Self-supervised learning methods have witnessed a recent surge of interest after proving successful in multiple application fields. In this work, we leverage these techniques, and we propose 3D versions for five different self-supervised methods, in the form of proxy tasks. Our methods facilitate neural network feature learning from unlabeled 3D images, aiming to reduce the required cost for expert annotation. The developed algorithms are 3D Contrastive Predictive Coding, 3D Rotation prediction, 3D Jigsaw puzzles, Relative 3D patch location, and 3D Exemplar networks. Our experiments show that pretraining models with our 3D tasks yields more powerful semantic representations, and enables solving downstream tasks more accurately and efficiently, compared to training the models from scratch and to pretraining them on 2D slices. We demonstrate the effectiveness of our methods on three downstream tasks from the medical imaging domain: i) Brain Tumor Segmentation from 3D MRI, ii) Pancreas Tumor Segmentation from 3D CT, and iii) Diabetic Retinopathy Detection from 2D Fundus images. In each task, we assess the gains in data-efficiency, performance, and speed of convergence. Interestingly, we also find gains when transferring the learned representations, by our methods, from a large unlabeled 3D corpus to a small downstream-specific dataset. We achieve results competitive to state-of-the-art solutions at a fraction of the computational expense. We publish our implementations for the developed algorithms (both 3D and 2D versions) as an open-source library, in an effort to allow other researchers to apply and extend our methods on their datasets.
We proposed a novel machine learning framework to conduct real-time multi-speaker diarization and recognition without prior registration and pretraining in a fully online learning setting. Our contributions are two-fold. First, we proposed a new benchmark to evaluate the rarely studied fully online speaker diarization problem. We built upon existing datasets of real world utterances to automatically curate MiniVox, an experimental environment which generates infinite configurations of continuous multi-speaker speech stream. Second, we considered the practical problem of online learning with episodically revealed rewards and introduced a solution based on semi-supervised and self-supervised learning methods. Additionally, we provided a workable web-based recognition system which interactively handles the cold start problem of new user's addition by transferring representations of old arms to new ones with an extendable contextual bandit. We demonstrated that our proposed method obtained robust performance in the online MiniVox framework.
Generalising deep networks to novel domains without manual labels is challenging to deep learning. This problem is intrinsically difficult due to unpredictable changing nature of imagery data distributions in novel domains. Pre-learned knowledge does not transfer well without making strong assumptions about the learned and the novel domains. Different methods have been studied to address the underlying problem based on different assumptions, e.g. from domain adaptation to zero-shot and few-shot learning. In this work, we address this problem by transfer clustering that aims to learn a discriminative latent space of the unlabelled target data in a novel domain by knowledge transfer from labelled related domains. Specifically, we want to leverage relative (pairwise) imagery information, which is freely available and intrinsic to a target domain, to model the target domain image distribution characteristics as well as the prior-knowledge learned from related labelled domains to enable more discriminative clustering of unlabelled target data. Our method mitigates nontransferrable prior-knowledge by self-supervision, benefiting from both transfer and self-supervised learning. Extensive experiments on four datasets for image clustering tasks reveal the superiority of our model over the state-of-the-art transfer clustering techniques. We further demonstrate its competitive transferability on four zero-shot learning benchmarks.
High-performance anti-spoofing models for automatic speaker verification (ASV), have been widely used to protect ASV by identifying and filtering spoofing audio that is deliberately generated by text-to-speech, voice conversion, audio replay, etc. However, it has been shown that high-performance anti-spoofing models are vulnerable to adversarial attacks. Adversarial attacks, that are indistinguishable from original data but result in the incorrect predictions, are dangerous for anti-spoofing models and not in dispute we should detect them at any cost. To explore this issue, we proposed to employ Mockingjay, a self-supervised learning based model, to protect anti-spoofing models against adversarial attacks in the black-box scenario. Self-supervised learning models are effective in improving downstream task performance like phone classification or ASR. However, their effect in defense for adversarial attacks has not been explored yet. In this work, we explore the robustness of self-supervised learned high-level representations by using them in the defense against adversarial attacks. A layerwise noise to signal ratio (LNSR) is proposed to quantize and measure the effectiveness of deep models in countering adversarial noise. Experimental results on the ASVspoof 2019 dataset demonstrate that high-level representations extracted by Mockingjay can prevent the transferability of adversarial examples, and successfully counter black-box attacks.
