3D reconstruction from a single 2D image was extensively covered in the literature but relies on depth supervision at training time, which limits its applicability. To relax the dependence to depth we propose SceneRF, a self-supervised monocular scene reconstruction method using only posed image sequences for training. Fueled by the recent progress in neural radiance fields (NeRF) we optimize a radiance field though with explicit depth optimization and a novel probabilistic sampling strategy to efficiently handle large scenes. At inference, a single input image suffices to hallucinate novel depth views which are fused together to obtain 3D scene reconstruction. Thorough experiments demonstrate that we outperform all baselines for novel depth views synthesis and scene reconstruction, on indoor BundleFusion and outdoor SemanticKITTI. Code is available at https://astra-vision.github.io/SceneRF .
Autoregressive language modeling (ALM) have been successfully used in self-supervised pre-training in Natural language processing (NLP). However, this paradigm has not achieved comparable results with other self-supervised approach in computer vision (e.g., contrastive learning, mask image modeling). In this paper, we try to find the reason why autoregressive modeling does not work well on vision tasks. To tackle this problem, we fully analyze the limitation of visual autoregressive methods and proposed a novel stochastic autoregressive image modeling (named SAIM) by the two simple designs. First, we employ stochastic permutation strategy to generate effective and robust image context which is critical for vision tasks. Second, we create a parallel encoder-decoder training process in which the encoder serves a similar role to the standard vision transformer focus on learning the whole contextual information, and meanwhile the decoder predicts the content of the current position, so that the encoder and decoder can reinforce each other. By introducing stochastic prediction and the parallel encoder-decoder, SAIM significantly improve the performance of autoregressive image modeling. Our method achieves the best accuracy (83.9%) on the vanilla ViT-Base model among methods using only ImageNet-1K data. Transfer performance in downstream tasks also show that our model achieves competitive performance.
Cross-modality magnetic resonance (MR) image synthesis can be used to generate missing modalities from given ones. Existing (supervised learning) methods often require a large number of paired multi-modal data to train an effective synthesis model. However, it is often challenging to obtain sufficient paired data for supervised training. In reality, we often have a small number of paired data while a large number of unpaired data. To take advantage of both paired and unpaired data, in this paper, we propose a Multi-scale Transformer Network (MT-Net) with edge-aware pre-training for cross-modality MR image synthesis. Specifically, an Edge-preserving Masked AutoEncoder (Edge-MAE) is first pre-trained in a self-supervised manner to simultaneously perform 1) image imputation for randomly masked patches in each image and 2) whole edge map estimation, which effectively learns both contextual and structural information. Besides, a novel patch-wise loss is proposed to enhance the performance of Edge-MAE by treating different masked patches differently according to the difficulties of their respective imputations. Based on this proposed pre-training, in the subsequent fine-tuning stage, a Dual-scale Selective Fusion (DSF) module is designed (in our MT-Net) to synthesize missing-modality images by integrating multi-scale features extracted from the encoder of the pre-trained Edge-MAE. Further, this pre-trained encoder is also employed to extract high-level features from the synthesized image and corresponding ground-truth image, which are required to be similar (consistent) in the training. Experimental results show that our MT-Net achieves comparable performance to the competing methods even using $70\%$ of all available paired data. Our code will be publicly available at https://github.com/lyhkevin/MT-Net.
Many Click-Through Rate (CTR) prediction works focused on designing advanced architectures to model complex feature interactions but neglected the importance of feature representation learning, e.g., adopting a plain embedding layer for each feature, which results in sub-optimal feature representations and thus inferior CTR prediction performance. For instance, low frequency features, which account for the majority of features in many CTR tasks, are less considered in standard supervised learning settings, leading to sub-optimal feature representations. In this paper, we introduce self-supervised learning to produce high-quality feature representations directly and propose a model-agnostic Contrastive Learning for CTR (CL4CTR) framework consisting of three self-supervised learning signals to regularize the feature representation learning: contrastive loss, feature alignment, and field uniformity. The contrastive module first constructs positive feature pairs by data augmentation and then minimizes the distance between the representations of each positive feature pair by the contrastive loss. The feature alignment constraint forces the representations of features from the same field to be close, and the field uniformity constraint forces the representations of features from different fields to be distant. Extensive experiments verify that CL4CTR achieves the best performance on four datasets and has excellent effectiveness and compatibility with various representative baselines.
