Recently, sequence-to-sequence models with attention have been successfully applied in Text-to-speech (TTS). These models can generate near-human speech with a large accurately-transcribed speech corpus. However, preparing such a large data-set is both expensive and laborious. To alleviate the problem of heavy data demand, we propose a novel unsupervised pre-training mechanism in this paper. Specifically, we first use Vector-quantization Variational-Autoencoder (VQ-VAE) to ex-tract the unsupervised linguistic units from large-scale, publicly found, and untranscribed speech. We then pre-train the sequence-to-sequence TTS model by using the<unsupervised linguistic units, audio>pairs. Finally, we fine-tune the model with a small amount of<text, audio>paired data from the target speaker. As a result, both objective and subjective evaluations show that our proposed method can synthesize more intelligible and natural speech with the same amount of paired training data. Besides, we extend our proposed method to the hypothesized low-resource languages and verify the effectiveness of the method using objective evaluation.
DEtection TRansformer (DETR) for object detection reaches competitive performance compared with Faster R-CNN via a transformer encoder-decoder architecture. However, trained with scratch transformers, DETR needs large-scale training data and an extreme long training schedule even on COCO dataset. Inspired by the great success of pre-training transformers in natural language processing, we propose a novel pretext task named random query patch detection in Unsupervised Pre-training DETR (UP-DETR). Specifically, we randomly crop patches from the given image and then feed them as queries to the decoder. The model is pre-trained to detect these query patches from the input image. During the pre-training, we address two critical issues: multi-task learning and multi-query localization. (1) To trade off classification and localization preferences in the pretext task, we find that freezing the CNN backbone is the prerequisite for the success of pre-training transformers. (2) To perform multi-query localization, we develop UP-DETR with multi-query patch detection with attention mask. Besides, UP-DETR also provides a unified perspective for fine-tuning object detection and one-shot detection tasks. In our experiments, UP-DETR significantly boosts the performance of DETR with faster convergence and higher average precision on object detection, one-shot detection and panoptic segmentation. Code and pre-training models: https://github.com/dddzg/up-detr.
It is common within the deep learning community to first pre-train a deep neural network from a large-scale dataset and then fine-tune the pre-trained model to a specific downstream task. Recently, both supervised and unsupervised pre-training approaches to learning representations have achieved remarkable advances, which exploit the discriminative knowledge of labels and the intrinsic structure of data, respectively. It follows natural intuition that both discriminative knowledge and intrinsic structure of the downstream task can be useful for fine-tuning, however, existing fine-tuning methods mainly leverage the former and discard the latter. A question arises: How to fully explore the intrinsic structure of data for boosting fine-tuning? In this paper, we propose Bi-tuning, a general learning framework to fine-tuning both supervised and unsupervised pre-trained representations to downstream tasks. Bi-tuning generalizes the vanilla fine-tuning by integrating two heads upon the backbone of pre-trained representations: a classifier head with an improved contrastive cross-entropy loss to better leverage the label information in an instance-contrast way, and a projector head with a newly-designed categorical contrastive learning loss to fully exploit the intrinsic structure of data in a category-consistent way. Comprehensive experiments confirm that Bi-tuning achieves state-of-the-art results for fine-tuning tasks of both supervised and unsupervised pre-trained models by large margins (e.g. 10.7\% absolute rise in accuracy on CUB in low-data regime).
In this work, we study a new image annotation task named Extractive Tags Summarization (ETS). The goal is to extract important tags from the context lying in an image and its corresponding tags. We adjust some state-of-the-art deep learning models to utilize both visual and textual information. Our proposed solution consists of different widely used blocks like convolutional and self-attention layers, together with a novel idea of combining auxiliary loss functions and the gating mechanism to glue and elevate these fundamental components and form a unified architecture. Besides, we introduce a loss function that aims to reduce the imbalance of the training data and a simple but effective data augmentation technique dedicated to alleviates the effect of outliers on the final results. Last but not least, we explore an unsupervised pre-training strategy to further boost the performance of the model by making use of the abundant amount of available unlabeled data. Our model shows the good results as 90% $F_\text{1}$ score on the public NUS-WIDE benchmark, and 50% $F_\text{1}$ score on a noisy large-scale real-world private dataset. Source code for reproducing the experiments is publicly available at: https://github.com/pixta-dev/labteam
This paper tackles the challenging problem of estimating the intensity of Facial Action Units with few labeled images. Contrary to previous works, our method does not require to manually select key frames, and produces state-of-the-art results with as little as $2\%$ of annotated frames, which are \textit{randomly chosen}. To this end, we propose a semi-supervised learning approach where a spatio-temporal model combining a feature extractor and a temporal module are learned in two stages. The first stage uses datasets of unlabeled videos to learn a strong spatio-temporal representation of facial behavior dynamics based on contrastive learning. To our knowledge we are the first to build upon this framework for modeling facial behavior in an unsupervised manner. The second stage uses another dataset of randomly chosen labeled frames to train a regressor on top of our spatio-temporal model for estimating the AU intensity. We show that although backpropagation through time is applied only with respect to the output of the network for extremely sparse and randomly chosen labeled frames, our model can be effectively trained to estimate AU intensity accurately, thanks to the unsupervised pre-training of the first stage. We experimentally validate that our method outperforms existing methods when working with as little as $2\%$ of randomly chosen data for both DISFA and BP4D datasets, without a careful choice of labeled frames, a time-consuming task still required in previous approaches.
