Self-supervised learning (SSL) achieves great success in speech recognition, while limited exploration has been attempted for other speech processing tasks. As speech signal contains multi-faceted information including speaker identity, paralinguistics, spoken content, etc., learning universal representations for all speech tasks is challenging. To tackle the problem, we propose a new pre-trained model, WavLM, to solve full-stack downstream speech tasks. WavLM jointly learns masked speech prediction and denoising in pre-training. By this means, WavLM does not only keep the speech content modeling capability by the masked speech prediction, but also improves the potential to non-ASR tasks by the speech denoising. In addition, WavLM employs gated relative position bias for the Transformer structure to better capture the sequence ordering of input speech. We also scale up the training dataset from 60k hours to 94k hours. WavLM Large achieves state-of-the-art performance on the SUPERB benchmark, and brings significant improvements for various speech processing tasks on their representative benchmarks. The code and pre-trained models are available at https://aka.ms/wavlm.
Recent contrastive methods show significant improvement in self-supervised learning in several domains. In particular, contrastive methods are most effective where data augmentation can be easily constructed e.g. in computer vision. However, they are less successful in domains without established data transformations such as time series data. In this paper, we propose a novel self-supervised learning framework that combines contrastive learning with neural processes. It relies on recent advances in neural processes to perform time series forecasting. This allows to generate augmented versions of data by employing a set of various sampling functions and, hence, avoid manually designed augmentations. We extend conventional neural processes and propose a new contrastive loss to learn times series representations in a self-supervised setup. Therefore, unlike previous self-supervised methods, our augmentation pipeline is task-agnostic, enabling our method to perform well across various applications. In particular, a ResNet with a linear classifier trained using our approach is able to outperform state-of-the-art techniques across industrial, medical and audio datasets improving accuracy over 10% in ECG periodic data. We further demonstrate that our self-supervised representations are more efficient in the latent space, improving multiple clustering indexes and that fine-tuning our method on 10% of labels achieves results competitive to fully-supervised learning.
Despite the large augmentation family, only a few cherry-picked robust augmentation policies are beneficial to self-supervised image representation learning. In this paper, we propose a directional self-supervised learning paradigm (DSSL), which is compatible with significantly more augmentations. Specifically, we adapt heavy augmentation policies after the views lightly augmented by standard augmentations, to generate harder view (HV). HV usually has a higher deviation from the original image than the lightly augmented standard view (SV). Unlike previous methods equally pairing all augmented views to symmetrically maximize their similarities, DSSL treats augmented views of the same instance as a partially ordered set (with directions as SV$\leftrightarrow $SV, SV$\leftarrow$HV), and then equips a directional objective function respecting to the derived relationships among views. DSSL can be easily implemented with a few lines of codes and is highly flexible to popular self-supervised learning frameworks, including SimCLR, SimSiam, BYOL. Extensive experimental results on CIFAR and ImageNet demonstrated that DSSL can stably improve various baselines with compatibility to a wider range of augmentations.
Self-supervised learning (SSL) has achieved great success in a variety of computer vision tasks. However, the mechanism of how SSL works in these tasks remains a mystery. In this paper, we study how SSL can enhance the performance of the out-of-distribution (OOD) detection task. We first point out two general properties that a good OOD detector should have: 1) the overall feature space should be large and 2) the inlier feature space should be small. Then we demonstrate that SSL can indeed increase the intrinsic dimension of the overall feature space. In the meantime, SSL even has the potential to shrink the inlier feature space. As a result, there will be more space spared for the outliers, making OOD detection much easier. The conditions when SSL can shrink the inlier feature space is also discussed and validated. By understanding the role of SSL in the OOD detection task, our study can provide a guideline for designing better OOD detection algorithms. Moreover, this work can also shed light to other tasks where SSL can improve the performance.