Probabilistic Latent Variable Models (LVMs) provide an alternative to self-supervised learning approaches for linguistic representation learning from speech. LVMs admit an intuitive probabilistic interpretation where the latent structure shapes the information extracted from the signal. Even though LVMs have recently seen a renewed interest due to the introduction of Variational Autoencoders (VAEs), their use for speech representation learning remains largely unexplored. In this work, we propose Convolutional Deep Markov Model (ConvDMM), a Gaussian state-space model with non-linear emission and transition functions modelled by deep neural networks. This unsupervised model is trained using black box variational inference. A deep convolutional neural network is used as an inference network for structured variational approximation. When trained on a large scale speech dataset (LibriSpeech), ConvDMM produces features that significantly outperform multiple self-supervised feature extracting methods on linear phone classification and recognition on the Wall Street Journal dataset. Furthermore, we found that ConvDMM complements self-supervised methods like Wav2Vec and PASE, improving on the results achieved with any of the methods alone. Lastly, we find that ConvDMM features enable learning better phone recognizers than any other features in an extreme low-resource regime with few labeled training examples.
Single-View depth estimation using the CNNs trained from unlabelled videos has shown significant promise. However, excellent results have mostly been obtained in street-scene driving scenarios, and such methods often fail in other settings, particularly indoor videos taken by handheld devices. In this work, we establish that the complex ego-motions exhibited in handheld settings are a critical obstacle for learning depth. Our fundamental analysis suggests that the rotation behaves as noise during training, as opposed to the translation (baseline) which provides supervision signals. To address the challenge, we propose a data pre-processing method that rectifies training images by removing their relative rotations for effective learning. The significantly improved performance validates our motivation. Towards end-to-end learning without requiring pre-processing, we propose an Auto-Rectify Network with novel loss functions, which can automatically learn to rectify images during training. Consequently, our results outperform the previous unsupervised SOTA method by a large margin on the challenging NYUv2 dataset. We also demonstrate the generalization of our trained model in ScanNet and Make3D, and the universality of our proposed learning method on 7-Scenes and KITTI datasets.
Graph Convolutional Networks (GCNs) have been successfully applied to analyze non-grid data, where the classical convolutional neural networks (CNNs) cannot be directly used. One similarity shared by GCNs and CNNs is the requirement of massive amount of labeled data for network training. In addition, GCNs need the adjacency matrix as input to define the relationship between those non-grid data, which leads to all of data including training, validation and test data typically forms only one graph structures data for training. Furthermore, the adjacency matrix is usually pre-defined and stationary, which makes the data augmentation strategies cannot be employed on the constructed graph structures data to augment the amount of training data. To further improve the learning capacity and model performance under the limited training data, in this paper, we propose two types of self-supervised learning strategies to exploit available information from the input graph structure data itself. Our proposed self-supervised learning strategies are examined on two representative GCN models with three public citation network datasets - Citeseer, Cora and Pubmed. The experimental results demonstrate the generalization ability as well as the portability of our proposed strategies, which can significantly improve the performance of GCNs with the power of self-supervised learning in improving feature learning.
Pretrained language models such as BERT, GPT have shown great effectiveness in language understanding. The auxiliary predictive tasks in existing pretraining approaches are mostly defined on tokens, thus may not be able to capture sentence-level semantics very well. To address this issue, we propose CERT: Contrastive self-supervised Encoder Representations from Transformers, which pretrains language representation models using contrastive self-supervised learning at the sentence level. CERT creates augmentations of original sentences using back-translation. Then it finetunes a pretrained language encoder (e.g., BERT) by predicting whether two augmented sentences originate from the same sentence. CERT is simple to use and can be flexibly plugged into any pretraining-finetuning NLP pipeline. We evaluate CERT on 11 natural language understanding tasks in the GLUE benchmark where CERT outperforms BERT on 7 tasks, achieves the same performance as BERT on 2 tasks, and performs worse than BERT on 2 tasks. On the averaged score of the 11 tasks, CERT outperforms BERT. The data and code are available at https://github.com/UCSD-AI4H/CERT
We consider the problem of learning to repair programs from diagnostic feedback (e.g., compiler error messages). Program repair is challenging for two reasons: First, it requires reasoning and tracking symbols across source code and diagnostic feedback. Second, labeled datasets available for program repair are relatively small. In this work, we propose novel solutions to these two challenges. First, we introduce a program-feedback graph, which connects symbols relevant to program repair in source code and diagnostic feedback, and then apply a graph neural network on top to model the reasoning process. Second, we present a self-supervised learning paradigm for program repair that leverages unlabeled programs available online to create a large amount of extra program repair examples, which we use to pre-train our models. We evaluate our proposed approach on two applications: correcting introductory programming assignments (DeepFix dataset) and correcting the outputs of program synthesis (SPoC dataset). Our final system, DrRepair, significantly outperforms prior work, achieving 68.2% full repair rate on DeepFix (+22.9% over the prior best), and 48.4% synthesis success rate on SPoC (+3.7% over the prior best).