Recent advances in machine learning leverage massive datasets of unlabeled images from the web to learn general-purpose image representations for tasks from image classification to face recognition. But do unsupervised computer vision models automatically learn implicit patterns and embed social biases that could have harmful downstream effects? We develop a novel method for quantifying biased associations between representations of social concepts and attributes in images. We find that state-of-the-art unsupervised models trained on ImageNet, a popular benchmark image dataset curated from internet images, automatically learn racial, gender, and intersectional biases. We replicate 8 previously documented human biases from social psychology, from the innocuous, as with insects and flowers, to the potentially harmful, as with race and gender. Our results closely match three hypotheses about intersectional bias from social psychology. For the first time in unsupervised computer vision, we also quantify implicit human biases about weight, disabilities, and several ethnicities. When compared with statistical patterns in online image datasets, our findings suggest that machine learning models can automatically learn bias from the way people are stereotypically portrayed on the web.
Pre-trained contextual vision-and-language (V&L) models have achieved impressive performance on various benchmarks. However, existing models require a large amount of parallel image-caption data for pre-training. Such data are costly to collect and require cumbersome curation. Inspired by unsupervised machine translation, we investigate if a strong V&L representation model can be learned through unsupervised pre-training without image-caption corpora. In particular, we propose to conduct ``mask-and-predict'' pre-training on text-only and image-only corpora and introduce the object tags detected by an object recognition model as anchor points to bridge two modalities. We find that such a simple approach achieves performance close to a model pre-trained with aligned data, on four English V&L benchmarks. Our work challenges the widely held notion that aligned data is necessary for V&L pre-training, while significantly reducing the amount of supervision needed for V&L models.
Recently, Transformer-based language models have demonstrated remarkable performance across many NLP domains. However, the unsupervised pre-training step of these models suffers from unbearable overall computational expenses. Current methods for accelerating the pre-training either rely on massive parallelism with advanced hardware or are not applicable to language modeling. In this work, we propose a method based on progressive layer dropping that speeds the training of Transformer-based language models, not at the cost of excessive hardware resources but from model architecture change and training technique boosted efficiency. Extensive experiments on BERT show that the proposed method achieves a 24% time reduction on average per sample and allows the pre-training to be 2.5 times faster than the baseline to get a similar accuracy on downstream tasks. While being faster, our pre-trained models are equipped with strong knowledge transferability, achieving comparable and sometimes higher GLUE score than the baseline when pre-trained with the same number of samples.
Side information of items, e.g., images and text description, has shown to be effective in contributing to accurate recommendations. Inspired by the recent success of pre-training models on natural language and images, we propose a pre-training strategy to learn item representations by considering both item side information and their relationships. We relate items by common user activities, e.g., co-purchase, and construct a homogeneous item graph. This graph provides a unified view of item relations and their associated side information in multimodality. We develop a novel sampling algorithm named MCNSampling to select contextual neighbors for each item. The proposed Pre-trained Multimodal Graph Transformer (PMGT) learns item representations with two objectives: 1) graph structure reconstruction, and 2) masked node feature reconstruction. Experimental results on real datasets demonstrate that the proposed PMGT model effectively exploits the multimodality side information to achieve better accuracies in downstream tasks including item recommendation, item classification, and click-through ratio prediction. We also report a case study of testing the proposed PMGT model in an online setting with 600 thousand users.
