Contrastive learning has revolutionized the field of computer vision, learning rich representations from unlabeled data, which generalize well to diverse vision tasks. Consequently, it has become increasingly important to explain these approaches and understand their inner workings mechanisms. Given that contrastive models are trained with interdependent and interacting inputs and aim to learn invariance through data augmentation, the existing methods for explaining single-image systems (e.g., image classification models) are inadequate as they fail to account for these factors and typically assume independent inputs. Additionally, there is a lack of evaluation metrics designed to assess pairs of explanations, and no analytical studies have been conducted to investigate the effectiveness of different techniques used to explaining contrastive learning. In this work, we design visual explanation methods that contribute towards understanding similarity learning tasks from pairs of images. We further adapt existing metrics, used to evaluate visual explanations of image classification systems, to suit pairs of explanations and evaluate our proposed methods with these metrics. Finally, we present a thorough analysis of visual explainability methods for contrastive learning, establish their correlation with downstream tasks and demonstrate the potential of our approaches to investigate their merits and drawbacks.
The Chinese University of HongKong, Shenzhen（香港中文大学（深圳））；JD AI Research（京东人工智能研究院）；Tongji University（同济大学）；The Hong Kong Polytechnic University（香港理工大学）
Contrastive learning, which aims at minimizing the distance between positive pairs while maximizing that of negative ones, has been widely and successfully applied in unsupervised feature learning, where the design of positive and negative (pos/neg) pairs is one of its keys. In this paper, we attempt to devise a feature-level data manipulation, differing from data augmentation, to enhance the generic contrastive self-supervised learning. To this end, we first design a visualization scheme for pos/neg score (Pos/neg score indicates cosine similarity of pos/neg pair.) distribution, which enables…
Large Convolutional Network models have recently demonstrated impressive classification performance on the ImageNet benchmark. However there is no clear understanding of why they perform so well, or how they might be improved. In this paper we address both issues. We introduce a novel visualization technique that gives insight into the function of intermediate feature layers and the operation of the classifier. We also perform an ablation study to discover the performance contribution from different model layers. This enables us to find model architectures that outperform Krizhevsky \etal on t…
Image and Point Clouds provide different information for robots. Finding the correspondences between data from different sensors is crucial for various tasks such as localization, mapping, and navigation. Learning-based descriptors have been developed for single sensors; there is little work on cross-modal features. This work treats learning cross-modal features as a dense contrastive learning problem. We propose a Tuple-Circle loss function for cross-modality feature learning. Furthermore, to learn good features and not lose generality, we developed a variant of widely used PointNet++ archite…
University of Massachusetts, Amherst（马萨诸塞大学阿默斯特分校）
A number of recent approaches have used deep convolutional neural networks (CNNs) to build texture representations. Nevertheless, it is still unclear how these models represent texture and invariances to categorical variations. This work conducts a systematic evaluation of recent CNN-based texture descriptors for recognition and attempts to understand the nature of invariances captured by these representations. First we show that the recently proposed bilinear CNN model [25] is an excellent general-purpose texture descriptor and compares favorably to other CNN-based descriptors on various text…
Understanding and interpreting the decisions made by deep learning models is valuable in many domains. In computer vision, computing heatmaps from a deep network is a popular approach for visualizing and understanding deep networks. However, heatmaps that do not correlate with the network may mislead human, hence the performance of heatmaps in providing a faithful explanation to the underlying deep network is crucial. In this paper, we propose I-GOS, which optimizes for a heatmap so that the classification scores on the masked image would maximally decrease. The main novelty of the approach is…
Guangxi Normal University（广西师范大学）；Huaqiao University（华侨大学）；JD Explore Academy（京东探索研究院）
Vision Transformer (ViT) has become a leading tool in various computer vision tasks, owing to its unique self-attention mechanism that learns visual representations explicitly through cross-patch information interactions. Despite having good success, the literature seldom explores the explainability of vision transformer, and there is no clear picture of how the attention mechanism with respect to the correlation across comprehensive patches will impact the performance and what is the further potential. In this work, we propose a novel explainable visualization approach to analyze and interpre…
Center for Research in Computer Vision（计算机视觉研究中心）；University of Exeter（埃克塞特大学）；University of Central Florida（中佛罗里达大学）
Deep convolutional neural networks have demonstrated high performances for fixation prediction in recent years. How they achieve this, however, is less explored and they remain to be black box models. Here, we attempt to shed light on the internal structure of deep saliency models and study what features they extract for fixation prediction. Specifically, we use a simple yet powerful architecture, consisting of only one CNN and a single resolution input, combined with a new loss function for pixel-wise fixation prediction during free viewing of natural scenes. We show that our simple method is…