A recent focal area in the space of graph neural networks (GNNs) is graph self-supervised learning (SSL), which aims to derive useful node representations without labeled data. Notably, many state-of-the-art graph SSL methods are contrastive methods, which use a combination of positive and negative samples to learn node representations. Owing to challenges in negative sampling (slowness and model sensitivity), recent literature introduced non-contrastive methods, which instead only use positive samples. Though such methods have shown promising performance in node-level tasks, their suitability for link prediction tasks, which are concerned with predicting link existence between pairs of nodes (and have broad applicability to recommendation systems contexts) is yet unexplored. In this work, we extensively evaluate the performance of existing non-contrastive methods for link prediction in both transductive and inductive settings. While most existing non-contrastive methods perform poorly overall, we find that, surprisingly, BGRL generally performs well in transductive settings. However, it performs poorly in the more realistic inductive settings where the model has to generalize to links to/from unseen nodes. We find that non-contrastive models tend to overfit to the training graph and use this analysis to propose T-BGRL, a novel non-contrastive framework that incorporates cheap corruptions to improve the generalization ability of the model. This simple modification strongly improves inductive performance in 5/6 of our datasets, with up to a 120% improvement in Hits@50--all with comparable speed to other non-contrastive baselines and up to 14x faster than the best-performing contrastive baseline. Our work imparts interesting findings about non-contrastive learning for link prediction and paves the way for future researchers to further expand upon this area.
Recently, deep neural networks have greatly advanced histopathology image segmentation but usually require abundant annotated data. However, due to the gigapixel scale of whole slide images and pathologists' heavy daily workload, obtaining pixel-level labels for supervised learning in clinical practice is often infeasible. Alternatively, weakly-supervised segmentation methods have been explored with less laborious image-level labels, but their performance is unsatisfactory due to the lack of dense supervision. Inspired by the recent success of self-supervised learning methods, we present a label-efficient tissue prototype dictionary building pipeline and propose to use the obtained prototypes to guide histopathology image segmentation. Particularly, taking advantage of self-supervised contrastive learning, an encoder is trained to project the unlabeled histopathology image patches into a discriminative embedding space where these patches are clustered to identify the tissue prototypes by efficient pathologists' visual examination. Then, the encoder is used to map the images into the embedding space and generate pixel-level pseudo tissue masks by querying the tissue prototype dictionary. Finally, the pseudo masks are used to train a segmentation network with dense supervision for better performance. Experiments on two public datasets demonstrate that our human-machine interactive tissue prototype learning method can achieve comparable segmentation performance as the fully-supervised baselines with less annotation burden and outperform other weakly-supervised methods. Codes will be available upon publication.
A hallmark of the deep learning era for computer vision is the successful use of large-scale labeled datasets to train feature representations for tasks ranging from object recognition and semantic segmentation to optical flow estimation and novel view synthesis of 3D scenes. In this work, we aim to learn dense discriminative object representations for low-shot category recognition without requiring any category labels. To this end, we propose Deep Object Patch Encodings (DOPE), which can be trained from multiple views of object instances without any category or semantic object part labels. To train DOPE, we assume access to sparse depths, foreground masks and known cameras, to obtain pixel-level correspondences between views of an object, and use this to formulate a self-supervised learning task to learn discriminative object patches. We find that DOPE can directly be used for low-shot classification of novel categories using local-part matching, and is competitive with and outperforms supervised and self-supervised learning baselines. Code and data available at https://github.com/rehg-lab/dope_selfsup.
Recently, masked image modeling (MIM), which learns visual representations by reconstructing the masked patches of an image, has dominated self-supervised learning in computer vision. However, the pre-training of MIM always takes massive time due to the large-scale data and large-size backbones. We mainly attribute it to the random patch masking in previous MIM works, which fails to leverage the crucial semantic information for effective visual representation learning. To tackle this issue, we propose the Frequency \& Attention-driven Masking and Throwing Strategy (FAMT), which can extract semantic patches and reduce the number of training patches to boost model performance and training efficiency simultaneously. Specifically, FAMT utilizes the self-attention mechanism to extract semantic information from the image for masking during training in an unsupervised manner. However, attention alone could sometimes focus on inappropriate areas regarding the semantic information. Thus, we are motivated to incorporate the information from the frequency domain into the self-attention mechanism to derive the sampling weights for masking, which captures semantic patches for visual representation learning. Furthermore, we introduce a patch throwing strategy based on the derived sampling weights to reduce the training cost. FAMT can be seamlessly integrated as a plug-and-play module and surpasses previous works, \emph{e.g.} reducing the training phase time by nearly $50\%$ and improving the linear probing accuracy of MAE by $1.3\% \sim 3.9\%$ across various datasets, including CIFAR-10/100, Tiny ImageNet, and ImageNet-1K. FAMT also demonstrates superior performance in downstream detection and segmentation tasks.
