Few-shot intent recognition aims to identify user’s intent from the utterance with limited training data. A considerable number of existing methods mainly rely on the generic knowledge acquired on the base classes to identify the novel classes. Such methods typically ignore the characteristics of each meta task itself, resulting in the inability to make full use of limited given samples when classifying unseen classes. To deal with such issues, we propose a Contrastive learning-based Task Adaptation model (CTA) for few-shot intent recognition. In detail, we leverage contrastive learning to help achieve task adaptation and make full use of the limited samples of novel classes. First, a self-attention layer is employed in the task adaptation module, which aims to establish interactions between samples of different categories so that new representations are task-specific rather than relying entirely on the base classes. Then, the contrastive-based loss functions and the semantics of the label name are respectively used for reducing the similarity between sample representations in different categories while increasing it in the same categories. Experimental results on a public dataset OOS verify the effectiveness of our proposal by beating the competitive baselines in terms of accuracy. Besides, we conduct the cross-domain experiments on three datasets, i.e., OOS, SNIPS as well as ATIS. We find that CTA gains obvious improvements in terms of accuracy in all cross-domain experiments, indicating that it has a better generalization ability than other competitive baselines in both cross-domain and single-domain settings. 

A Contrastive learning-based Task Adaptation model for few-shot intent recognition

Contextually aware intelligent agents are often required to understand the users and their surroundings in real-time. Our goal is to build Artificial Intelligence (AI) systems that can assist children in their learning process. Within such complex frameworks, Spoken Dialogue Systems (SDS) are crucial building blocks to handle efficient task-oriented communication with children in game-based learning settings. We are working towards a multimodal dialogue system for younger kids learning basic math concepts. Our focus is on improving the Natural Language Understanding (NLU) module of the task-oriented SDS pipeline with limited datasets. This work explores the potential benefits of data augmentation with paraphrase generation for the NLU models trained on small task-specific datasets. We also investigate the effects of extracting entities for conceivably further data expansion. We have shown that paraphrasing with model-in-the-loop (MITL) strategies using small seed data is a promising approach yielding improved performance results for the Intent Recognition task.

Data Augmentation with Paraphrase Generation and Entity Extraction for Multimodal Dialogue System

We present LINGUIST, a method for generating annotated data for Intent Classification and Slot Tagging (IC+ST), via fine-tuning AlexaTM 5B, a 5-billion-parameter multilingual sequence-to-sequence (seq2seq) model, on a flexible instruction prompt. In a 10-shot novel intent setting for the SNIPS dataset, LINGUIST surpasses state-of-the-art approaches (Back-Translation and Example Extrapolation) by a wide margin, showing absolute improvement for the target intents of +1.9 points on IC Recall and +2.5 points on ST F1 Score. In the zero-shot cross-lingual setting of the mATIS++ dataset, LINGUIST out-performs a strong baseline of Machine Translation with Slot Alignment by +4.14 points absolute on ST F1 Score across 6 languages, while matching performance on IC. Finally, we verify our results on an internal large-scale multilingual dataset for conversational agent IC+ST and show significant improvements over a baseline which uses Back-Translation, Paraphrasing and Slot Catalog Resampling. To our knowledge, we are the first to demonstrate instruction fine-tuning of a large-scale seq2seq model to control the outputs of multilingual intent- and slot-labeled data generation.

LINGUIST: Language Model Instruction Tuning to Generate Annotated Utterances for Intent Classification and Slot Tagging

A significant part of natural language processing (NLP) techniques for sentiment analysis is based on supervised methods, which are affected by the quality of data. Therefore, sentiment analysis needs to be prepared for data quality issues, such as imbalance and lack of labeled data. Data augmentation methods, widely adopted in image classification tasks, include data-space solutions to tackle the problem of limited data and enhance the size and quality of training datasets to provide better models. In this work, we study the advantages and drawbacks of text augmentation methods such as easy data augmentation, back-translation, BART, and pretrained data augmentor) with recent classification algorithms (long short-term memory, convolutional neural network, bidirectional encoder representations of transformers, support vector machine, gated recurrent units, random forests, and enhanced language representation with informative entities, that have attracted sentiment-analysis researchers and industry applications. We explored seven sentiment-analysis datasets to provide scenarios of imbalanced datasets and limited data to discuss the influence of a given classifier in overcoming these problems, and provide insights into promising combinations of transformation, paraphrasing, and generation methods of sentence augmentation. The results revealed improvements from the augmented dataset, mainly for reduced datasets. Furthermore, when balanced by augmenting the minority class, the datasets were found to have improved quality, leading to more robust classifiers. The contributions to this article include the taxonomy of NLP augmentation methods and their efficiency over several classifiers from recent research trends in sentiment analysis and related fields. 