We introduce algorithms to visualize feature spaces used by object detectors. Our method works by inverting a visual feature back to multiple natural images. We found that these visualizations allow us to analyze object detection systems in new ways and gain new insight into the detector's failures. For example, when we visualize the features for high scoring false alarms, we discovered that, although they are clearly wrong in image space, they do look deceptively similar to true positives in feature space. This result suggests that many of these false alarms are caused by our choice of featur…
How the human vision system addresses the object identity-preserving recognition problem is largely unknown. Here, we use a vision recognition-reconstruction network (RRN) to investigate the development, recognition, learning and forgetting mechanisms, and achieve similar characteristics to electrophysiological measurements in monkeys. First, in network development study, the RRN also experiences critical developmental stages characterized by specificities in neuron types, synapse and activation patterns, and visual task performance from the early stage of coarse salience map recognition to ma…
We present a new approach to instill 4D dynamic object priors into learned 3D representations by unsupervised pre-training. We observe that dynamic movement of an object through an environment provides important cues about its objectness, and thus propose to imbue learned 3D representations with such dynamic understanding, that can then be effectively transferred to improved performance in downstream 3D semantic scene understanding tasks. We propose a new data augmentation scheme leveraging synthetic 3D shapes moving in static 3D environments, and employ contrastive learning under 3D-4D constr…
University of Exeter（埃克塞特大学）；Aalto University（阿尔托大学）；MarkableAI（MarkableAI公司）
Recently, data-driven deep saliency models have achieved high performance and have outperformed classical saliency models, as demonstrated by results on datasets such as the MIT300 and SALICON. Yet, there remains a large gap between the performance of these models and the inter-human baseline. Some outstanding questions include what have these models learned, how and where they fail, and how they can be improved. This article attempts to answer these questions by analyzing the representations learned by individual neurons located at the intermediate layers of deep saliency models. To this end,…
City University of Hong Kong（香港城市大学）；Microsoft Research Asia（微软亚洲研究院）
Recent progress in contrastive learning has revolutionized unsupervised representation learning. Concretely, multiple views (augmentations) from the same image are encouraged to map to the similar embeddings, while views from different images are pulled apart. In this paper, through visualizing and diagnosing classification errors, we observe that current contrastive models are ineffective at localizing the foreground object, limiting their ability to extract discriminative high-level features. This is due to the fact that view generation process considers pixels in an image uniformly. To addr…
Technical University of Munich（慕尼黑工业大学）；Facebook AI Research（脸书人工智能研究院）
The rapid progress in 3D scene understanding has come with growing demand for data; however, collecting and annotating 3D scenes (e.g. point clouds) are notoriously hard. For example, the number of scenes (e.g. indoor rooms) that can be accessed and scanned might be limited; even given sufficient data, acquiring 3D labels (e.g. instance masks) requires intensive human labor. In this paper, we explore data-efficient learning for 3D point cloud. As a first step towards this direction, we propose Contrastive Scene Contexts, a 3D pre-training method that makes use of both point-level correspondenc…
Vision-Language Pre-training with Object Contrastive Learning for 3D Scene Understanding
Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；College of Computer Science and Software Engineering, Shenzhen University（深圳大学计算机与软件学院）；Harbin Institute of Technology, Shenzhen（哈尔滨工业大学（深圳））；Research Center of Artificial Intelligence, Peng Cheng Laboratory（鹏城实验室人工智能研究中心）
Affinity Contrastive Learning for Skeleton-based Human Activity Understanding
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；School of Electronic, Electrical and Communication Engineering, Univeristy of Chinese Academy of Sciences（中国科学院大学电子电气与通信工程学院）；School of Artificial Intelligence, Beijing University of Posts and Telecommunications（北京邮电大学人工智能学院）；Moonshot AI
Carnegie Mellon University（卡内基梅隆大学）；University of Michigan（密歇根大学）；Georgia Tech（佐治亚理工学院）；HKUST（香港科技大学）
The promise of multimodal models for real-world applications has inspired research in visualizing and understanding their internal mechanics with the end goal of empowering stakeholders to visualize model behavior, perform model debugging, and promote trust in machine learning models. However, modern multimodal models are typically black-box neural networks, which makes it challenging to understand their internal mechanics. How can we visualize the internal modeling of multimodal interactions in these models? Our paper aims to fill this gap by proposing MultiViz, a method for analyzing the beh…
