Due to the constraints of the imaging device and high cost in operation time, computer tomography (CT) scans are usually acquired with low intra-slice resolution. Improving the intra-slice resolution is beneficial to the disease diagnosis for both human experts and computer-aided systems. To this end, this paper builds a novel medical slice synthesis to increase the between-slice resolution. Considering that the ground-truth intermediate medical slices are always absent in clinical practice, we introduce the incremental cross-view mutual distillation strategy to accomplish this task in the self-supervised learning manner. Specifically, we model this problem from three different views: slice-wise interpolation from axial view and pixel-wise interpolation from coronal and sagittal views. Under this circumstance, the models learned from different views can distill valuable knowledge to guide the learning processes of each other. We can repeat this process to make the models synthesize intermediate slice data with increasing inter-slice resolution. To demonstrate the effectiveness of the proposed approach, we conduct comprehensive experiments on a large-scale CT dataset. Quantitative and qualitative comparison results show that our method outperforms state-of-the-art algorithms by clear margins.
Effective representation learning is the key in improving model performance for medical image analysis. In training deep learning models, a compromise often must be made between performance and trust, both of which are essential for medical applications. Moreover, models optimized with cross-entropy loss tend to suffer from unwarranted overconfidence in the majority class and over-cautiousness in the minority class. In this work, we integrate a new surrogate loss with self-supervised learning for computer-aided screening of COVID-19 patients using radiography images. In addition, we adopt a new quantification score to measure a model's trustworthiness. Ablation study is conducted for both the performance and the trust on feature learning methods and loss functions. Comparisons show that leveraging the new surrogate loss on self-supervised models can produce label-efficient networks that are both high-performing and trustworthy.
Contrastive learning has become a key component of self-supervised learning approaches for graph-structured data. Despite their success, existing graph contrastive learning methods are incapable of uncertainty quantification for node representations or their downstream tasks, limiting their application in high-stakes domains. In this paper, we propose a novel Bayesian perspective of graph contrastive learning methods showing random augmentations leads to stochastic encoders. As a result, our proposed method represents each node by a distribution in the latent space in contrast to existing techniques which embed each node to a deterministic vector. By learning distributional representations, we provide uncertainty estimates in downstream graph analytics tasks and increase the expressive power of the predictive model. In addition, we propose a Bayesian framework to infer the probability of perturbations in each view of the contrastive model, eliminating the need for a computationally expensive search for hyperparameter tuning. We empirically show a considerable improvement in performance compared to existing state-of-the-art methods on several benchmark datasets.
Improving generalization is a major challenge in audio classification due to labeled data scarcity. Self-supervised learning (SSL) methods tackle this by leveraging unlabeled data to learn useful features for downstream classification tasks. In this work, we propose an augmented contrastive SSL framework to learn invariant representations from unlabeled data. Our method applies various perturbations to the unlabeled input data and utilizes contrastive learning to learn representations robust to such perturbations. Experimental results on the Audioset and DESED datasets show that our framework significantly outperforms state-of-the-art SSL and supervised learning methods on sound/event classification tasks.
Recently, pioneer work finds that speech pre-trained models can solve full-stack speech processing tasks, because the model utilizes bottom layers to learn speaker-related information and top layers to encode content-related information. Since the network capacity is limited, we believe the speech recognition performance could be further improved if the model is dedicated to audio content information learning. To this end, we propose Intermediate Layer Supervision for Self-Supervised Learning (ILS-SSL), which forces the model to concentrate on content information as much as possible by adding an additional SSL loss on the intermediate layers. Experiments on LibriSpeech test-other set show that our method outperforms HuBERT significantly, which achieves a 23.5%/11.6% relative word error rate reduction in the w/o language model setting for base/large models. Detailed analysis shows the bottom layers of our model have a better correlation with phonetic units, which is consistent with our intuition and explains the success of our method for ASR.
Web-based interactions can be frequently represented by an attributed graph, and node clustering in such graphs has received much attention lately. Multiple efforts have successfully applied Graph Convolutional Networks (GCN), though with some limits on accuracy as GCNs have been shown to suffer from over-smoothing issues. Though other methods (particularly those based on Laplacian Smoothing) have reported better accuracy, a fundamental limitation of all the work is a lack of scalability. This paper addresses this open problem by relating the Laplacian smoothing to the Generalized PageRank and applying a random-walk based algorithm as a scalable graph filter. This forms the basis for our scalable deep clustering algorithm, RwSL, where through a self-supervised mini-batch training mechanism, we simultaneously optimize a deep neural network for sample-cluster assignment distribution and an autoencoder for a clustering-oriented embedding. Using 6 real-world datasets and 6 clustering metrics, we show that RwSL achieved improved results over several recent baselines. Most notably, we show that RwSL, unlike all other deep clustering frameworks, can continue to scale beyond graphs with more than one million nodes, i.e., handle web-scale. We also demonstrate how RwSL could perform node clustering on a graph with 1.8 billion edges using only a single GPU.