Large pre-trained language models such as BERT have been the driving force behind recent improvements across many NLP tasks. However, BERT is only trained to predict missing words - either behind masks or in the next sentence - and has no knowledge of lexical, syntactic or semantic information beyond what it picks up through unsupervised pre-training. We propose a novel method to explicitly inject linguistic knowledge in the form of word embeddings into any layer of a pre-trained BERT. Our performance improvements on multiple semantic similarity datasets when injecting dependency-based and counter-fitted embeddings indicate that such information is beneficial and currently missing from the original model. Our qualitative analysis shows that counter-fitted embedding injection particularly helps with cases involving synonym pairs.
Self-training and unsupervised pre-training have emerged as effective approaches to improve speech recognition systems using unlabeled data. However, it is not clear whether they learn similar patterns or if they can be effectively combined. In this paper, we show that pseudo-labeling and pre-training with wav2vec 2.0 are complementary in a variety of labeled data setups. Using just 10 minutes of labeled data from Libri-light as well as 53k hours of unlabeled data from LibriVox achieves WERs of 3.0%/5.2% on the clean and other test sets of Librispeech - rivaling the best published systems trained on 960 hours of labeled data only a year ago. Training on all labeled data of Librispeech achieves WERs of 1.5%/3.1%.
Existing approaches for automated essay scoring and document representation learning typically rely on discourse parsers to incorporate discourse structure into text representation. However, the performance of parsers is not always adequate, especially when they are used on noisy texts, such as student essays. In this paper, we propose an unsupervised pre-training approach to capture discourse structure of essays in terms of coherence and cohesion that does not require any discourse parser or annotation. We introduce several types of token, sentence and paragraph-level corruption techniques for our proposed pre-training approach and augment masked language modeling pre-training with our pre-training method to leverage both contextualized and discourse information. Our proposed unsupervised approach achieves new state-of-the-art result on essay Organization scoring task.
In this work we propose for the first time a transformer-based framework for unsupervised representation learning of multivariate time series. Pre-trained models can be potentially used for downstream tasks such as regression and classification, forecasting and missing value imputation. By evaluating our models on several benchmark datasets for multivariate time series regression and classification, we show that not only does our modeling approach represent the most successful method employing unsupervised learning of multivariate time series presented to date, but also that it exceeds the current state-of-the-art performance of supervised methods; it does so even when the number of training samples is very limited, while offering computational efficiency. Finally, we demonstrate that unsupervised pre-training of our transformer models offers a substantial performance benefit over fully supervised learning, even without leveraging additional unlabeled data, i.e., by reusing the same data samples through the unsupervised objective.
Unsupervised pre-training has led to much recent progress in natural language understanding. In this paper, we study self-training as another way to leverage unlabeled data through semi-supervised learning. To obtain additional data for a specific task, we introduce SentAugment, a data augmentation method which computes task-specific query embeddings from labeled data to retrieve sentences from a bank of billions of unlabeled sentences crawled from the web. Unlike previous semi-supervised methods, our approach does not require in-domain unlabeled data and is therefore more generally applicable. Experiments show that self-training is complementary to strong RoBERTa baselines on a variety of tasks. Our augmentation approach leads to scalable and effective self-training with improvements of up to 2.6% on standard text classification benchmarks. Finally, we also show strong gains on knowledge-distillation and few-shot learning.
We explore the suitability of unsupervised representation learning methods on biomedical text -- BioBERT, SciBERT, and BioSentVec -- for biomedical question answering. To further improve unsupervised representations for biomedical QA, we introduce a new pre-training task from unlabeled data designed to reason about biomedical entities in the context. Our pre-training method consists of corrupting a given context by randomly replacing some mention of a biomedical entity with a random entity mention and then querying the model with the correct entity mention in order to locate the corrupted part of the context. This de-noising task enables the model to learn good representations from abundant, unlabeled biomedical text that helps QA tasks and minimizes the train-test mismatch between the pre-training task and the downstream QA tasks by requiring the model to predict spans. Our experiments show that pre-training BioBERT on the proposed pre-training task significantly boosts performance and outperforms the previous best model from the 7th BioASQ Task 7b-Phase B challenge.