Deep neural network (DNN) models have recently obtained state-of-the-art prediction accuracy for the transcription factor binding (TFBS) site classification task. However, it remains unclear how these approaches identify meaningful DNA sequence signals and give insights as to why TFs bind to certain locations. In this paper, we propose a toolkit called the Deep Motif Dashboard (DeMo Dashboard) which provides a suite of visualization strategies to extract motifs, or sequence patterns from deep neural network models for TFBS classification. We demonstrate how to visualize and understand three im…
We present VideoCLIP, a contrastive approach to pre-train a unified model for zero-shot video and text understanding, without using any labels on downstream tasks. VideoCLIP trains a transformer for video and text by contrasting temporally overlapping positive video-text pairs with hard negatives from nearest neighbor retrieval. Our experiments on a diverse series of downstream tasks, including sequence-level text-video retrieval, VideoQA, token-level action localization, and action segmentation reveal state-of-the-art performance, surpassing prior work, and in some cases even outperforming su…
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）
Visual Dialog is a challenging vision-language task since the visual dialog agent needs to answer a series of questions after reasoning over both the image content and dialog history. Though existing methods try to deal with the cross-modal understanding in visual dialog, they are still not enough in ranking candidate answers based on their understanding of visual and textual contexts. In this paper, we analyze the cross-modal understanding in visual dialog based on the vision-language pre-training model VD-BERT and propose a novel approach to improve the cross-modal understanding for visual d…
Massachusetts Institute of Technology（麻省理工学院）；MIT-IBM Watson AI Lab（MIT-IBM沃森人工智能实验室）；IBM Research（IBM研究院）；The Chinese University of Hong Kong（香港中文大学）
Generative Adversarial Networks (GANs) have recently achieved impressive results for many real-world applications, and many GAN variants have emerged with improvements in sample quality and training stability. However, they have not been well visualized or understood. How does a GAN represent our visual world internally? What causes the artifacts in GAN results? How do architectural choices affect GAN learning? Answering such questions could enable us to develop new insights and better models. In this work, we present an analytic framework to visualize and understand GANs at the unit-, object-…
MIT Computer Science & Artificial Intelligence Lab（麻省理工学院计算机科学与人工智能实验室）
Contrastive representation learning has been outstandingly successful in practice. In this work, we identify two key properties related to the contrastive loss: (1) alignment (closeness) of features from positive pairs, and (2) uniformity of the induced distribution of the (normalized) features on the hypersphere. We prove that, asymptotically, the contrastive loss optimizes these properties, and analyze their positive effects on downstream tasks. Empirically, we introduce an optimizable metric to quantify each property. Extensive experiments on standard vision and language datasets confirm th…
A recent popular approach to out-of-distribution (OOD) detection is based on a self-supervised learning technique referred to as contrastive learning. There are two main variants of contrastive learning, namely instance and class discrimination, targeting features that can discriminate between different instances for the former, and different classes for the latter. In this paper, we aim to understand the effectiveness and limitation of existing contrastive learning methods for OOD detection. We approach this in 3 ways. First, we systematically study the performance difference between the inst…
Montreal Institute for Learning Algorithms（蒙特利尔学习算法研究所）；ConscientAI Labs（ConscientAI实验室）；Microsoft Research New England（微软新英格兰研究院）
We present a project that aims to generate images that depict accurate, vivid, and personalized outcomes of climate change using Cycle-Consistent Adversarial Networks (CycleGANs). By training our CycleGAN model on street-view images of houses before and after extreme weather events (e.g. floods, forest fires, etc.), we learn a mapping that can then be applied to images of locations that have not yet experienced these events. This visual transformation is paired with climate model predictions to assess likelihood and type of climate-related events in the long term (50 years) in order to bring t…
TreeGaussian: Tree-Guided Cascaded Contrastive Learning for Hierarchical Consistent 3D Gaussian Scene Segmentation and Understanding
Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Beihang University（北京航空航天大学）
Recently, contrastive learning has risen to be a promising approach for large-scale self-supervised learning. However, theoretical understanding of how it works is still unclear. In this paper, we propose a new guarantee on the downstream performance without resorting to the conditional independence assumption that is widely adopted in previous work but hardly holds in practice. Our new theory hinges on the insight that the support of different intra-class samples will become more overlapped under aggressive data augmentations, thus simply aligning the positive samples (augmented views of the …
We show that Contrastive Learning (CL) under a broad family of loss functions (including InfoNCE) has a unified formulation of coordinate-wise optimization on the network parameter $\boldsymbolθ$ and pairwise importance $α$, where the \emph{max player} $\boldsymbolθ$ learns representation for contrastiveness, and the \emph{min player} $α$ puts more weights on pairs of distinct samples that share similar representations. The resulting formulation, called $α$-CL, unifies not only various existing contrastive losses, which differ by how sample-pair importance $α$ is constructed, but also is able …
