ICLR 2023 论文列表

We introduce a regularization loss based on kernel mean embeddings with rotation-invariant kernels on the hypersphere (also known as dot-product kernels) for self-supervised learning of image representations. Besides being fully competitive with the state of the art, our method significantly reduces time and memory complexity for self-supervised training, making it implementable for very large embedding dimensions on existing devices and more easily adjustable than previous methods to settings with limited resources. Our work follows the major paradigm where the model learns to be invariant to some predefined image transformations (cropping, blurring, color jittering, etc.), while avoiding a degenerate solution by regularizing the embedding distribution. Our particular contribution is to propose a loss family promoting the embedding distribution to be close to the uniform distribution on the hypersphere, with respect to the maximum mean discrepancy pseudometric. We demonstrate that this family encompasses several regularizers of former methods, including uniformity-based and information-maximization methods, which are variants of our flexible regularization loss with different kernels. Beyond its practical consequences for state of the art self-supervised learning with limited resources, the proposed generic regularization approach opens perspectives to leverage more widely the literature on kernel methods in order to improve self-supervised learning methods.

Recently, a few self-supervised representation learning (SSL) methods have outperformed the ImageNet classification pre-training for vision tasks such as object detection. However, its effects on 3D human body pose and shape estimation (3DHPSE) are open to question, whose target is fixed to a unique class, the human, and has an inherent task gap with SSL. We empirically study and analyze the effects of SSL and further compare it with other pre-training alternatives for 3DHPSE. The alternatives are 2D annotation-based pre-training and synthetic data pre-training, which share the motivation of SSL that aims to reduce the labeling cost. They have been widely utilized as a source of weak-supervision or fine-tuning, but have not been remarked as a pre-training source. SSL methods underperform the conventional ImageNet classification pre-training on multiple 3DHPSE benchmarks by 7.7% on average. In contrast, despite a much less amount of pre-training data, the 2D annotation-based pre-training improves accuracy on all benchmarks and shows faster convergence during fine-tuning. Our observations challenge the naive application of the current SSL pre-training to 3DHPSE and relight the value of other data types in the pre-training aspect.

As a typical self-supervised learning strategy, Masked Image Modeling (MIM) is driven by recovering all masked patches from visible ones. However, patches from the same image are highly correlated and it is redundant to reconstruct all the masked patches. We find that this redundancy is neglected by existing MIM based methods and causes non-negligible overheads in computation that do not necessarily benefit self-supervised representation. In this paper, we present a novel approach named PCAE, short for Progressively Compressed AutoEncoder, to address the redundant reconstruction issue by progressively compacting tokens and only retaining necessary information for forward propagation and reconstruction. In particular, we identify those redundant tokens in an image via a simple yet effective similarity metric between each token with the mean of the token sequence. Those redundant tokens that other ones can probably represent are progressively dropped accordingly during the forward propagation, and importantly, we only focus on reconstructing these retained tokens. As a result, we are able to achieve a better trade-off between performance and efficiency for pre-training. Besides, benefitting from the flexible strategy, PCAE can be also directly employed for downstream fine-tuning tasks and enable scalable deployment. Experiments show that PCAE achieves comparable performance to MAE with only 1/8 GPU days. The code is available at https://github.com/caddyless/PCAE/.

Masked image modeling (MIM) has become a popular strategy for self-supervised learning (SSL) of visual representations with Vision Transformers. A representative MIM model, the masked auto-encoder (MAE), randomly masks a subset of image patches and reconstructs the masked patches given the unmasked patches. Concurrently, many recent works in self-supervised learning utilize the student/teacher paradigm which provides the student with an additional target based on the output of a teacher composed of an exponential moving average (EMA) of previous students. Although common, relatively little is known about the dynamics of the interaction between the student and teacher. Through analysis on a simple linear model, we find that the teacher conditionally removes previous gradient directions based on feature similarities which effectively acts as a conditional momentum regularizer. From this analysis, we present a simple SSL method, the Reconstruction-Consistent Masked Auto-Encoder (RC-MAE) by adding an EMA teacher to MAE. We find that RC-MAE converges faster and requires less memory usage than state-of-the-art self-distillation methods during pre-training, which may provide a way to enhance the practicality of prohibitively expensive self-supervised learning of Vision Transformer models. Additionally, we show that RC-MAE achieves more robustness and better performance compared to MAE on downstream tasks such as ImageNet-1K classification, object detection, and instance segmentation.