Recent self-supervised learning methods are able to learn high-quality image representations and are closing the gap with supervised approaches. However, these methods are unable to acquire new knowledge incrementally -- they are, in fact, mostly used only as a pre-training phase over IID data. In this work we investigate self-supervised methods in continual learning regimes without any replay mechanism. We show that naive functional regularization, also known as feature distillation, leads to lower plasticity and limits continual learning performance. Instead, we propose Projected Functional Regularization in which a separate temporal projection network ensures that the newly learned feature space preserves information of the previous one, while at the same time allowing for the learning of new features. This prevents forgetting while maintaining the plasticity of the learner. Comparison with other incremental learning approaches applied to self-supervision demonstrates that our method obtains competitive performance in different scenarios and on multiple datasets.
This paper advocates the use of implicit surface representation in autoencoder-based self-supervised 3D representation learning. The most popular and accessible 3D representation, i.e., point clouds, involves discrete samples of the underlying continuous 3D surface. This discretization process introduces sampling variations on the 3D shape, making it challenging to develop transferable knowledge of the true 3D geometry. In the standard autoencoding paradigm, the encoder is compelled to encode not only the 3D geometry but also information on the specific discrete sampling of the 3D shape into the latent code. This is because the point cloud reconstructed by the decoder is considered unacceptable unless there is a perfect mapping between the original and the reconstructed point clouds. This paper introduces the Implicit AutoEncoder (IAE), a simple yet effective method that addresses the sampling variation issue by replacing the commonly-used point-cloud decoder with an implicit decoder. The implicit decoder reconstructs a continuous representation of the 3D shape, independent of the imperfections in the discrete samples. Extensive experiments demonstrate that the proposed IAE achieves state-of-the-art performance across various self-supervised learning benchmarks.
Unsupervised anomaly detection aims to build models to effectively detect unseen anomalies by only training on the normal data. Although previous reconstruction-based methods have made fruitful progress, their generalization ability is limited due to two critical challenges. First, the training dataset only contains normal patterns, which limits the model generalization ability. Second, the feature representations learned by existing models often lack representativeness which hampers the ability to preserve the diversity of normal patterns. In this paper, we propose a novel approach called Adaptive Memory Network with Self-supervised Learning (AMSL) to address these challenges and enhance the generalization ability in unsupervised anomaly detection. Based on the convolutional autoencoder structure, AMSL incorporates a self-supervised learning module to learn general normal patterns and an adaptive memory fusion module to learn rich feature representations. Experiments on four public multivariate time series datasets demonstrate that AMSL significantly improves the performance compared to other state-of-the-art methods. Specifically, on the largest CAP sleep stage detection dataset with 900 million samples, AMSL outperforms the second-best baseline by \textbf{4}\%+ in both accuracy and F1 score. Apart from the enhanced generalization ability, AMSL is also more robust against input noise.
Machine learning systems often experience a distribution shift between training and testing. In this paper, we introduce a simple variational objective whose optima are exactly the set of all representations on which risk minimizers are guaranteed to be robust to any distribution shift that preserves the Bayes predictor, e.g., covariate shifts. Our objective has two components. First, a representation must remain discriminative for the task, i.e., some predictor must be able to simultaneously minimize the source and target risk. Second, the representation's marginal support needs to be the same across source and target. We make this practical by designing self-supervised objectives that only use unlabelled data and augmentations to train robust representations. Our objectives give insights into the robustness of CLIP, and further improve CLIP's representations to achieve SOTA results on DomainBed.
Solar energy is now the cheapest form of electricity in history. Unfortunately, significantly increasing the grid's fraction of solar energy remains challenging due to its variability, which makes balancing electricity's supply and demand more difficult. While thermal generators' ramp rate -- the maximum rate that they can change their output -- is finite, solar's ramp rate is essentially infinite. Thus, accurate near-term solar forecasting, or nowcasting, is important to provide advance warning to adjust thermal generator output in response to solar variations to ensure a balanced supply and demand. To address the problem, this paper develops a general model for solar nowcasting from abundant and readily available multispectral satellite data using self-supervised learning. Specifically, we develop deep auto-regressive models using convolutional neural networks (CNN) and long short-term memory networks (LSTM) that are globally trained across multiple locations to predict raw future observations of the spatio-temporal data collected by the recently launched GOES-R series of satellites. Our model estimates a location's future solar irradiance based on satellite observations, which we feed to a regression model trained on smaller site-specific solar data to provide near-term solar photovoltaic (PV) forecasts that account for site-specific characteristics. We evaluate our approach for different coverage areas and forecast horizons across 25 solar sites and show that our approach yields errors close to that of a model using ground-truth observations.
