An important challenge in emotion recognition is to develop methods that can leverage unlabeled training data. In this paper, we propose the VQ-MAE-AV model, a self-supervised multimodal model that leverages masked autoencoders to learn representations of audiovisual speech without labels. The model includes vector quantized variational autoencoders that compress raw audio and visual speech data into discrete tokens. The audiovisual speech tokens are used to train a multimodal masked autoencoder that consists of an encoder-decoder architecture with attention mechanisms. The model is designed to extract both local (i.e., at the frame level) and global (i.e., at the sequence level) representations of audiovisual speech. During self-supervised pre-training, the VQ-MAE-AV model is trained on a large-scale unlabeled dataset of audiovisual speech, for the task of reconstructing randomly masked audiovisual speech tokens and with a contrastive learning strategy. During this pre-training, the encoder learns to extract a representation of audiovisual speech that can be subsequently leveraged for emotion recognition. During the supervised fine-tuning stage, a small classification model is trained on top of the VQ-MAE-AV encoder for an emotion recognition task. The proposed approach achieves state-of-the-art emotion recognition results across several datasets in both controlled and in-the-wild conditions.
The purpose of this study is to develop a computer-aided diagnosis system for classifying benign and malignant lung lesions, and to assist physicians in real-time analysis of radial probe endobronchial ultrasound (EBUS) videos. During the biopsy process of lung cancer, physicians use real-time ultrasound images to find suitable lesion locations for sampling. However, most of these images are difficult to classify and contain a lot of noise. Previous studies have employed 2D convolutional neural networks to effectively differentiate between benign and malignant lung lesions, but doctors still need to manually select good-quality images, which can result in additional labor costs. In addition, the 2D neural network has no ability to capture the temporal information of the ultrasound video, so it is difficult to obtain the relationship between the features of the continuous images. This study designs an automatic diagnosis system based on a 3D neural network, uses the SlowFast architecture as the backbone to fuse temporal and spatial features, and uses the SwAV method of contrastive learning to enhance the noise robustness of the model. The method we propose includes the following advantages, such as (1) using clinical ultrasound films as model input, thereby reducing the need for high-quality image selection by physicians, (2) high-accuracy classification of benign and malignant lung lesions can assist doctors in clinical diagnosis and reduce the time and risk of surgery, and (3) the capability to classify well even in the presence of significant image noise. The AUC, accuracy, precision, recall and specificity of our proposed method on the validation set reached 0.87, 83.87%, 86.96%, 90.91% and 66.67%, respectively. The results have verified the importance of incorporating temporal information and the effectiveness of using the method of contrastive learning on feature extraction.
In this study, our goal is to show the impact of self-supervised pre-training of transformers for organ at risk (OAR) and tumor segmentation as compared to costly fully-supervised learning. The proposed algorithm is called Monte Carlo Transformer based U-Net (MC-Swin-U). Unlike many other available models, our approach presents uncertainty quantification with Monte Carlo dropout strategy while generating its voxel-wise prediction. We test and validate the proposed model on both public and one private datasets and evaluate the gross tumor volume (GTV) as well as nearby risky organs' boundaries. We show that self-supervised pre-training approach improves the segmentation scores significantly while providing additional benefits for avoiding large-scale annotation costs.