Self-supervised learning has significantly improved the performance of many NLP tasks. However, how can self-supervised learning discover useful features, and why is it better than traditional approaches such as probabilistic models are still largely unknown. In this paper, we focus on the context of topic modeling and highlight a key advantage of self-supervised learning - when applied to data generated by topic models, self-supervised learning can be oblivious to the specific model, and hence is less susceptible to model misspecification. In particular, we prove that commonly used self-supervised objectives based on reconstruction or contrastive samples can both recover useful posterior information for general topic models. Empirically, we show that the same objectives can perform on par with posterior inference using the correct model, while outperforming posterior inference using misspecified models.

Partitioning a set of elements into subsets of a priori unknown sizes is essential in many applications. These subset sizes are rarely explicitly learned - be it the cluster sizes in clustering applications or the number of shared versus independent generative latent factors in weakly-supervised learning. Probability distributions over correct combinations of subset sizes are non-differentiable due to hard constraints, which prohibit gradient-based optimization. In this work, we propose the differentiable hypergeometric distribution. The hypergeometric distribution models the probability of different group sizes based on their relative importance. We introduce reparameterizable gradients to learn the importance between groups and highlight the advantage of explicitly learning the size of subsets in two typical applications: weakly-supervised learning and clustering. In both applications, we outperform previous approaches, which rely on suboptimal heuristics to model the unknown size of groups.

Disentangling complex data to its latent factors of variation is a fundamental task in representation learning. Existing work on sequential disentanglement mostly provides two factor representations, i.e., it separates the data to time-varying and time-invariant factors. In contrast, we consider multifactor disentanglement in which multiple (more than two) semantic disentangled components are generated. Key to our approach is a strong inductive bias where we assume that the underlying dynamics can be represented linearly in the latent space. Under this assumption, it becomes natural to exploit the recently introduced Koopman autoencoder models. However, disentangled representations are not guaranteed in Koopman approaches, and thus we propose a novel spectral loss term which leads to structured Koopman matrices and disentanglement. Overall, we propose a simple and easy to code new deep model that is fully unsupervised and it supports multifactor disentanglement. We showcase new disentangling abilities such as swapping of individual static factors between characters, and an incremental swap of disentangled factors from the source to the target. Moreover, we evaluate our method extensively on two factor standard benchmark tasks where we significantly improve over competing unsupervised approaches, and we perform competitively in comparison to weakly- and self-supervised state-of-the-art approaches. The code is available at https://github.com/azencot-group/SKD.

We tackle the problem of Selective Classification where the objective is to achieve the best performance on a predetermined ratio (coverage) of the dataset. Recent state-of-the-art selective methods come with architectural changes either via introducing a separate selection head or an extra abstention logit. In this paper, we challenge the aforementioned methods. The results suggest that the superior performance of state-of-the-art methods is owed to training a more generalizable classifier rather than their proposed selection mechanisms. We argue that the best performing selection mechanism should instead be rooted in the classifier itself. Our proposed selection strategy uses the classification scores and achieves better results by a significant margin, consistently, across all coverages and all datasets, without any added compute cost. Furthermore, inspired by semi-supervised learning, we propose an entropy-based regularizer that improves the performance of selective classification methods. Our proposed selection mechanism with the proposed entropy-based regularizer achieves new state-of-the-art results.