We present the use of self-supervised learning to explore and exploit large unlabeled datasets. Focusing on 42 million galaxy images from the latest data release of the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Surveys, we first train a self-supervised model to distill low-dimensional representations that are robust to symmetries, uncertainties, and noise in each image. We then use the representations to construct and publicly release an interactive semantic similarity search tool. We demonstrate how our tool can be used to rapidly discover rare objects given only a single example, increase the speed of crowd-sourcing campaigns, and construct and improve training sets for supervised applications. While we focus on images from sky surveys, the technique is straightforward to apply to any scientific dataset of any dimensionality. The similarity search web app can be found at https://github.com/georgestein/galaxy_search
Self-supervised learning has gradually emerged as a powerful technique for graph representation learning. However, transferable, generalizable, and robust representation learning on graph data still remains a challenge for pre-training graph neural networks. In this paper, we propose a simple and effective self-supervised pre-training strategy, named Pairwise Half-graph Discrimination (PHD), that explicitly pre-trains a graph neural network at graph-level. PHD is designed as a simple binary classification task to discriminate whether two half-graphs come from the same source. Experiments demonstrate that the PHD is an effective pre-training strategy that offers comparable or superior performance on 13 graph classification tasks compared with state-of-the-art strategies, and achieves notable improvements when combined with node-level strategies. Moreover, the visualization of learned representation revealed that PHD strategy indeed empowers the model to learn graph-level knowledge like the molecular scaffold. These results have established PHD as a powerful and effective self-supervised learning strategy in graph-level representation learning.
Recent years have witnessed the significant success of applying graph neural networks (GNNs) in learning effective node representations for classification. However, current GNNs are mostly built under the balanced data-splitting, which is inconsistent with many real-world networks where the number of training nodes can be extremely imbalanced among the classes. Thus, directly utilizing current GNNs on imbalanced data would generate coarse representations of nodes in minority classes and ultimately compromise the classification performance. This therefore portends the importance of developing effective GNNs for handling imbalanced graph data. In this work, we propose a novel Distance-wise Prototypical Graph Neural Network (DPGNN), which proposes a class prototype-driven training to balance the training loss between majority and minority classes and then leverages distance metric learning to differentiate the contributions of different dimensions of representations and fully encode the relative position of each node to each class prototype. Moreover, we design a new imbalanced label propagation mechanism to derive extra supervision from unlabeled nodes and employ self-supervised learning to smooth representations of adjacent nodes while separating inter-class prototypes. Comprehensive node classification experiments and parameter analysis on multiple networks are conducted and the proposed DPGNN almost always significantly outperforms all other baselines, which demonstrates its effectiveness in imbalanced node classification. The implementation of DPGNN is available at \url{https://github.com/YuWVandy/DPGNN}.
Active learning focuses on choosing a subset of unlabeled data to be labeled. However, most such methods assume that a large subset of the data can be annotated. We are interested in low-budget active learning where only a small subset (e.g., 0.2% of ImageNet) can be annotated. Instead of proposing a new query strategy to iteratively sample batches of unlabeled data given an initial pool, we learn rich features by an off-the-shelf self-supervised learning method only once, and then study the effectiveness of different sampling strategies given a low labeling budget on a variety of datasets including ImageNet. We show that although the state-of-the-art active learning methods work well given a large labeling budget, a simple K-means clustering algorithm can outperform them on low budgets. We believe this method can be used as a simple baseline for low-budget active learning on image classification. Code is available at: https://github.com/UCDvision/low-budget-al
In this technical report, we present our approaches for the continual object detection track of the SODA10M challenge. We adapt ResNet50-FPN as the baseline and try several improvements for the final submission model. We find that task-specific replay scheme, learning rate scheduling, model calibration, and using original image scale helps to improve performance for both large and small objects in images. Our team `hypertune28' secured the second position among 52 participants in the challenge. This work will be presented at the ICCV 2021 Workshop on Self-supervised Learning for Next-Generation Industry-level Autonomous Driving (SSLAD).
We are interested in representation learning from labeled or unlabeled data. Inspired by recent success of self-supervised learning (SSL), we develop a non-contrastive representation learning method that can exploit additional knowledge. This additional knowledge may come from annotated labels in the supervised setting or an SSL model from another modality in the SSL setting. Our main idea is to generalize the mean-shift algorithm by constraining the search space of nearest neighbors, resulting in semantically purer representations. Our method simply pulls the embedding of an instance closer to its nearest neighbors in a search space that is constrained using the additional knowledge. By leveraging this non-contrastive loss, we show that the supervised ImageNet-1k pretraining with our method results in better transfer performance as compared to the baselines. Further, we demonstrate that our method is relatively robust to label noise. Finally, we show that it is possible to use the noisy constraint across modalities to train self-supervised video models.