We introduce Masked Trajectory Models (MTM) as a generic abstraction for sequential decision making. MTM takes a trajectory, such as a state-action sequence, and aims to reconstruct the trajectory conditioned on random subsets of the same trajectory. By training with a highly randomized masking pattern, MTM learns versatile networks that can take on different roles or capabilities, by simply choosing appropriate masks at inference time. For example, the same MTM network can be used as a forward dynamics model, inverse dynamics model, or even an offline RL agent. Through extensive experiments in several continuous control tasks, we show that the same MTM network -- i.e. same weights -- can match or outperform specialized networks trained for the aforementioned capabilities. Additionally, we find that state representations learned by MTM can significantly accelerate the learning speed of traditional RL algorithms. Finally, in offline RL benchmarks, we find that MTM is competitive with specialized offline RL algorithms, despite MTM being a generic self-supervised learning method without any explicit RL components. Code is available at https://github.com/facebookresearch/mtm
Modelling of early language acquisition aims to understand how infants bootstrap their language skills. The modelling encompasses properties of the input data used for training the models, the cognitive hypotheses and their algorithmic implementations being tested, and the evaluation methodologies to compare models to human data. Recent developments have enabled the use of more naturalistic training data for computational models. This also motivates development of more naturalistic tests of model behaviour. A crucial step towards such an aim is to develop representative speech datasets consisting of speech heard by infants in their natural environments. However, a major drawback of such recordings is that they are typically noisy, and it is currently unclear how the sound quality could affect analyses and modelling experiments conducted on such data. In this paper, we explore this aspect for the case of infant-directed speech (IDS) and adult-directed speech (ADS) analysis. First, we manually and automatically annotated audio quality of utterances extracted from two corpora of child-centred long-form recordings (in English and French). We then compared acoustic features of IDS and ADS in an in-lab dataset and across different audio quality subsets of naturalistic data. Finally, we assessed how the audio quality and recording environment may change the conclusions of a modelling analysis using a recent self-supervised learning model. Our results show that the use of modest and high audio quality naturalistic speech data result in largely similar conclusions on IDS and ADS in terms of acoustic analyses and modelling experiments. We also found that an automatic sound quality assessment tool can be used to screen out useful parts of long-form recordings for a closer analysis with comparable results to that of manual quality annotation.
In this paper, we explore self-supervised learning (SSL) for analyzing a first-of-its-kind database of cry recordings containing clinical indications of more than a thousand newborns. Specifically, we target cry-based detection of neurological injury as well as identification of cry triggers such as pain, hunger, and discomfort. Annotating a large database in the medical setting is expensive and time-consuming, typically requiring the collaboration of several experts over years. Leveraging large amounts of unlabeled audio data to learn useful representations can lower the cost of building robust models and, ultimately, clinical solutions. In this work, we experiment with self-supervised pre-training of a convolutional neural network on large audio datasets. We show that pre-training with SSL contrastive loss (SimCLR) performs significantly better than supervised pre-training for both neuro injury and cry triggers. In addition, we demonstrate further performance gains through SSL-based domain adaptation using unlabeled infant cries. We also show that using such SSL-based pre-training for adaptation to cry sounds decreases the need for labeled data of the overall system.
Geo-tagged images are publicly available in large quantities, whereas labels such as object classes are rather scarce and expensive to collect. Meanwhile, contrastive learning has achieved tremendous success in various natural image and language tasks with limited labeled data. However, existing methods fail to fully leverage geospatial information, which can be paramount to distinguishing objects that are visually similar. To directly leverage the abundant geospatial information associated with images in pre-training, fine-tuning, and inference stages, we present Contrastive Spatial Pre-Training (CSP), a self-supervised learning framework for geo-tagged images. We use a dual-encoder to separately encode the images and their corresponding geo-locations, and use contrastive objectives to learn effective location representations from images, which can be transferred to downstream supervised tasks such as image classification. Experiments show that CSP can improve model performance on both iNat2018 and fMoW datasets. Especially, on iNat2018, CSP significantly boosts the model performance with 10-34% relative improvement with various labeled training data sampling ratios.
Recently, self-supervised learning (SSL) from unlabelled speech data has gained increased attention in the automatic speech recognition (ASR) community. Typical SSL methods include autoregressive predictive coding (APC), Wav2vec2.0, and hidden unit BERT (HuBERT). However, SSL models are biased to the pretraining data. When SSL models are finetuned with data from another domain, domain shifting occurs and might cause limited knowledge transfer for downstream tasks. In this paper, we propose a novel framework, domain responsible adaptation and finetuning (DRAFT), to reduce domain shifting in pretrained speech models, and evaluate it for a causal and non-causal transformer. For the causal transformer, an extension of APC (E-APC) is proposed to learn richer information from unlabelled data by using multiple temporally-shifted sequences to perform prediction. For the non-causal transformer, various solutions for using the bidirectional APC (Bi-APC) are investigated. In addition, the DRAFT framework is examined for Wav2vec2.0 and HuBERT methods, which use non-causal transformers as the backbone. The experiments are conducted on child ASR (using the OGI and MyST databases) using SSL models trained with unlabelled adult speech data from Librispeech. The relative WER improvements of up to 19.7% on the two child tasks are observed when compared to the pretrained models without adaptation. With the proposed methods (E-APC and DRAFT), the relative WER improvements are even larger (30% and 19% on the OGI and MyST data, respectively) when compared to the models without using pretraining methods.