Self-supervised visual representation learning aims to learn useful representations without relying on human annotations. Joint embedding approach bases on maximizing the agreement between embedding vectors from different views of the same image. Various methods have been proposed to solve the collapsing problem where all embedding vectors collapse to a trivial constant solution. Among these methods, contrastive learning prevents collapse via negative sample pairs. It has been shown that non-contrastive methods suffer from a lesser collapse problem of a different nature: dimensional collapse, …
Monash University（莫纳什大学）；Defence Science and Technology Group, Australia（澳大利亚国防科技集团）；CSIRO Data61, Australia（澳大利亚联邦科学与工业研究组织Data61）
Contrastive learning (CL) has recently emerged as an effective approach to learning representation in a range of downstream tasks. Central to this approach is the selection of positive (similar) and negative (dissimilar) sets to provide the model the opportunity to `contrast' between data and class representation in the latent space. In this paper, we investigate CL for improving model robustness using adversarial samples. We first designed and performed a comprehensive study to understand how adversarial vulnerability behaves in the latent space. Based on this empirical evidence, we propose a…
Adversarial Training (AT) is known as an effective approach to enhance the robustness of deep neural networks. Recently researchers notice that robust models with AT have good generative ability and can synthesize realistic images, while the reason behind it is yet under-explored. In this paper, we demystify this phenomenon by developing a unified probabilistic framework, called Contrastive Energy-based Models (CEM). On the one hand, we provide the first probabilistic characterization of AT through a unified understanding of robustness and generative ability. On the other hand, our unified fra…
Facebook AI Research（脸书人工智能研究院）；Stanford University（斯坦福大学）
While contrastive approaches of self-supervised learning (SSL) learn representations by minimizing the distance between two augmented views of the same data point (positive pairs) and maximizing views from different data points (negative pairs), recent \emph{non-contrastive} SSL (e.g., BYOL and SimSiam) show remarkable performance {\it without} negative pairs, with an extra learnable predictor and a stop-gradient operation. A fundamental question arises: why do these methods not collapse into trivial representations? We answer this question via a simple theoretical study and propose a novel ap…
How can neural networks trained by contrastive learning extract features from the unlabeled data? Why does contrastive learning usually need much stronger data augmentations than supervised learning to ensure good representations? These questions involve both the optimization and statistical aspects of deep learning, but can hardly be answered by analyzing supervised learning, where the target functions are the highest pursuit. Indeed, in self-supervised learning, it is inevitable to relate to the optimization/generalization of neural networks to how they can encode the latent structures in th…
In recent years, contrastive learning achieves impressive results on self-supervised visual representation learning, but there still lacks a rigorous understanding of its learning dynamics. In this paper, we show that if we cast a contrastive objective equivalently into the feature space, then its learning dynamics admits an interpretable form. Specifically, we show that its gradient descent corresponds to a specific message passing scheme on the corresponding augmentation graph. Based on this perspective, we theoretically characterize how contrastive learning gradually learns discriminative f…
Carnegie Mellon University（卡内基梅隆大学）；University of California, Berkeley（加州大学伯克利分校）
Contrastive methods have led a recent surge in the performance of self-supervised representation learning (SSL). Recent methods like BYOL or SimSiam purportedly distill these contrastive methods down to their essence, removing bells and whistles, including the negative examples, that do not contribute to downstream performance. These "non-contrastive" methods work surprisingly well without using negatives even though the global minimum lies at trivial collapse. We empirically analyze these non-contrastive methods and find that SimSiam is extraordinarily sensitive to dataset and model size. In …
Multiple modalities for certain information provide a variety of perspectives on that information, which can improve the understanding of the information. Thus, it may be crucial to generate data of different modality from the existing data to enhance the understanding. In this paper, we investigate the cross-modal audio-to-image generation problem and propose Cross-Modal Contrastive Representation Learning (CMCRL) to extract useful features from audios and use it in the generation phase. Experimental results show that CMCRL enhances quality of images generated than previous research.
We present modality gap, an intriguing geometric phenomenon of the representation space of multi-modal models. Specifically, we show that different data modalities (e.g. images and text) are embedded at arm's length in their shared representation in multi-modal models such as CLIP. Our systematic analysis demonstrates that this gap is caused by a combination of model initialization and contrastive learning optimization. In model initialization, we show empirically and theoretically that the representation of a common deep neural network is restricted to a narrow cone. As a consequence, in a mu…
