs in biological knowledge databases such as IntAct, mainly performed through manual curation, which is neither time- nor cost-effective. Here we aim to facilitate annotation by extracting PPIs along with their pairwise PTM from the literature by using distantly supervised training data using deep learning to aid human curation. METHOD: We use the IntAct PPI database to create a distant supervised dataset annotated with interacting protein pairs, their corresponding PTM type, and associated abstracts from the PubMed database. We train an ensemble of BioBERT models-dubbed PPI-BioBERT-x10-to improve confidence calibration. We extend the use of ensemble average confidence approach with confidence variation to counteract the effects of class imbalance to extract high confidence predictions. RESULTS AND CONCLUSION: The PPI-BioBERT-x10 model evaluated on the test set resulted in a modest F1-micro 41.3 (P =5 8.1, R = 32.1). However, by combining high confidence and low variation to identify high quality predictions, tuning the predictions for precision, we retained 19% of the test predictions with 100% precision. We evaluated PPI-BioBERT-x10 on 18 million PubMed abstracts and extracted 1.6 million (546507 unique PTM-PPI triplets) PTM-PPI predictions, and filter [Formula: see text] (4584 unique) high confidence predictions. Of the 5700, human evaluation on a small randomly sampled subset shows that the precision drops to 33.7% despite confidence calibration and highlights the challenges of generalisability beyond the test set even with confidence calibration. We circumvent the problem by only including predictions associated with multiple papers, improving the precision to 58.8%. In this work, we highlight the benefits and challenges of deep learning-based text mining in practice, and the need for increased emphasis on confidence calibration to facilitate human curation efforts.

In cross-lingual transfer, NLP models over one or more source languages are applied to a lowresource target language.While most prior work has used a single source model or a few carefully selected models, here we consider a "massive" setting with many such models.This setting raises the problem of poor transfer, particularly from distant languages.We propose two techniques for modulating the transfer, suitable for zero-shot or few-shot learning, respectively.Evaluating on named entity recognition, we show that our techniques are much more effective than strong baselines, including standard ensembling, and our unsupervised method rivals oracle selection of the single best individual model. 1

Crowdsourcing has emerged as a core component of data science pipelines. From collected noisy worker labels, aggregation models that incorporate worker reliability parameters aim to infer a latent true annotation. In this paper, we argue that existing crowdsourcing approaches do not sufficiently model worker correlations observed in practical settings; we propose in response an enhanced Bayesian classifier combination (EBCC) model, with inference based on a mean-field variational approach. An introduced mixture of intra-class reliabilities—connected to tensor decomposition and item clustering—induces inter-worker correlation. EBCC does not suffer the limitations of existing correlation models: intractable marginalisation of missing labels and poor scaling to large worker cohorts. Extensive empirical comparison on 17 real-world datasets sees EBCC achieving the highest mean accuracy across 10 benchmark crowdsourcing methods.

Crowd-sourcing is a cheap and popular means of creating training and evaluation datasets for machine learning, however it poses the problem of 'truth inference', as individual workers cannot be wholly trusted to provide reliable annotations. Research into models of annotation aggregation attempts to infer a latent 'true' annotation, which has been shown to improve the utility of crowd-sourced data. However, existing techniques beat simple baselines only in low redundancy settings, where the number of annotations per instance is low (= 3), or in situations where workers are unreliable and produce low quality annotations (e.g., through spamming, random, or adversarial behaviours.) As we show, datasets produced by crowd-sourcing are often not of this type: the data is highly redundantly annotated (= 5 annotations per instance), and the vast majority of workers produce high quality outputs. In these settings, the majority vote heuristic performs very well, and most truth inference models underperform this simple baseline. We propose a novel technique, based on a Bayesian graphical model with conjugate priors, and simple iterative expectation-maximisation inference. Our technique produces competitive performance to the state-of-the-art benchmark methods, and is the only method that significantly outperforms the majority vote heuristic at one-sided level 0.025, shown by significance tests. Moreover, our technique is simple, is implemented in only 50 lines of code, and trains in seconds. 1

Active learning aims to select a small subset of data for annotation such that a classifier learned on the data is highly accurate.This is usually done using heuristic selection methods, however the effectiveness of such methods is limited and moreover, the performance of heuristics varies between datasets.To address these shortcomings, we introduce a novel formulation by reframing the active learning as a reinforcement learning problem and explicitly learning a data selection policy, where the policy takes the role of the active learning heuristic.Importantly, our method allows the selection policy learned using simulation on one language to be transferred to other languages.We demonstrate our method using cross-lingual named entity recognition, observing uniform improvements over traditional active learning.