Distribution shift occurs when the test distribution differs from the training distribution, and it can considerably degrade performance of machine learning models deployed in the real world. Temporal shifts -- distribution shifts arising from the passage of time -- often occur gradually and have the additional structure of timestamp metadata. By leveraging timestamp metadata, models can potentially learn from trends in past distribution shifts and extrapolate into the future. While recent works have studied distribution shifts, temporal shifts remain underexplored. To address this gap, we curate Wild-Time, a benchmark of 5 datasets that reflect temporal distribution shifts arising in a variety of real-world applications, including patient prognosis and news classification. On these datasets, we systematically benchmark 13 prior approaches, including methods in domain generalization, continual learning, self-supervised learning, and ensemble learning. We use two evaluation strategies: evaluation with a fixed time split (Eval-Fix) and evaluation with a data stream (Eval-Stream). Eval-Fix, our primary evaluation strategy, aims to provide a simple evaluation protocol, while Eval-Stream is more realistic for certain real-world applications. Under both evaluation strategies, we observe an average performance drop of 20% from in-distribution to out-of-distribution data. Existing methods are unable to close this gap. Code is available at https://wild-time.github.io/.
Although substantial efforts have been made using graph neural networks (GNNs) for AI-driven drug discovery (AIDD), effective molecular representation learning remains an open challenge, especially in the case of insufficient labeled molecules. Recent studies suggest that big GNN models pre-trained by self-supervised learning on unlabeled datasets enable better transfer performance in downstream molecular property prediction tasks. However, the approaches in these studies require multiple complex self-supervised tasks and large-scale datasets, which are time-consuming, computationally expensive, and difficult to pre-train end-to-end. Here, we design a simple yet effective self-supervised strategy to simultaneously learn local and global information about molecules, and further propose a novel bi-branch masked graph transformer autoencoder (BatmanNet) to learn molecular representations. BatmanNet features two tailored complementary and asymmetric graph autoencoders to reconstruct the missing nodes and edges, respectively, from a masked molecular graph. With this design, BatmanNet can effectively capture the underlying structure and semantic information of molecules, thus improving the performance of molecular representation. BatmanNet achieves state-of-the-art results for multiple drug discovery tasks, including molecular properties prediction, drug-drug interaction, and drug-target interaction, on 13 benchmark datasets, demonstrating its great potential and superiority in molecular representation learning.
We aim to learn to temporally localize object state changes and the corresponding state-modifying actions by observing people interacting with objects in long uncurated web videos. We introduce three principal contributions. First, we explore alternative multi-task network architectures and identify a model that enables efficient joint learning of multiple object states and actions such as pouring water and pouring coffee. Second, we design a multi-task self-supervised learning procedure that exploits different types of constraints between objects and state-modifying actions enabling end-to-end training of a model for temporal localization of object states and actions in videos from only noisy video-level supervision. Third, we report results on the large-scale ChangeIt and COIN datasets containing tens of thousands of long (un)curated web videos depicting various interactions such as hole drilling, cream whisking, or paper plane folding. We show that our multi-task model achieves a relative improvement of 40% over the prior single-task methods and significantly outperforms both image-based and video-based zero-shot models for this problem. We also test our method on long egocentric videos of the EPIC-KITCHENS and the Ego4D datasets in a zero-shot setup demonstrating the robustness of our learned model.