Self-supervised learning has shown very promising results for monocular depth estimation. Scene structure and local details both are significant clues for high-quality depth estimation. Recent works suffer from the lack of explicit modeling of scene structure and proper handling of details information, which leads to a performance bottleneck and blurry artefacts in predicted results. In this paper, we propose the Channel-wise Attention-based Depth Estimation Network (CADepth-Net) with two effective contributions: 1) The structure perception module employs the self-attention mechanism to capture long-range dependencies and aggregates discriminative features in channel dimensions, explicitly enhances the perception of scene structure, obtains the better scene understanding and rich feature representation. 2) The detail emphasis module re-calibrates channel-wise feature maps and selectively emphasizes the informative features, aiming to highlight crucial local details information and fuse different level features more efficiently, resulting in more precise and sharper depth prediction. Furthermore, the extensive experiments validate the effectiveness of our method and show that our model achieves the state-of-the-art results on the KITTI benchmark and Make3D datasets.
Automatic speech recognition (ASR) has shown rapid advances in recent years but still degrades significantly in far-field and noisy environments. The recent development of self-supervised learning (SSL) technology can improve the ASR performance by pre-training the model with additional unlabeled speech and the SSL pre-trained model has achieved the state-of-the-art result on several speech benchmarks. Nevertheless, most of the previous SSL methods ignore the influence of the background noise or reverberation, which is crucial to deploying ASR systems in real-world speech applications. This study addresses the robust ASR by introducing a multi-variant consistency (MVC) based SSL method that adapts to different environments. The MVC-SSL is a robust SSL pre-training method designed for noisy and distant-talking speech in real-world applications. Compared to the previous SSL method, the MVC-SSL can calculate the contrastive loss among audios from different acoustic conditions or channels and can learn invariant representations with the change in the environment or the recording equipment. We also explore different SSL training pipelines to balance the noisy distant-talking speech and extra high resource clean speech. We evaluate the proposed method on the commercially-motivated dataset, CHiME-4, and the meeting dataset, AMI. With the help of the MVC-SSL and appropriate training pipeline, we can achieve up to 30% relative word error rate reductions over the baseline wav2vec2.0, one of the most successful SSL methods for ASR.
Machine learning is gaining growing momentum in various recent models for the dynamic analysis of information flows in data communications networks. These preliminary models often rely on off-the-shelf learning models to predict from historical statistics while disregarding the physics governing the generating behaviors of these flows. This paper instead introduces Flow Neural Network (FlowNN) to improve the feature representation with learned physical bias. This is implemented by an induction layer, working upon the embedding layer, to impose the physics connected data correlations, and a self-supervised learning strategy with stop-gradient to make the learned physics universal. For the short-timescale network prediction tasks, FlowNN achieves 17% - 71% of loss decrease than the state-of-the-art baselines on both synthetic and real-world networking datasets, which shows the strength of this new approach.
State-of-the-art deep learning approaches for skin lesion recognition often require pretraining on larger and more varied datasets, to overcome the generalization limitations derived from the reduced size of the skin lesion imaging datasets. ImageNet is often used as the pretraining dataset, but its transferring potential is hindered by the domain gap between the source dataset and the target dermatoscopic scenario. In this work, we introduce a novel pretraining approach that sequentially trains a series of Self-Supervised Learning pretext tasks and only requires the unlabeled skin lesion imaging data. We present a simple methodology to establish an ordering that defines a pretext task curriculum. For the multi-class skin lesion classification problem, and ISIC-2019 dataset, we provide experimental evidence showing that: i) a model pretrained by a curriculum of pretext tasks outperforms models pretrained by individual pretext tasks, and ii) a model pretrained by the optimal pretext task curriculum outperforms a model pretrained on ImageNet. We demonstrate that this performance gain is related to the fact that the curriculum of pretext tasks better focuses the attention of the final model on the skin lesion. Beyond performance improvement, this strategy allows for a large reduction in the training time with respect to ImageNet pretraining, which is especially advantageous for network architectures tailored for a specific problem.