This paper describes the system designed by ERNIE Team which achieved the first place in SemEval-2020 Task 10: Emphasis Selection For Written Text in Visual Media. Given a sentence, we are asked to find out the most important words as the suggestion for automated design. We leverage the unsupervised pre-training model and finetune these models on our task. After our investigation, we found that the following models achieved an excellent performance in this task: ERNIE 2.0, XLM-ROBERTA, ROBERTA and ALBERT. We combine a pointwise regression loss and a pairwise ranking loss which is more close to the final M atchm metric to finetune our models. And we also find that additional feature engineering and data augmentation can help improve the performance. Our best model achieves the highest score of 0.823 and ranks first for all kinds of metrics
Autonomous surgical procedures, in particular minimal invasive surgeries, are the next frontier for Artificial Intelligence research. However, the existing challenges include precise identification of the human anatomy and the surgical settings, and modeling the environment for training of an autonomous agent. To address the identification of human anatomy and the surgical settings, we propose a deep learning based semantic segmentation algorithm to identify and label the tissues and organs in the endoscopic video feed of the human torso region. We present an annotated dataset, m2caiSeg, created from endoscopic video feeds of real-world surgical procedures. Overall, the data consists of 307 images, each of which is annotated for the organs and different surgical instruments present in the scene. We propose and train a deep convolutional neural network for the semantic segmentation task. To cater for the low quantity of annotated data, we use unsupervised pre-training and data augmentation. The trained model is evaluated on an independent test set of the proposed dataset. We obtained a F1 score of 0.33 while using all the labeled categories for the semantic segmentation task. Secondly, we labeled all instruments into an 'Instruments' superclass to evaluate the model's performance on discerning the various organs and obtained a F1 score of 0.57. We propose a new dataset and a deep learning method for pixel level identification of various organs and instruments in a endoscopic surgical scene. Surgical scene understanding is one of the first steps towards automating surgical procedures.
We present a self-supervised Contrastive Video Representation Learning (CVRL) method to learn spatiotemporal visual representations from unlabeled videos. Our representations are learned using a contrastive loss, where two augmented clips from the same short video are pulled together in the embedding space, while clips from different videos are pushed away. We study what makes for good data augmentations for video self-supervised learning and find that both spatial and temporal information are crucial. We carefully design data augmentations involving spatial and temporal cues. Concretely, we propose a temporally consistent spatial augmentation method to impose strong spatial augmentations on each frame of the video while maintaining the temporal consistency across frames. We also propose a sampling-based temporal augmentation method to avoid overly enforcing invariance on clips that are distant in time. On Kinetics-600, a linear classifier trained on the representations learned by CVRL achieves 70.4% top-1 accuracy with a 3D-ResNet-50 (R3D-50) backbone, outperforming ImageNet supervised pre-training by 15.7% and SimCLR unsupervised pre-training by 18.8% using the same inflated R3D-50. The performance of CVRL can be further improved to 72.9% with a larger R3D-152 (2x filters) backbone, significantly closing the gap between unsupervised and supervised video representation learning. Our code and models will be available at https://github.com/tensorflow/models/tree/master/official/.
Unsupervised and self-supervised learning approaches have become a crucial tool to learn representations for downstream prediction tasks. While these approaches are widely used in practice and achieve impressive empirical gains, their theoretical understanding largely lags behind. Towards bridging this gap, we present a unifying perspective where several such approaches can be viewed as imposing a regularization on the representation via a learnable function using unlabeled data. We propose a discriminative theoretical framework for analyzing the sample complexity of these approaches, which generalizes the framework of (Balcan and Blum, 2010) to allow learnable regularization functions. Our sample complexity bounds show that, with carefully chosen hypothesis classes to exploit the structure in the data, these learnable regularization functions can prune the hypothesis space, and help reduce the amount of labeled data needed. We then provide two concrete examples of functional regularization, one using auto-encoders and the other using masked self-supervision, and apply our framework to quantify the reduction in the sample complexity bound of labeled data. We also provide complementary empirical results to support our analysis.
Building Automatic Speech Recognition (ASR) systems from scratch is significantly challenging, mostly due to the time-consuming and financially-expensive process of annotating a large amount of audio data with transcripts. Although several unsupervised pre-training models have been proposed, applying such models directly might still be sub-optimal if more labeled, training data could be obtained without a large cost. In this paper, we present a weakly supervised framework for constructing ASR systems with massive video data. As videos often contain human-speech audios aligned with subtitles, we consider videos as an important knowledge source, and propose an effective approach to extract high-quality audios aligned with transcripts from videos based on Optical Character Recognition (OCR). The underlying ASR model can be fine-tuned to fit any domain-specific target training datasets after weakly supervised pre-training. Extensive experiments show that our framework can easily produce state-of-the-art results on six public datasets for Mandarin speech recognition.