In the context of mobile sensing environments, various sensors on mobile devices continually generate a vast amount of data. Analyzing this ever-increasing data presents several challenges, including limited access to annotated data and a constantly changing environment. Recent advancements in self-supervised learning have been utilized as a pre-training step to enhance the performance of conventional supervised models to address the absence of labelled datasets. This research examines the impact of using a self-supervised representation learning model for time series classification tasks in which data is incrementally available. We proposed and evaluated a workflow in which a model learns to extract informative features using a corpus of unlabeled time series data and then conducts classification on labelled data using features extracted by the model. We analyzed the effect of varying the size, distribution, and source of the unlabeled data on the final classification performance across four public datasets, including various types of sensors in diverse applications.
This paper presents an unsupervised domain adaptation (UDA) method for predicting unlabeled target domain data, specific to complex UDA tasks where the domain gap is significant. Mainstream UDA models aim to learn from both domains and improve target discrimination by utilizing labeled source domain data. However, the performance boost may be limited when the discrepancy between the source and target domains is large or the target domain contains outliers. To explicitly address this issue, we propose the Adversarial self-superVised domain Adaptation network for the TARget domain (AVATAR) algorithm. It outperforms state-of-the-art UDA models by concurrently reducing domain discrepancy while enhancing discrimination through domain adversarial learning, self-supervised learning, and sample selection strategy for the target domain, all guided by deep clustering. Our proposed model significantly outperforms state-of-the-art methods on three UDA benchmarks, and extensive ablation studies and experiments demonstrate the effectiveness of our approach for addressing complex UDA tasks.
Industrial image anomaly detection under the setting of one-class classification has significant practical value. However, most existing models struggle to extract separable feature representations when performing feature embedding and struggle to build compact descriptions of normal features when performing one-class classification. One direct consequence of this is that most models perform poorly in detecting logical anomalies which violate contextual relationships. Focusing on more effective and comprehensive anomaly detection, we propose a network based on self-supervised learning and self-attentive graph convolution (SLSG) for anomaly detection. SLSG uses a generative pre-training network to assist the encoder in learning the embedding of normal patterns and the reasoning of position relationships. Subsequently, SLSG introduces the pseudo-prior knowledge of anomaly through simulated abnormal samples. By comparing the simulated anomalies, SLSG can better summarize the normal features and narrow down the hypersphere used for one-class classification. In addition, with the construction of a more general graph structure, SLSG comprehensively models the dense and sparse relationships among elements in the image, which further strengthens the detection of logical anomalies. Extensive experiments on benchmark datasets show that SLSG achieves superior anomaly detection performance, demonstrating the effectiveness of our method.
Biparametric magnetic resonance imaging (bpMRI) has demonstrated promising results in prostate cancer (PCa) detection using convolutional neural networks (CNNs). Recently, transformers have achieved competitive performance compared to CNNs in computer vision. Large scale transformers need abundant annotated data for training, which are difficult to obtain in medical imaging. Self-supervised learning (SSL) utilizes unlabeled data to generate meaningful semantic representations without the need for costly annotations, enhancing model performance on tasks with limited labeled data. We introduce a novel end-to-end Cross-Shaped windows (CSwin) transformer UNet model, CSwin UNet, to detect clinically significant prostate cancer (csPCa) in prostate bi-parametric MR imaging (bpMRI) and demonstrate the effectiveness of our proposed self-supervised pre-training framework. Using a large prostate bpMRI dataset with 1500 patients, we first pretrain CSwin transformer using multi-task self-supervised learning to improve data-efficiency and network generalizability. We then finetune using lesion annotations to perform csPCa detection. Five-fold cross validation shows that self-supervised CSwin UNet achieves 0.888 AUC and 0.545 Average Precision (AP), significantly outperforming four comparable models (Swin UNETR, DynUNet, Attention UNet, UNet). Using a separate bpMRI dataset with 158 patients, we evaluate our method robustness to external hold-out data. Self-supervised CSwin UNet achieves 0.79 AUC and 0.45 AP, still outperforming all other comparable methods and demonstrating good generalization to external data.