Recent Self-Supervised Learning (SSL) methods are able to learn feature representations that are invariant to different data augmentations, which can then be transferred to downstream tasks of interest. However, different downstream tasks require different invariances for their best performance, so the optimal choice of augmentations for SSL depends on the target task. In this paper, we aim to learn self-supervised features that generalize well across a variety of downstream tasks (e.g., object classification, detection and instance segmentation) without knowing any task information beforehand. We do so by Masked Augmentation Subspace Training (or MAST) to encode in the single feature space the priors from different data augmentations in a factorized way. Specifically, we disentangle the feature space into separate subspaces, each induced by a learnable mask that selects relevant feature dimensions to model invariance to a specific augmentation. We show the success of MAST in jointly capturing generalizable priors from different augmentations, using both unique and shared features across the subspaces. We further show that MAST benefits from uncertainty modeling to reweight ambiguous samples from strong augmentations that may cause similarity mismatch in each subspace. Experiments demonstrate that MAST consistently improves generalization on various downstream tasks, while being task-agnostic and efficient during SSL. We also provide interesting insights about how different augmentations are related and how uncertainty reflects learning difficulty.

Prevention of complete and dimensional collapse of representations has recently become a design principle for self-supervised learning (SSL). However, questions remain in our theoretical understanding: When do those collapses occur? What are the mechanisms and causes? We answer these questions by deriving and thoroughly analyzing an analytically tractable theory of SSL loss landscapes. In this theory, we identify the causes of the dimensional collapse and study the effect of normalization and bias. Finally, we leverage the interpretability afforded by the analytical theory to understand how dimensional collapse can be beneficial and what affects the robustness of SSL against data imbalance.

We present Bit Diffusion: a simple and generic approach for generating discrete data with continuous state and continuous time diffusion models. The main idea behind our approach is to first represent the discrete data as binary bits, and then train a continuous diffusion model to model these bits as real numbers which we call analog bits. To generate samples, the model first generates the analog bits, which are then thresholded to obtain the bits that represent the discrete variables. We further propose two simple techniques, namely Self-Conditioning and Asymmetric Time Intervals, which lead to a significant improvement in sample quality. Despite its simplicity, the proposed approach can achieve strong performance in both discrete image generation and image captioning tasks. For discrete image generation, we significantly improve previous state-of-the-art on both CIFAR-10 (which has 3K discrete 8-bit tokens) and ImageNet-64x64 (which has 12K discrete 8-bit tokens), outperforming the best autoregressive model in both sample quality (measured by FID) and efficiency. For image captioning on MS-COCO dataset, our approach achieves competitive results compared to autoregressive models.

We consider the problem of training a deep neural network on a given classification task, e.g., ImageNet-1K (IN1K), so that it excels at both the training task as well as at other (future) transfer tasks. These two seemingly contradictory properties impose a trade-off between improving the model’s generalization and maintaining its performance on the original task. Models trained with self-supervised learning tend to generalize better than their supervised counterparts for transfer learning; yet, they still lag behind supervised models on IN1K. In this paper, we propose a supervised learning setup that leverages the best of both worlds. We extensively analyze supervised training using multi-scale crops for data augmentation and an expendable projector head, and reveal that the design of the projector allows us to control the trade-off between performance on the training task and transferability. We further replace the last layer of class weights with class prototypes computed on the fly using a memory bank and derive two models: t-ReX that achieves a new state of the art for transfer learning and outperforms top methods such as DINO and PAWS on IN1K, and t-ReX* that matches the highly optimized RSB-A1 model on IN1K while performing better on transfer tasks. Code and pretrained models: https://europe.naverlabs.com/t-rex

There has been a debate on the choice of plain vs. hierarchical vision transformers, where researchers often believe that the former (e.g., ViT) has a simpler design but the latter (e.g., Swin) enjoys higher recognition accuracy. Recently, the emerge of masked image modeling (MIM), a self-supervised visual pre-training method, raised a new challenge to vision transformers in terms of flexibility, i.e., part of image patches or tokens are to be discarded, which seems to claim the advantages of plain vision transformers. In this paper, we delve deep into the comparison between ViT and Swin, revealing that (i) the performance gain of Swin is mainly brought by a deepened backbone and relative positional encoding, (ii) the hierarchical design of Swin can be simplified into hierarchical patch embedding (proposed in this work), and (iii) other designs such as shifted-window attentions can be removed. By removing the unnecessary operations, we come up with a new architecture named HiViT (short for hierarchical ViT), which is simpler and more efficient than Swin yet further improves its performance on fully-supervised and self-supervised visual representation learning. In particular, after pre-trained using masked autoencoder (MAE) on ImageNet-1K, HiViT-B reports a 84.6% accuracy on ImageNet-1K classification, a 53.3% box AP on COCO detection, and a 52.8% mIoU on ADE20K segmentation, significantly surpassing the baseline. Code is available at https://github.com/zhangxiaosong18/hivit.

