2022-07-29 更新
Time to augment contrastive learning
Authors:Arthur Aubret, Markus Ernst, Céline Teulière, Jochen Triesch
Biological vision systems are unparalleled in their ability to learn visual representations without supervision. In machine learning, contrastive learning (CL) has led to major advances in forming object representations in an unsupervised fashion. These systems learn representations invariant to augmentation operations over images, like cropping or flipping. In contrast, biological vision systems exploit the temporal structure of the visual experience. This gives access to augmentations not commonly used in CL, like watching the same object from multiple viewpoints or against different backgrounds. Here, we systematically investigate and compare the potential benefits of such time-based augmentations for learning object categories. Our results show that time-based augmentations achieve large performance gains over state-of-the-art image augmentations. Specifically, our analyses reveal that: 1) 3-D object rotations drastically improve the learning of object categories; 2) viewing objects against changing backgrounds is vital for learning to discard background-related information. Overall, we conclude that time-based augmentations can greatly improve contrastive learning, narrowing the gap between artificial and biological vision systems.
PDF
点此查看论文截图
VICTOR: Visual Incompatibility Detection with Transformers and Fashion-specific contrastive pre-training
Authors:Stefanos-Iordanis Papadopoulos, Christos Koutlis, Symeon Papadopoulos, Ioannis Kompatsiaris
In order to consider fashion outfits as aesthetically pleasing, the garments that constitute them need to be compatible in terms of visual aspects, such as style, category and color. With the advent and omnipresence of computer vision deep learning models, increased interest has also emerged for the task of visual compatibility detection with the aim to develop quality fashion outfit recommendation systems. Previous works have defined visual compatibility as a binary classification task with items in a garment being considered as fully compatible or fully incompatible. However, this is not applicable to Outfit Maker applications where users create their own outfits and need to know which specific items may be incompatible with the rest of the outfit. To address this, we propose the Visual InCompatibility TransfORmer (VICTOR) that is optimized for two tasks: 1) overall compatibility as regression and 2) the detection of mismatching items. Unlike previous works that either rely on feature extraction from ImageNet-pretrained models or by end-to-end fine tuning, we utilize fashion-specific contrastive language-image pre-training for fine tuning computer vision neural networks on fashion imagery. Moreover, we build upon the Polyvore outfit benchmark to generate partially mismatching outfits, creating a new dataset termed Polyvore-MISFITs, that is used to train VICTOR. A series of ablation and comparative analyses show that the proposed architecture can compete and even surpass the current state-of-the-art on Polyvore datasets while reducing the instance-wise floating operations by 88%, striking a balance between high performance and efficiency.
PDF 14 pages, 5 figures
点此查看论文截图
Contrastive Attraction and Contrastive Repulsion for Representation Learning
Authors:Huangjie Zheng, Xu Chen, Jiangchao Yao, Hongxia Yang, Chunyuan Li, Ya Zhang, Hao Zhang, Ivor Tsang, Jingren Zhou, Mingyuan Zhou
Contrastive learning (CL) methods effectively learn data representations without label supervision, where the encoder contrasts each positive sample over multiple negative samples via a one-vs-many softmax cross-entropy loss. By leveraging large amounts of unlabeled image data, recent CL methods have achieved promising results when pre-trained on ImageNet, a well-curated data set with balanced image classes. However, they tend to yield worse performance when pre-trained on images in the wild. In this paper, to further improve the performance of CL and enhance its robustness on uncurated data sets, we propose a doubly CL strategy that contrasts the positive (negative) samples of a query within themselves before deciding how strongly to pull (push) them. We realize this strategy with contrastive attraction and contrastive repulsion (CACR), which makes the query not only exert a greater force to attract more distant positive samples but also do so to repel closer negative samples. Theoretical analysis reveals that CACR generalizes CL’s behavior by taking into consideration the differences between the distributions of the positive/negative samples, which in general are sampled independently of the query, and their true conditional distributions given the query. We demonstrate this unique intra-positive attraction and intra-negative repulsion mechanism, which helps remove the need to assume uniform prior distributions on both the data and their latent representation, is particularly beneficial when data sets are less curated. Extensive large-scale experiments on a number of standard vision tasks show that CACR not only consistently outperforms existing CL methods on benchmark data sets in representation learning, but also shows better robustness when generalized to pre-training on imbalanced image data sets.