Contrastive learning has recently established itself as a powerful self-supervised learning framework for extracting rich and versatile data representations. Broadly speaking, contrastive learning relies on a data augmentation scheme to generate two versions of the input data and learns low-dimensional representations by maximizing a normalized temperature-scaled cross entropy loss (NT-Xent) to identify augmented samples corresponding to the same original entity. In this paper, we investigate the potential of deploying contrastive learning in combination with Graph Neural Networks for embedding nodes in a graph. Specifically, we show that the quality of the resulting embeddings and training time can be significantly improved by a simple column-wise postprocessing of the embedding matrix, instead of the row-wise postprocessing via multilayer perceptrons (MLPs) that is adopted by the majority of peer methods. This modification yields improvements in downstream classification tasks of up to 1.5% and even beats existing state-of-the-art approaches on 6 out of 8 different benchmarks. We justify our choices of postprocessing by revisiting the "alignment vs. uniformity paradigm", and show that column-wise post-processing improves both "alignment" and "uniformity" of the embeddings.
Humans interact with the environment using a combination of perception - transforming sensory inputs from their environment into symbols, and cognition - mapping symbols to knowledge about the environment for supporting abstraction, reasoning by analogy, and long-term planning. Human perception-inspired machine perception, in the context of AI, refers to large-scale pattern recognition from raw data using neural networks trained using self-supervised learning objectives such as next-word prediction or object recognition. On the other hand, machine cognition encompasses more complex computations, such as using knowledge of the environment to guide reasoning, analogy, and long-term planning. Humans can also control and explain their cognitive functions. This seems to require the retention of symbolic mappings from perception outputs to knowledge about their environment. For example, humans can follow and explain the guidelines and safety constraints driving their decision-making in safety-critical applications such as healthcare, criminal justice, and autonomous driving. This article introduces the rapidly emerging paradigm of Neurosymbolic AI combines neural networks and knowledge-guided symbolic approaches to create more capable and flexible AI systems. These systems have immense potential to advance both algorithm-level (e.g., abstraction, analogy, reasoning) and application-level (e.g., explainable and safety-constrained decision-making) capabilities of AI systems.
Objective technical skill assessment is crucial for effective training of new surgeons in robot-assisted surgery. With advancements in surgical training programs in both physical and virtual environments, it is imperative to develop generalizable methods for automatically assessing skills. In this paper, we propose a novel approach for skill assessment by transferring domain knowledge from labeled kinematic data to unlabeled data. Our approach leverages labeled data from common surgical training tasks such as Suturing, Needle Passing, and Knot Tying to jointly train a model with both labeled and unlabeled data. Pseudo labels are generated for the unlabeled data through an iterative manner that incorporates uncertainty estimation to ensure accurate labeling. We evaluate our method on a virtual reality simulated training task (Ring Transfer) using data from the da Vinci Research Kit (dVRK). The results show that trainees with robotic assistance have significantly higher expert probability compared to these without any assistance, p < 0.05, which aligns with previous studies showing the benefits of robotic assistance in improving training proficiency. Our method offers a significant advantage over other existing works as it does not require manual labeling or prior knowledge of the surgical training task for robot-assisted surgery.
Deep Learning (DL) can diagnose faults and assess machine health from raw condition monitoring data without manually designed statistical features. However, practical manufacturing applications remain extremely difficult for existing DL methods. Machine data is often unlabeled and from very few health conditions (e.g., only normal operating data). Furthermore, models often encounter shifts in domain as process parameters change and new categories of faults emerge. Traditional supervised learning may struggle to learn compact, discriminative representations that generalize to these unseen target domains since it depends on having plentiful classes to partition the feature space with decision boundaries. Transfer Learning (TL) with domain adaptation attempts to adapt these models to unlabeled target domains but assumes similar underlying structure that may not be present if new faults emerge. This study proposes focusing on maximizing the feature generality on the source domain and applying TL via weight transfer to copy the model to the target domain. Specifically, Self-Supervised Learning (SSL) with Barlow Twins may produce more discriminative features for monitoring health condition than supervised learning by focusing on semantic properties of the data. Furthermore, Federated Learning (FL) for distributed training may also improve generalization by efficiently expanding the effective size and diversity of training data by sharing information across multiple client machines. Results show that Barlow Twins outperforms supervised learning in an unlabeled target domain with emerging motor faults when the source training data contains very few distinct categories. Incorporating FL may also provide a slight advantage by diffusing knowledge of health conditions between machines.