Previous researches of sketches often considered sketches in pixel format and leveraged CNN based models in the sketch understanding. Fundamentally, a sketch is stored as a sequence of data points, a vector format representation, rather than the photo-realistic image of pixels. SketchRNN studied a generative neural representation for sketches of vector format by Long Short Term Memory networks (LSTM). Unfortunately, the representation learned by SketchRNN is primarily for the generation tasks, rather than the other tasks of recognition and retrieval of sketches. To this end and inspired by the recent BERT model, we present a model of learning Sketch Bidirectional Encoder Representation from Transformer (Sketch-BERT). We generalize BERT to sketch domain, with the novel proposed components and pre-training algorithms, including the newly designed sketch embedding networks, and the self-supervised learning of sketch gestalt. Particularly, towards the pre-training task, we present a novel Sketch Gestalt Model (SGM) to help train the Sketch-BERT. Experimentally, we show that the learned representation of Sketch-BERT can help and improve the performance of the downstream tasks of sketch recognition, sketch retrieval, and sketch gestalt.
For self-supervised speech processing, it is crucial to use pretrained models as speech representation extractors. In recent works, increasing the size of the model has been utilized in acoustic model training in order to achieve better performance. In this paper, we propose Audio ALBERT, a lite version of the self-supervised speech representation model. We use the representations with two downstream tasks, speaker identification, and phoneme classification. We show that Audio ALBERT is capable of achieving competitive performance with those huge models in the downstream tasks while utilizing 91\% fewer parameters. Moreover, we use some simple probing models to measure how much the information of the speaker and phoneme is encoded in latent representations. In probing experiments, we find that the latent representations encode richer information of both phoneme and speaker than that of the last layer.
Previous methods on estimating detailed human depth often require supervised training with `ground truth' depth data. This paper presents a self-supervised method that can be trained on YouTube videos without known depth, which makes training data collection simple and improves the generalization of the learned network. The self-supervised learning is achieved by minimizing a photo-consistency loss, which is evaluated between a video frame and its neighboring frames warped according to the estimated depth and the 3D non-rigid motion of the human body. To solve this non-rigid motion, we first estimate a rough SMPL model at each video frame and compute the non-rigid body motion accordingly, which enables self-supervised learning on estimating the shape details. Experiments demonstrate that our method enjoys better generalization and performs much better on data in the wild.
Whereas adversarial training is employed as the main defence strategy against specific adversarial samples, it has limited generalization capability and incurs excessive time complexity. In this paper, we propose an attack-agnostic defence framework to enhance the intrinsic robustness of neural networks, without jeopardizing the ability of generalizing clean samples. Our Feature Pyramid Decoder (FPD) framework applies to all block-based convolutional neural networks (CNNs). It implants denoising and image restoration modules into a targeted CNN, and it also constraints the Lipschitz constant of the classification layer. Moreover, we propose a two-phase strategy to train the FPD-enhanced CNN, utilizing $\epsilon$-neighbourhood noisy images with multi-task and self-supervised learning. Evaluated against a variety of white-box and black-box attacks, we demonstrate that FPD-enhanced CNNs gain sufficient robustness against general adversarial samples on MNIST, SVHN and CALTECH. In addition, if we further conduct adversarial training, the FPD-enhanced CNNs perform better than their non-enhanced versions.
We propose a self-supervised method to solve Pronoun Disambiguation and Winograd Schema Challenge problems. Our approach exploits the characteristic structure of training corpora related to so-called "trigger" words, which are responsible for flipping the answer in pronoun disambiguation. We achieve such commonsense reasoning by constructing pair-wise contrastive auxiliary predictions. To this end, we leverage a mutual exclusive loss regularized by a contrastive margin. Our architecture is based on the recently introduced transformer networks, BERT, that exhibits strong performance on many NLP benchmarks. Empirical results show that our method alleviates the limitation of current supervised approaches for commonsense reasoning. This study opens up avenues for exploiting inexpensive self-supervision to achieve performance gain in commonsense reasoning tasks.