Toward Text Data Augmentation for Sentiment Analysis

Use of synthetic data is rapidly emerging as a realistic alternative to manually annotating live traffic for industry-scale model building. Manual data annotation is slow, expensive and not preferred for meeting customer privacy expectations. Further, commercial natural language applications are required to support continuously evolving features as well as newly added experiences. To address these requirements, we propose a targeted synthetic data generation technique by inserting tokens into a given semantic signature. The generated data are used as additional training samples in the tasks of intent classification and named entity recognition. We evaluate on a real-world voice assistant dataset, and using only 33% of the available training set, we achieve the same accuracy as training with all available data. Further, we analyze the effects of data generation across varied real-world applications and propose heuristics that improve the task performance further.

Controlled Data Generation via Insertion Operations for NLU

Recent progress through advanced neural models pushed the performance of task-oriented dialog systems to almost perfect accuracy on existing benchmark datasets for intent classification and slot labeling. However, in evolving real-world dialog systems, where new functionality is regularly added, a major additional challenge is the lack of annotated training data for such new functionality, as the necessary data collection efforts are laborious and time-consuming. A potential solution to reduce the effort is to augment initial seed data by paraphrasing existing utterances automatically. In this paper, we propose a new, data-efficient approach following this idea. Using an interpretation-to-text model for paraphrase generation, we are able to rely on existing dialog system training data, and, in combination with shuffling-based sampling techniques, we can obtain diverse and novel paraphrases from small amounts of seed data. In experiments on a public dataset and with a real-world dialog system, we observe improvements for both intent classification and slot labeling, demonstrating the usefulness of our approach.

/pdf/data-efficient-paraphrase-generation-to-bootstrap-intent-7le812hnur.pdf

Data-efficient paraphrase generation to bootstrap intent classification and slot labeling for new features in task-oriented dialog systems

In large-scale commercial dialog systems, users express the same request in a wide variety of alternative ways with a long tail of less frequent alternatives. Handling the full range of this distribution is challenging, in particular when relying on manual annotations. However, the same users also provide useful implicit feedback as they often paraphrase an utterance if the dialog system failed to understand it. We propose MARUPA, a method to leverage this type of feedback by creating annotated training examples from it. MARUPA creates new data in a fully automatic way, without manual intervention or effort from annotators, and specifically for currently failing utterances. By re-training the dialog system on this new data, accuracy and coverage for long-tail utterances can be improved. In experiments, we study the effectiveness of this approach in a commercial dialog system across various domains and three languages.

/pdf/leveraging-user-paraphrasing-behavior-in-dialog-systems-to-4jrid9jjgq.pdf

Leveraging User Paraphrasing Behavior In Dialog Systems To Automatically Collect Annotations For Long-Tail Utterances

The lack of labeled training data for new features is a common problem in rapidly changing real-world dialog systems. As a solution, we propose a multilingual paraphrase generation model that can be used to generate novel utterances for a target feature and target language. The generated utterances can be used to augment existing training data to improve intent classification and slot labeling models. We evaluate the quality of generated utterances using intrinsic evaluation metrics and by conducting downstream evaluation experiments with English as the source language and nine different target languages. Our method shows promise across languages, even in a zero-shot setting where no seed data is available.

/pdf/multilingual-paraphrase-generation-for-bootstrapping-new-2szx0mfur9.pdf

Caglar Tirkaz

Papers

Data-efficient paraphrase generation to bootstrap intent classification and slot labeling for new features in task-oriented dialog systems

Leveraging User Paraphrasing Behavior In Dialog Systems To Automatically Collect Annotations For Long-Tail Utterances

Multilingual Paraphrase Generation For Bootstrapping New Features in Task-Oriented Dialog Systems