PDF
点此查看论文截图
Learning Visual Representation from Modality-Shared Contrastive Language-Image Pre-training
Authors:Haoxuan You, Luowei Zhou, Bin Xiao, Noel Codella, Yu Cheng, Ruochen Xu, Shih-Fu Chang, Lu Yuan
Large-scale multi-modal contrastive pre-training has demonstrated great utility to learn transferable features for a range of downstream tasks by mapping multiple modalities into a shared embedding space. Typically, this has employed separate encoders for each modality. However, recent work suggests that transformers can support learning across multiple modalities and allow knowledge sharing. Inspired by this, we investigate a variety of Modality-Shared Contrastive Language-Image Pre-training (MS-CLIP) frameworks. More specifically, we question how many parameters of a transformer model can be shared across modalities during contrastive pre-training, and rigorously examine architectural design choices that position the proportion of parameters shared along a spectrum. In studied conditions, we observe that a mostly unified encoder for vision and language signals outperforms all other variations that separate more parameters. Additionally, we find that light-weight modality-specific parallel modules further improve performance. Experimental results show that the proposed MS-CLIP approach outperforms vanilla CLIP by up to 13\% relative in zero-shot ImageNet classification (pre-trained on YFCC-100M), while simultaneously supporting a reduction of parameters. In addition, our approach outperforms vanilla CLIP by 1.6 points in linear probing on a collection of 24 downstream vision tasks. Furthermore, we discover that sharing parameters leads to semantic concepts from different modalities being encoded more closely in the embedding space, facilitating the transferring of common semantic structure (e.g., attention patterns) from language to vision. Code is available at \href{https://github.com/Hxyou/MSCLIP}{URL}.
PDF Accepted by ECCV 2022, 22 pages, 4 figures
点此查看论文截图
Contrastive Masked Autoencoders are Stronger Vision Learners
Authors:Zhicheng Huang, Xiaojie Jin, Chengze Lu, Qibin Hou, Ming-Ming Cheng, Dongmei Fu, Xiaohui Shen, Jiashi Feng
Masked image modeling (MIM) has achieved promising results on various vision tasks. However, the limited discriminability of learned representation manifests there is still plenty to go for making a stronger vision learner. Towards this goal, we propose Contrastive Masked Autoencoders (CMAE), a new self-supervised pre-training method for learning more comprehensive and capable vision representations. By elaboratively unifying contrastive learning (CL) and masked image model (MIM) through novel designs, CMAE leverages their respective advantages and learns representations with both strong instance discriminability and local perceptibility. Specifically, CMAE consists of two branches where the online branch is an asymmetric encoder-decoder and the target branch is a momentum updated encoder. During training, the online encoder reconstructs original images from latent representations of masked images to learn holistic features. The target encoder, fed with the full images, enhances the feature discriminability via contrastive learning with its online counterpart. To make CL compatible with MIM, CMAE introduces two new components, i.e. pixel shift for generating plausible positive views and feature decoder for complementing features of contrastive pairs. Thanks to these novel designs, CMAE effectively improves the representation quality and transfer performance over its MIM counterpart. CMAE achieves the state-of-the-art performance on highly competitive benchmarks of image classification, semantic segmentation and object detection. Notably, CMAE-Base achieves $85.3\%$ top-1 accuracy on ImageNet and $52.5\%$ mIoU on ADE20k, surpassing previous best results by $0.7\%$ and $1.8\%$ respectively. Codes will be made publicly available.
PDF
点此查看论文截图
Contrastive Image Synthesis and Self-supervised Feature Adaptation for Cross-Modality Biomedical Image Segmentation
Authors:Xinrong Hu, Corey Wang, Yiyu Shi
This work presents a novel framework CISFA (Contrastive Image synthesis and Self-supervised Feature Adaptation)that builds on image domain translation and unsupervised feature adaptation for cross-modality biomedical image segmentation. Different from existing works, we use a one-sided generative model and add a weighted patch-wise contrastive loss between sampled patches of the input image and the corresponding synthetic image, which serves as shape constraints. Moreover, we notice that the generated images and input images share similar structural information but are in different modalities. As such, we enforce contrastive losses on the generated images and the input images to train the encoder of a segmentation model to minimize the discrepancy between paired images in the learned embedding space. Compared with existing works that rely on adversarial learning for feature adaptation, such a method enables the encoder to learn domain-independent features in a more explicit way. We extensively evaluate our methods on segmentation tasks containing CT and MRI images for abdominal cavities and whole hearts. Experimental results show that the proposed framework not only outputs synthetic images with less distortion of organ shapes, but also outperforms state-of-the-art domain adaptation methods by a large margin.
PDF
点此查看论文截图