Recent work has shown that self-supervised pre-training leads to improvements over supervised learning on challenging visual recognition tasks. CLIP, an exciting new approach to learning with language supervision, demonstrates promising performance on a wide variety of benchmarks. In this work, we explore whether self-supervised learning can aid in the use of language supervision for visual representation learning. We introduce SLIP, a multi-task learning framework for combining self-supervised learning and CLIP pre-training. After pre-training with Vision Transformers, we thoroughly evaluate representation quality and compare performance to both CLIP and self-supervised learning under three distinct settings: zero-shot transfer, linear classification, and end-to-end finetuning. Across ImageNet and a battery of additional datasets, we find that SLIP improves accuracy by a large margin. We validate our results further with experiments on different model sizes, training schedules, and pre-training datasets. Our findings show that SLIP enjoys the best of both worlds: better performance than self-supervision (+8.1% linear accuracy) and language supervision (+5.2% zero-shot accuracy).
Unsupervised graph representation learning has recently gained interest in several application domains such as neuroscience, where modeling the diverse morphology of cell types in the brain is one of the key challenges. It is currently unknown how many excitatory cortical cell types exist and what their defining morphological features are. Here we present GraphDINO, a purely data-driven approach to learn low-dimensional representations of 3D neuronal morphologies from unlabeled large-scale datasets. GraphDINO is a novel transformer-based representation learning method for spatially-embedded graphs. To enable self-supervised learning on transformers, we (1) developed data augmentation strategies for spatially-embedded graphs, (2) adapted the positional encoding and (3) introduced a novel attention mechanism, AC-Attention, which combines attention-based global interaction between nodes and classic graph convolutional processing. We show, in two different species and across multiple brain areas, that this method yields morphological cell type clusterings that are on par with manual feature-based classification by experts, but without using prior knowledge about the structural features of neurons. Moreover, it outperforms previous approaches on quantitative benchmarks predicting expert labels. Our method could potentially enable data-driven discovery of novel morphological features and cell types in large-scale datasets. It is applicable beyond neuroscience in settings where samples in a dataset are graphs and graph-level embeddings are desired.
Measures of Activity of Daily Living (ADL) are an important indicator of overall health but difficult to measure in-clinic. Automated and accurate human activity recognition (HAR) using wrist-worn accelerometers enables practical and cost efficient remote monitoring of ADL. Key obstacles in developing high quality HAR is the lack of large labeled datasets and the performance loss when applying models trained on small curated datasets to the continuous stream of heterogeneous data in real-life. In this work we design a self-supervised learning paradigm to create a robust representation of accelerometer data that can generalize across devices and subjects. We demonstrate that this representation can separate activities of daily living and achieve strong HAR accuracy (on multiple benchmark datasets) using very few labels. We also propose a segmentation algorithm which can identify segments of salient activity and boost HAR accuracy on continuous real-life data.
Learnable keypoint detectors and descriptors are beginning to outperform classical hand-crafted feature extraction methods. Recent studies on self-supervised learning of visual representations have driven the increasing performance of learnable models based on deep networks. By leveraging traditional data augmentations and homography transformations, these networks learn to detect corners under adverse conditions such as extreme illumination changes. However, their generalization capabilities are limited to corner-like features detected a priori by classical methods or synthetically generated data. In this paper, we propose the Correspondence Network (CorrNet) that learns to detect repeatable keypoints and to extract discriminative descriptions via unsupervised contrastive learning under spatial constraints. Our experiments show that CorrNet is not only able to detect low-level features such as corners, but also high-level features that represent similar objects present in a pair of input images through our proposed joint guided backpropagation of their latent space. Our approach obtains competitive results under viewpoint changes and achieves state-of-the-art performance under illumination changes.
Recent advances in deep learning, in particular enabled by hardware advances and big data, have provided impressive results across a wide range of computational problems such as computer vision, natural language, or reinforcement learning. Many of these improvements are however constrained to problems with large-scale curated data-sets which require a lot of human labor to gather. Additionally, these models tend to generalize poorly under both slight distributional shifts and low-data regimes. In recent years, emerging fields such as meta-learning or self-supervised learning have been closing the gap between proof-of-concept results and real-life applications of machine learning by extending deep-learning to the semi-supervised and few-shot domains. We follow this line of work and explore spatio-temporal structure in a recently introduced image-to-image translation problem in order to: i) formulate a novel multi-task few-shot image generation benchmark and ii) explore data augmentations in contrastive pre-training for image translation downstream tasks. We present several baselines for the few-shot problem and discuss trade-offs between different approaches. Our code is available at https://github.com/irugina/meta-image-translation.