Self-supervised learning has attracted plenty of recent research interest. However, most works for self-supervision in speech are typically unimodal and there has been limited work that studies the interaction between audio and visual modalities for cross-modal self-supervision. This work (1) investigates visual self-supervision via face reconstruction to guide the learning of audio representations; (2) proposes an audio-only self-supervision approach for speech representation learning; (3) shows that a multi-task combination of the proposed visual and audio self-supervision is beneficial for learning richer features that are more robust in noisy conditions; (4) shows that self-supervised pretraining can outperform fully supervised training and is especially useful to prevent overfitting on smaller sized datasets. We evaluate our learned audio representations for discrete emotion recognition, continuous affect recognition and automatic speech recognition. We outperform existing self-supervised methods for all tested downstream tasks. Our results demonstrate the potential of visual self-supervision for audio feature learning and suggest that joint visual and audio self-supervision leads to more informative audio representations for speech and emotion recognition.
The goal of this work is to train discriminative cross-modal embeddings without access to manually annotated data. Recent advances in self-supervised learning have shown that effective representations can be learnt from natural cross-modal synchrony. We build on earlier work to train embeddings that are more discriminative for uni-modal downstream tasks. To this end, we propose a novel training strategy that not only optimises metrics across modalities, but also enforces intra-class feature separation within each of the modalities. The effectiveness of the method is demonstrated on two downstream tasks: lip reading using the features trained on audio-visual synchronisation, and speaker recognition using the features trained for cross-modal biometric matching. The proposed method outperforms state-of-the-art self-supervised baselines by a signficant margin.
Transfer learning has yielded state-of-the-art (SoTA) results in many supervised NLP tasks. However, annotated data for every target task in every target language is rare, especially for low-resource languages. We propose UXLA, a novel unsupervised data augmentation framework for zero-resource transfer learning scenarios. In particular, UXLA aims to solve cross-lingual adaptation problems from a source language task distribution to an unknown target language task distribution, assuming no training label in the target language. At its core, UXLA performs simultaneous self-training with data augmentation and unsupervised sample selection. To show its effectiveness, we conduct extensive experiments on three diverse zero-resource cross-lingual transfer tasks. UXLA achieves SoTA results in all the tasks, outperforming the baselines by a good margin. With an in-depth framework dissection, we demonstrate the cumulative contributions of different components to its success.
We present an introspection of an audiovisual speech enhancement model. In particular, we focus on interpreting how a neural audiovisual speech enhancement model uses visual cues to improve the quality of the target speech signal. We show that visual cues provide not only high-level information about speech activity, i.e., speech/silence, but also fine-grained visual information about the place of articulation. One byproduct of this finding is that the learned visual embeddings can be used as features for other visual speech applications. We demonstrate the effectiveness of the learned visual embeddings for classifying visemes (the visual analogy to phonemes). Our results provide insight into important aspects of audiovisual speech enhancement and demonstrate how such models can be used for self-supervision tasks for visual speech applications.
Self-supervised representation learning has achieved impressive results in recent years, with experiments primarily coming on ImageNet or other similarly large internet imagery datasets. There has been little to no work with these methods on other smaller domains, such as satellite, textural, or biological imagery. We experiment with several popular methods on an unprecedented variety of domains. We discover, among other findings, that Rotation is by far the most semantically meaningful task, with much of the performance of Jigsaw and Instance Discrimination being attributable to the nature of their induced distribution rather than semantic understanding. Additionally, there are several areas, such as fine-grain classification, where all tasks underperform. We quantitatively and qualitatively diagnose the reasons for these failures and successes via novel experiments studying pretext generalization, random labelings, and implicit dimensionality. Code and models are available at https://github.com/BramSW/Extending_SSRL_Across_Domains/.
We present a self-supervised learning approach to learn audio-visual representations from video and audio. Our method uses contrastive learning for cross-modal discrimination of video from audio and vice-versa. We show that optimizing for cross-modal discrimination, rather than within-modal discrimination, is important to learn good representations from video and audio. With this simple but powerful insight, our method achieves highly competitive performance when finetuned on action recognition tasks. Furthermore, while recent work in contrastive learning defines positive and negative samples as individual instances, we generalize this definition by exploring cross-modal agreement. We group together multiple instances as positives by measuring their similarity in both the video and audio feature spaces. Cross-modal agreement creates better positive and negative sets, which allows us to calibrate visual similarities by seeking within-modal discrimination of positive instances, and achieve significant gains on downstream tasks.