In self-supervised learning, a model is trained to solve a pretext task, using a data set whose annotations are created by a machine. The objective is to transfer the trained weights to perform a downstream task in the target domain. We critically examine the most notable pretext tasks to extract features from image data and further go on to conduct experiments on resource constrained networks, which aid faster experimentation and deployment. We study the performance of various self-supervised techniques keeping all other parameters uniform. We study the patterns that emerge by varying model type, size and amount of pre-training done for the backbone as well as establish a standard to compare against for future research. We also conduct comprehensive studies to understand the quality of representations learned by different architectures.
Self-supervised monocular depth estimation has been widely investigated to estimate depth images and relative poses from RGB images. This framework is attractive for researchers because the depth and pose networks can be trained from just time sequence images without the need for the ground truth depth and poses. In this work, we estimate the depth around a robot (360 degree view) using time sequence spherical camera images, from a camera whose parameters are unknown. We propose a learnable axisymmetric camera model which accepts distorted spherical camera images with two fisheye camera images. In addition, we trained our models with a photo-realistic simulator to generate ground truth depth images to provide supervision. Moreover, we introduced loss functions to provide floor constraints to reduce artifacts that can result from reflective floor surfaces. We demonstrate the efficacy of our method using the spherical camera images from the GO Stanford dataset and pinhole camera images from the KITTI dataset to compare our method's performance with that of baseline method in learning the camera parameters.
CLIP yielded impressive results on zero-shot transfer learning tasks and is considered as a foundation model like BERT or GPT3. CLIP vision models that have a rich representation are pre-trained using the InfoNCE objective and natural language supervision before they are fine-tuned on particular tasks. Though CLIP excels at zero-shot transfer learning, it suffers from an explaining away problem, that is, it focuses on one or few features, while neglecting other relevant features. This problem is caused by insufficiently extracting the covariance structure in the original multi-modal data. We suggest to use modern Hopfield networks to tackle the problem of explaining away. Their retrieved embeddings have an enriched covariance structure derived from co-occurrences of features in the stored embeddings. However, modern Hopfield networks increase the saturation effect of the InfoNCE objective which hampers learning. We propose to use the InfoLOOB objective to mitigate this saturation effect. We introduce the novel "Contrastive Leave One Out Boost" (CLOOB), which uses modern Hopfield networks for covariance enrichment together with the InfoLOOB objective. In experiments we compare CLOOB to CLIP after pre-training on the Conceptual Captions and the YFCC dataset with respect to their zero-shot transfer learning performance on other datasets. CLOOB consistently outperforms CLIP at zero-shot transfer learning across all considered architectures and datasets.
The remarkable performance of the pre-trained language model (LM) using self-supervised learning has led to a major paradigm shift in the study of natural language processing. In line with these changes, leveraging the performance of speech recognition systems with massive deep learning-based LMs is a major topic of speech recognition research. Among the various methods of applying LMs to speech recognition systems, in this paper, we focus on a cross-modal knowledge distillation method that transfers knowledge between two types of deep neural networks with different modalities. We propose an acoustic model structure with multiple auxiliary output layers for cross-modal distillation and demonstrate that the proposed method effectively compensates for the shortcomings of the existing label-interpolation-based distillation method. In addition, we extend the proposed method to a hierarchical distillation method using LMs trained in different units (senones, monophones, and subwords) and reveal the effectiveness of the hierarchical distillation method through an ablation study.
The objective of this paper is to learn semantic representations for sleep stage classification from raw physiological time series. Although supervised methods have gained remarkable performance, they are limited in clinical situations due to the requirement of fully labeled data. Self-supervised learning (SSL) based on contrasting semantically similar (positive) and dissimilar (negative) pairs of samples have achieved promising success. However, existing SSL methods suffer the problem that many semantically similar positives are still uncovered and even treated as negatives. In this paper, we propose a novel SSL approach named SleepPriorCL to alleviate the above problem. Advances of our approach over existing SSL methods are two-fold: 1) by incorporating prior domain knowledge into the training regime of SSL, more semantically similar positives are discovered without accessing ground-truth labels; 2) via investigating the influence of the temperature in contrastive loss, an adaptive temperature mechanism for each sample according to prior domain knowledge is further proposed, leading to better performance. Extensive experiments demonstrate that our method achieves state-of-the-art performance and consistently outperforms baselines.