Self-paced learning has been beneficial for tasks where some initial knowledge is available, such as weakly supervised learning and domain adaptation, to select and order the training sample sequence, from easy to complex. However its applicability remains unexplored in unsupervised learning, whereby the knowledge of the task matures during training. We propose a novel HYperbolic Self-Paced model (HYSP) for learning skeletonbased action representations. HYSP adopts self-supervision: it uses data augmentations to generate two views of the same sample, and it learns by matching one (named online) to the other (the target). We propose to use hyperbolic uncertainty to determine the algorithmic learning pace, under the assumption that less uncertain samples should be more strongly driving the training, with a larger weight and pace. Hyperbolic uncertainty is a by-product of the adopted hyperbolic neural networks, it matures during training and it comes with no extra cost, compared to the established Euclidean SSL framework counterparts. When tested on three established skeleton-based action recognition datasets, HYSP outperforms the state-of-the-art on PKU-MMD I, as well as on 2 out of 3 downstream tasks on NTU-60 and NTU-120. Additionally, HYSP only uses positive pairs and bypasses therefore the complex and computationally-demanding mining procedures required for the negatives in contrastive techniques. Code is available at https://github.com/paolomandica/HYSP.

Recently, great success has been made in learning visual representations from text supervision, facilitating the emergence of text-supervised semantic segmentation. However, existing works focus on pixel grouping and cross-modal semantic alignment, while ignoring the correspondence among multiple augmented views of the same image. To overcome such limitation, we propose multi-View Consistent learning (ViewCo) for text-supervised semantic segmentation. Specifically, we first propose text-to-views consistency modeling to learn correspondence for multiple views of the same input image. Additionally, we propose cross-view segmentation consistency modeling to address the ambiguity issue of text supervision by contrasting the segment features of Siamese visual encoders. The text-to-views consistency benefits dense assignment of the visual features by encouraging different crops to align with the same text, while the cross-view segmentation consistency modeling provides additional self-supervision, overcoming the limitation of ambiguous text supervision for segmentation masks. Trained with large-scale image-text data, our model can directly segment objects of arbitrary categories in a zero-shot manner. Extensive experiments show that ViewCo outperforms state-of-the-art methods on average by up to 2.9%, 1.6%, and 2.4% mIoU on PASCAL VOC2012, PASCAL Context, and COCO, respectively.

Existing collaborative self-supervised learning (SSL) schemes are not suitable for cross-client applications because of their expensive computation and large local data requirements. To address these issues, we propose MocoSFL, a collaborative SSL framework based on Split Federated Learning (SFL) and Momentum Contrast (MoCo). In MocoSFL, the large backbone model is split into a small client-side model and a large server-side model, and only the small client-side model is processed locally on the client's local devices. MocoSFL has three key components: (i) vector concatenation which enables the use of small batch size and reduces computation and memory requirements by orders of magnitude; (ii) feature sharing that helps achieve high accuracy regardless of the quality and volume of local data; (iii) frequent synchronization that helps achieve better non-IID performance because of smaller local model divergence. For a 1,000-client case with non-IID data (each client only has data from 2 random classes of CIFAR-10), MocoSFL can achieve over 84% accuracy with ResNet-18 model. Next we present TAResSFL module that significantly improves the resistance to privacy threats and communication overhead with small sacrifice in accuracy for a MocoSFL system. On a Raspberry Pi 4B device, the MocoSFL-based scheme requires less than 1MB of memory and less than 40MB of communication, and consumes less than 5W power. The code is available at https://github.com/SonyAI/MocoSFL.