Self-supervised facial representation has recently attracted increasing attention due to its ability to perform face understanding without relying on large-scale annotated datasets heavily. However, analytically, current contrastive-based self-supervised learning (SSL) still performs unsatisfactorily for learning facial representation. More specifically, existing contrastive learning (CL) tends to learn pose-invariant features that cannot depict the pose details of faces, compromising the learning performance. To conquer the above limitation of CL, we propose a novel Pose-disentangled Contrastive Learning (PCL) method for general self-supervised facial representation. Our PCL first devises a pose-disentangled decoder (PDD) with a delicately designed orthogonalizing regulation, which disentangles the pose-related features from the face-aware features; therefore, pose-related and other pose-unrelated facial information could be performed in individual subnetworks and do not affect each other's training. Furthermore, we introduce a pose-related contrastive learning scheme that learns pose-related information based on data augmentation of the same image, which would deliver more effective face-aware representation for various downstream tasks. We conducted linear evaluation on four challenging downstream facial understanding tasks, ie, facial expression recognition, face recognition, AU detection and head pose estimation. Experimental results demonstrate that our method significantly outperforms state-of-the-art SSL methods. Code is available at https://github.com/DreamMr/PCL}{https://github.com/DreamMr/PCL
Recent advances have indicated the strengths of self-supervised pre-training for improving representation learning on downstream tasks. Existing works often utilize self-supervised pre-trained models by fine-tuning on downstream tasks. However, fine-tuning does not generalize to the case when one needs to build a customized model architecture different from the self-supervised model. In this work, we formulate a new knowledge distillation framework to transfer the knowledge from self-supervised pre-trained models to any other student network by a novel approach named Embedding Graph Alignment. Specifically, inspired by the spirit of instance discrimination in self-supervised learning, we model the instance-instance relations by a graph formulation in the feature embedding space and distill the self-supervised teacher knowledge to a student network by aligning the teacher graph and the student graph. Our distillation scheme can be flexibly applied to transfer the self-supervised knowledge to enhance representation learning on various student networks. We demonstrate that our model outperforms multiple representative knowledge distillation methods on three benchmark datasets, including CIFAR100, STL10, and TinyImageNet. Code is here: https://github.com/yccm/EGA.
Image inpainting is the process of regenerating lost parts of the image. Supervised algorithm-based methods have shown excellent results but have two significant drawbacks. They do not perform well when tested with unseen data. They fail to capture the global context of the image, resulting in a visually unappealing result. We propose a novel self-supervised learning framework for image-inpainting: Weighted Self-Supervised Learning (WSSL) to tackle these problems. We designed WSSL to learn features from multiple weighted pretext tasks. These features are then utilized for the downstream task, image-inpainting. To improve the performance of our framework and produce more visually appealing images, we also present a novel loss function for image inpainting. The loss function takes advantage of both reconstruction loss and perceptual loss functions to regenerate the image. Our experimentation shows WSSL outperforms previous methods, and our loss function helps produce better results.
Transformers, which were originally developed for natural language processing, have recently generated significant interest in the computer vision and audio communities due to their flexibility in learning long-range relationships. Constrained by the data hungry nature of transformers and the limited amount of labelled data, most transformer-based models for audio tasks are finetuned from ImageNet pretrained models, despite the huge gap between the domain of natural images and audio. This has motivated the research in self-supervised pretraining of audio transformers, which reduces the dependency on large amounts of labeled data and focuses on extracting concise representations of audio spectrograms. In this paper, we propose \textbf{L}ocal-\textbf{G}lobal \textbf{A}udio \textbf{S}pectrogram v\textbf{I}sion \textbf{T}ransformer, namely ASiT, a novel self-supervised learning framework that captures local and global contextual information by employing group masked model learning and self-distillation. We evaluate our pretrained models on both audio and speech classification tasks, including audio event classification, keyword spotting, and speaker identification. We further conduct comprehensive ablation studies, including evaluations of different pretraining strategies. The proposed ASiT framework significantly boosts the performance on all tasks and sets a new state-of-the-art performance in five audio and speech classification tasks, outperforming recent methods, including the approaches that use additional datasets for pretraining.
Existing Graph Neural Networks (GNNs) follow the message-passing mechanism that conducts information interaction among nodes iteratively. While considerable progress has been made, such node interaction paradigms still have the following limitation. First, the scalability limitation precludes the broad application of GNNs in large-scale industrial settings since the node interaction among rapidly expanding neighbors incurs high computation and memory costs. Second, the over-smoothing problem restricts the discrimination ability of nodes, i.e., node representations of different classes will converge to indistinguishable after repeated node interactions. In this work, we propose a novel hop interaction paradigm to address these limitations simultaneously. The core idea is to convert the interaction target among nodes to pre-processed multi-hop features inside each node. We design a simple yet effective HopGNN framework that can easily utilize existing GNNs to achieve hop interaction. Furthermore, we propose a multi-task learning strategy with a self-supervised learning objective to enhance HopGNN. We conduct extensive experiments on 12 benchmark datasets in a wide range of domains, scales, and smoothness of graphs. Experimental results show that our methods achieve superior performance while maintaining high scalability and efficiency. The code is at https://github.com/JC-202/HopGNN.