Traditionally, research in automated speech recognition has focused on local-first encoding of audio representations to predict the spoken phonemes in an utterance. Unfortunately, approaches relying on such hyper-local information tend to be vulnerable to both local-level corruption (such as audio-frame drops, or loud noises) and global-level noise (such as environmental noise, or background noise) that has not been seen during training. In this work, we introduce a novel approach which leverages a self-supervised learning technique based on masked language modeling to compute a global, multi-modal encoding of the environment in which the utterance occurs. We then use a new deep-fusion framework to integrate this global context into a traditional ASR method, and demonstrate that the resulting method can outperform baseline methods by up to 7% on Librispeech; gains on internal datasets range from 6% (on larger models) to 45% (on smaller models).
Speech representation learning plays a vital role in speech processing. Among them, self-supervised learning (SSL) has become an important research direction. It has been shown that an SSL pretraining model can achieve excellent performance in various downstream tasks of speech processing. On the other hand, supervised multi-task learning (MTL) is another representation learning paradigm, which has been proven effective in computer vision (CV) and natural language processing (NLP). However, there is no systematic research on the general representation learning model trained by supervised MTL in speech processing. In this paper, we show that MTL finetuning can further improve SSL pretraining. We analyze the generalizability of supervised MTL finetuning to examine if the speech representation learned by MTL finetuning can generalize to unseen new tasks.
Recently, neural networks based purely on self-attention, such as the Vision Transformer (ViT), have been shown to outperform deep learning models constructed with convolutional neural networks (CNNs) on various vision tasks, thus extending the success of Transformers, which were originally developed for language processing, to the vision domain. A recent study showed that a similar methodology can also be applied to the audio domain. Specifically, the Audio Spectrogram Transformer (AST) achieves state-of-the-art results on various audio classification benchmarks. However, pure Transformer models tend to require more training data compared to CNNs, and the success of the AST relies on supervised pretraining that requires a large amount of labeled data and a complex training pipeline, thus limiting the practical usage of AST. This paper focuses on audio and speech classification, and aims to reduce the need for large amounts of labeled data for AST by leveraging self-supervised learning using unlabeled data. Specifically, we propose to pretrain the AST model with joint discriminative and generative masked spectrogram patch modeling (MSPM) using unlabeled audio from AudioSet and Librispeech. We evaluate our pretrained models on both audio and speech classification tasks including audio event classification, keyword spotting, emotion recognition, and speaker identification. The proposed self-supervised framework significantly boosts AST performance on all tasks, with an average improvement of 60.9%, leading to similar or even better results than a supervised pretrained AST. To the best of our knowledge, it is the first patch-based self-supervised learning framework in the audio and speech domain, and also the first self-supervised learning framework for AST.
We introduce DECAR, a self-supervised pre-training approach for learning general-purpose audio representations. Our system is based on clustering: it utilizes an offline clustering step to provide target labels that act as pseudo-labels for solving a prediction task. We develop on top of recent advances in self-supervised learning for computer vision and design a lightweight, easy-to-use self-supervised pre-training scheme. We pre-train DECAR embeddings on a balanced subset of the large-scale Audioset dataset and transfer those representations to 9 downstream classification tasks, including speech, music, animal sounds, and acoustic scenes. Furthermore, we conduct ablation studies identifying key design choices and also make all our code and pre-trained models publicly available.
SOTA Transformer and DNN short text sentiment classifiers report over 97% accuracy on narrow domains like IMDB movie reviews. Real-world performance is significantly lower because traditional models overfit benchmarks and generalize poorly to different or more open domain texts. This paper introduces SentimentArcs, a new self-supervised time series sentiment analysis methodology that addresses the two main limitations of traditional supervised sentiment analysis: limited labeled training datasets and poor generalization. A large ensemble of diverse models provides a synthetic ground truth for self-supervised learning. Novel metrics jointly optimize an exhaustive search across every possible corpus:model combination. The joint optimization over both the corpus and model solves the generalization problem. Simple visualizations exploit the temporal structure in narratives so domain experts can quickly spot trends, identify key features, and note anomalies over hundreds of arcs and millions of data points. To our knowledge, this is the first self-supervised method for time series sentiment analysis and the largest survey directly comparing real-world model performance on long-form narratives.