The success of machine learning relies heavily on massive amounts of data, which are usually generated and stored across a range of diverse and distributed data sources. Decentralized learning has thus been advocated and widely deployed to make efficient use of distributed datasets, with an extensive focus on supervised learning (SL) problems. Unfortunately, the majority of real-world data are unlabeled and can be highly heterogeneous across sources. In this work, we carefully study decentralized learning with unlabeled data through the lens of self-supervised learning (SSL), specifically contrastive visual representation learning. We study the effectiveness of a range of contrastive learning algorithms under a decentralized learning setting, on relatively large-scale datasets including ImageNet-100, MS-COCO, and a new real-world robotic warehouse dataset. Our experiments show that the decentralized SSL (Dec-SSL) approach is robust to the heterogeneity of decentralized datasets, and learns useful representation for object classification, detection, and segmentation tasks, even when combined with the simple and standard decentralized learning algorithm of Federated Averaging (FedAvg). This robustness makes it possible to significantly reduce communication and to reduce the participation ratio of data sources with only minimal drops in performance. Interestingly, using the same amount of data, the representation learned by Dec-SSL can not only perform on par with that learned by centralized SSL which requires communication and excessive data storage costs, but also sometimes outperform representations extracted from decentralized SL which requires extra knowledge about the data labels. Finally, we provide theoretical insights into understanding why data heterogeneity is less of a concern for Dec-SSL objectives, and introduce feature alignment and clustering techniques to develop a new Dec-SSL algorithm that further improves the performance, in the face of highly non-IID data. Our study presents positive evidence to embrace unlabeled data in decentralized learning, and we hope to provide new insights into whether and why decentralized SSL is effective and/or even advantageous.

Category-level articulated object pose estimation aims to estimate a hierarchy of articulation-aware object poses of an unseen articulated object from a known category. To reduce the heavy annotations needed for supervised learning methods, we present a novel self-supervised strategy that solves this problem without any human labels. Our key idea is to factorize canonical shapes and articulated object poses from input articulated shapes through part-level equivariant shape analysis. Specifically, we first introduce the concept of part-level SE(3) equivariance and devise a network to learn features of such property. Then, through a carefully designed fine-grained pose-shape disentanglement strategy, we expect that canonical spaces to support pose estimation could be induced automatically. Thus, we could further predict articulated object poses as per-part rigid transformations describing how parts transform from their canonical part spaces to the camera space. Extensive experiments demonstrate the effectiveness of our method on both complete and partial point clouds from synthetic and real articulated object datasets.

Self-supervised learning (SSL) for graph neural networks (GNNs) has attracted increasing attention from the graph machine learning community in recent years, owing to its capability to learn performant node embeddings without costly label information. One weakness of conventional SSL frameworks for GNNs is that they learn through a single philosophy, such as mutual information maximization or generative reconstruction. When applied to various downstream tasks, these frameworks rarely perform equally well for every task, because one philosophy may not span the extensive knowledge required for all tasks. To enhance the task generalization across tasks, as an important first step forward in exploring fundamental graph models, we introduce PARETOGNN, a multi-task SSL framework for node representation learning over graphs. Specifically, PARETOGNN is self-supervised by manifold pretext tasks observing multiple philosophies. To reconcile different philosophies, we explore a multiple-gradient descent algorithm, such that PARETOGNN actively learns from every pretext task while minimizing potential conflicts. We conduct comprehensive experiments over four downstream tasks (i.e., node classification, node clustering, link prediction, and partition prediction), and our proposal achieves the best overall performance across tasks on 11 widely adopted benchmark datasets. Besides, we observe that learning from multiple philosophies enhances not only the task generalization but also the single task performances, demonstrating that PARETOGNN achieves better task generalization via the disjoint yet complementary knowledge learned from different philosophies. Our code is publicly available at https://github.com/jumxglhf/ParetoGNN.