Generalized Zero-Shot Learning (GZSL) aims to train a classifier that can generalize to unseen classes, using a set of attributes as auxiliary information, and the visual features extracted from a pre-trained convolutional neural network. While recent GZSL methods have explored various techniques to leverage the capacity of these features, there has been an extensive growth of representation learning techniques that remain under-explored. In this work, we investigate the utility of different GZSL methods when using different feature extractors, and examine how these models' pre-training objectives, datasets, and architecture design affect their feature representation ability. Our results indicate that 1) methods using generative components for GZSL provide more advantages when using recent feature extractors; 2) feature extractors pre-trained using self-supervised learning objectives and knowledge distillation provide better feature representations, increasing up to 15% performance when used with recent GZSL techniques; 3) specific feature extractors pre-trained with larger datasets do not necessarily boost the performance of GZSL methods. In addition, we investigate how GZSL methods fare against CLIP, a more recent multi-modal pre-trained model with strong zero-shot performance. We found that GZSL tasks still benefit from generative-based GZSL methods along with CLIP's internet-scale pre-training to achieve state-of-the-art performance in fine-grained datasets. We release a modular framework for analyzing representation learning issues in GZSL here: https://github.com/uvavision/TV-GZSL
We integrate contrastive learning (CL) with adversarial learning to co-optimize the robustness and accuracy of code models. Different from existing works, we show that code obfuscation, a standard code transformation operation, provides novel means to generate complementary `views' of a code that enable us to achieve both robust and accurate code models. To the best of our knowledge, this is the first systematic study to explore and exploit the robustness and accuracy benefits of (multi-view) code obfuscations in code models. Specifically, we first adopt adversarial codes as robustness-promoting views in CL at the self-supervised pre-training phase. This yields improved robustness and transferability for downstream tasks. Next, at the supervised fine-tuning stage, we show that adversarial training with a proper temporally-staggered schedule of adversarial code generation can further improve robustness and accuracy of the pre-trained code model. Built on the above two modules, we develop CLAWSAT, a novel self-supervised learning (SSL) framework for code by integrating $\underline{\textrm{CL}}$ with $\underline{\textrm{a}}$dversarial vie$\underline{\textrm{w}}$s (CLAW) with $\underline{\textrm{s}}$taggered $\underline{\textrm{a}}$dversarial $\underline{\textrm{t}}$raining (SAT). On evaluating three downstream tasks across Python and Java, we show that CLAWSAT consistently yields the best robustness and accuracy ($\textit{e.g.}$ 11$\%$ in robustness and 6$\%$ in accuracy on the code summarization task in Python). We additionally demonstrate the effectiveness of adversarial learning in CLAW by analyzing the characteristics of the loss landscape and interpretability of the pre-trained models.
Drawings are powerful means of pictorial abstraction and communication. Understanding diverse forms of drawings, including digital arts, cartoons, and comics, has been a major problem of interest for the computer vision and computer graphics communities. Although there are large amounts of digitized drawings from comic books and cartoons, they contain vast stylistic variations, which necessitate expensive manual labeling for training domain-specific recognizers. In this work, we show how self-supervised learning, based on a teacher-student network with a modified student network update design, can be used to build face and body detectors. Our setup allows exploiting large amounts of unlabeled data from the target domain when labels are provided for only a small subset of it. We further demonstrate that style transfer can be incorporated into our learning pipeline to bootstrap detectors using a vast amount of out-of-domain labeled images from natural images (i.e., images from the real world). Our combined architecture yields detectors with state-of-the-art (SOTA) and near-SOTA performance using minimal annotation effort. Our code can be accessed from https://github.com/barisbatuhan/DASS_Detector.
Masked Video Autoencoder (MVA) approaches have demonstrated their potential by significantly outperforming previous video representation learning methods. However, they waste an excessive amount of computations and memory in predicting uninformative tokens/frames due to random masking strategies. (e.g., over 16 nodes with 128 NVIDIA A100 GPUs). To resolve this issue, we exploit the unequal information density among the patches in videos and propose EVEREST, a surprisingly efficient MVA approach for video representation learning that finds tokens containing rich motion features and discards uninformative ones during both pre-training and fine-tuning. We further present an information-intensive frame selection strategy that allows the model to focus on informative and causal frames with minimal redundancy. Our method significantly reduces the computation and memory requirements of MVA, enabling the pre-training and fine-tuning on a single machine with 8 GPUs while achieving comparable performance to computation- and memory-heavy baselines on multiple benchmarks and the uncurated Ego4D dataset. We hope that our work contributes to reducing the barrier to further research on video understanding.