With the recent progress in large-scale vision and language representation learning, Vision Language Pre-training (VLP) models have achieved promising improvements on various multi-modal downstream tasks. Albeit powerful, these models have not fully leveraged world knowledge to their advantage. A key challenge of knowledge-augmented VLP is the lack of clear connections between knowledge and multi-modal data. Moreover, not all knowledge present in images/texts is useful, therefore prior approaches often struggle to effectively integrate knowledge, visual, and textual information. In this study, we propose REtrieval-based knowledge Augmented Vision Language (REAVL), a novel knowledge-augmented pre-training framework to address the above issues. For the first time, we introduce a knowledge-aware self-supervised learning scheme that efficiently establishes the correspondence between knowledge and multi-modal data and identifies informative knowledge to improve the modeling of alignment and interactions between visual and textual modalities. By adaptively integrating informative knowledge with visual and textual information, REAVL achieves new state-of-the-art performance uniformly on knowledge-based vision-language understanding and multi-modal entity linking tasks, as well as competitive results on general vision-language tasks while only using 0.2% pre-training data of the best models. Our model shows strong sample efficiency and effective knowledge utilization.
Machine learning methods for satellite data have a range of societally relevant applications, but labels used to train models can be difficult or impossible to acquire. Self-supervision is a natural solution in settings with limited labeled data, but current self-supervised models for satellite data fail to take advantage of the characteristics of that data, including the temporal dimension (which is critical for many applications, such as monitoring crop growth) and availability of data from many complementary sensors (which can significantly improve a model's predictive performance). We present Presto (the Pretrained Remote Sensing Transformer), a model pre-trained on remote sensing pixel-timeseries data. By designing Presto specifically for remote sensing data, we can create a significantly smaller but performant model. Presto excels at a wide variety of globally distributed remote sensing tasks and performs competitively with much larger models while requiring far less compute. Presto can be used for transfer learning or as a feature extractor for simple models, enabling efficient deployment at scale.
While Hopfield networks are known as paradigmatic models for memory storage and retrieval, modern artificial intelligence systems mainly stand on the machine learning paradigm. We show that it is possible to formulate a teacher-student self-supervised learning problem with Boltzmann machines in terms of a suitable generalization of the Hopfield model with structured patterns, where the spin variables are the machine weights and patterns correspond to the training set's examples. We analyze the learning performance by studying the phase diagram in terms of the training set size, the dataset noise and the inference temperature (i.e. the weight regularization). With a small but informative dataset the machine can learn by memorization. With a noisy dataset, an extensive number of examples above a critical threshold is needed. In this regime the memory storage limits of the system becomes an opportunity for the occurrence of a learning regime in which the system can generalize.
Self-supervised multi-object trackers have tremendous potential as they enable learning from raw domain-specific data. However, their re-identification accuracy still falls short compared to their supervised counterparts. We hypothesize that this drawback results from formulating self-supervised objectives that are limited to single frames or frame pairs. Such formulations do not capture sufficient visual appearance variations to facilitate learning consistent re-identification features for autonomous driving when the frame rate is low or object dynamics are high. In this work, we propose a training objective that enables self-supervised learning of re-identification features from multiple sequential frames by enforcing consistent association scores across short and long timescales. We perform extensive evaluations demonstrating that re-identification features trained from longer sequences significantly reduce ID switches on standard autonomous driving datasets compared to existing self-supervised learning methods, which are limited to training on frame pairs. Using our proposed SubCo loss function, we set the new state-of-the-art among self-supervised methods and even perform on par with fully supervised learning methods.