Using ``Annotator Rationales'' to Improve Machine Learning for Text Categorization (original) (raw)

Machine Learning with Annotator Rationales to Reduce Annotation Cost

We review two novel methods for text categorization, based on a new framework that utilizes richer annotations that we call annotator rationales. A human annotator provides hints to a machine learner by highlighting contextual "rationales" in support of each of his or her annotations. We have collected such rationales, in the form of substrings, for an existing document sentiment classification dataset [1]. We have developed two methods, one discriminative [2] and one generative , that use these rationales during training to obtain significant accuracy improvements over two strong baselines. Our generative model in particular could be adapted to help learn other kinds of probabilistic classifiers for quite different tasks. Based on a small study of annotation speed, we posit that for some tasks, providing rationales can be a more fruitful use of an annotator's time than annotating more examples.

Automatically Generating Annotator Rationales to Improve Sentiment Classification

2010

One of the central challenges in sentimentbased text categorization is that not every portion of a document is equally informative for inferring the overall sentiment of the document. Previous research has shown that enriching the sentiment labels with human annotators' "rationales" can produce substantial improvements in categorization performance . We explore methods to automatically generate annotator rationales for document-level sentiment classification. Rather unexpectedly, we find the automatically generated rationales just as helpful as human rationales.

Learning with rationales for document classification

Machine Learning

We present a simple and yet effective approach for document classification to incorporate rationales elicited from annotators into the training of any off-the-shelf classifier. We empirically show on several document classification datasets that our classifier-agnostic approach, which makes no assumptions about the underlying classifier, can effectively incorporate rationales into the training of multinomial naïve Bayes, logistic regression, and support vector machines. In addition to being classifier-agnostic, we show that our method has comparable performance to previous classifier-specific approaches developed for incorporating rationales and feature annotations. Additionally, we propose and evaluate an active learning method tailored specifically for the learning with rationales framework.

Towards Learning with Feature-Based Explanations for Document Classification

2016

Traditional supervised learning approaches for document classification ask human labelers to provide labels for documents. Humans know more than just the labels and can provide extra information such as domain knowledge, feature annotations, rules, and rationales for classification. Researchers indeed tried to utilize this extra information for labeling of features and documents in tandem, and more recently for incorporating rationales for the provided labels. We extend this approach further by allowing the human to provide explanations, in the form of domain-specific features that support and oppose the classification of documents, and present an approach to incorporate explanations into the training of any off-the-shelf classifier to speed-up the learning process.

MARTA: Leveraging Human Rationales for Explainable Text Classification

2021

Explainability is a key requirement for text classification in many application domains ranging from sentiment analysis to medical diagnosis or legal reviews. Existing methods often rely on “attention” mechanisms for explaining classification results by estimating the relative importance of input units. However, recent studies have shown that such mechanisms tend to mis-identify irrelevant input units in their explanation. In this work, we propose a hybrid human-AI approach that incorporates human rationales into attention-based text classification models to improve the explainability of classification results. Specifically, we ask workers to provide rationales for their annotation by selecting relevant pieces of text. We introduce MARTA, a Bayesian framework that jointly learns an attention-based model and the reliability of workers while injecting human rationales into model training. We derive a principled optimization algorithm based on variational inference with efficient updat...

Explaining Sentiment Classification

Interspeech 2019

This paper presents a novel 1-D sentiment classifier trained on the benchmark IMDB dataset. The classifier is a 1-D convolutional neural network with repeated convolution and max pooling layers. The main contribution of this work is the demonstration of a deconvolution technique for 1-D convolutional neural networks that is agnostic to specific architecture types. This deconvolution technique enables text classification to be explained, a feature that is important for NLP-based decision support systems, as well as being an invaluable diagnostic tool.

Active Learning with Rationales for Text Classification

Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 2015

We present a simple and yet effective approach that can incorporate rationales elicited from annotators into the training of any offthe-shelf classifier. We show that our simple approach is effective for multinomial naïve Bayes, logistic regression, and support vector machines. We additionally present an active learning method tailored specifically for the learning with rationales framework.

The Many Benefits of Annotator Rationales for Relevance Judgments

Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence (IJCAI-17), 2017

When collecting subjective human ratings of items, it can be difficult to mea- sure and enforce data quality due to task subjectivity and lack of insight into how judges arrive at each rating decision. To address this, we propose requiring judges to provide a specific type of rationale underlying each rating decision. We evaluate this approach in the domain of In- formation Retrieval, where human judges rate the relevance of Webpages. Cost- benefit analysis over 10,000 judgments collected on Mechanical Turk suggests a win-win: experienced crowd workers provide rationales with no increase in task completion time while providing further benefits, including more reliable judgments and greater transparency.

Interpreting Text Classifiers by Learning Context-sensitive Influence of Words

Proceedings of the First Workshop on Trustworthy Natural Language Processing, 2021

Many existing approaches for interpreting text classification models focus on providing importance scores for parts of the input text, such as words, but without a way to test or improve the interpretation method itself. This has the effect of compounding the problem of understanding or building trust in the model, with the interpretation method itself adding to the opacity of the model. Further, importance scores on individual examples are usually not enough to provide a sufficient picture of model behavior. To address these concerns, we propose MOXIE (MOdeling conteXt-sensitive InfluencE of words) with an aim to enable a richer interface for a user to interact with the model being interpreted and to produce testable predictions. In particular, we aim to make predictions for importance scores, counterfactuals and learned biases with MOXIE. In addition, with a global learning objective, MOXIE provides a clear path for testing and improving itself. We evaluate the reliability and efficiency of MOXIE on the task of sentiment analysis.

Knowledge-Guided Sentiment Analysis Via Learning From Natural Language Explanations

IEEE Access, 2021

Sentiment analysis is crucial for studying public opinion since it can provide us with valuable information. Existing sentiment analysis methods rely on finding the sentiment element from the content of user-generated. However, the question of why a message produces certain emotions has not been well explored or utilized in previous works. To address this challenge, we propose a natural language explanation framework for sentiment analysis that provides sufficient domain knowledge for generating additional labelled data for each new labelling decision. A rule-based semantic parser transforms these explanations into programmatic labelling functions that generate noisy labels for an arbitrary amount of unlabelled sentiment information to train a sentiment analysis classifier. Experiments on two sentiment analysis datasets demonstrate the superiority it achieves over baseline methods by leveraging explanations as external knowledge to joint training a sentiment analysis model rather th...