There is an increasing number of real-world problems in computer vision and machine learning requiring to take into consideration multiple interpretation layers (modalities or views) of the world and learn how they relate to each other. For example, in the case of Earth Observations from satellite data, it is important to be able to predict one observation layer (e.g. vegetation index) from other layers (e.g. water vapor, snow cover, temperature etc), in order to best understand how the Earth System functions and also be able to reliably predict information for one layer when the data is missing (e.g. due to measurement failure or error).
Institute of Software Chinese Academy of Sciences（中国科学院软件研究所）；University of Chinese Academy of Sciences（中国科学院大学）；JD Explore Academy（京东探索研究院）；Renmin University of China（中国人民大学）
Recently, significant progress has been made in masked image modeling to catch up to masked language modeling. However, unlike words in NLP, the lack of semantic decomposition of images still makes masked autoencoding (MAE) different between vision and language. In this paper, we explore a potential visual analogue of words, i.e., semantic parts, and we integrate semantic information into the training process of MAE by proposing a Semantic-Guided Masking strategy. Compared to widely adopted random masking, our masking strategy can gradually guide the network to learn various information, i.e.,…
Department of Computer Science and Engineering（计算机科学与工程系）；Lehigh University（里海大学）
Many deep neural networks are susceptible to minute perturbations of images that have been carefully crafted to cause misclassification. Ideally, a robust classifier would be immune to small variations in input images, and a number of defensive approaches have been created as a result. One method would be to discern a latent representation which could ignore small changes to the input. However, typical autoencoders easily mingle inter-class latent representations when there are strong similarities between classes, making it harder for a decoder to accurately project the image back to the origi…
Masked Autoencoders (MAEs) learn generalizable representations for image, text, audio, video, etc., by reconstructing masked input data from tokens of the visible data. Current MAE approaches for videos rely on random patch, tube, or frame-based masking strategies to select these tokens. This paper proposes AdaMAE, an adaptive masking strategy for MAEs that is end-to-end trainable. Our adaptive masking strategy samples visible tokens based on the semantic context using an auxiliary sampling network. This network estimates a categorical distribution over spacetime-patch tokens. The tokens that …
Resource-Efficient Multiview Perception: Integrating Semantic Masking with Masked Autoencoders
Faculty of Engineering, The University of Sydney, Australia（悉尼大学工程学院）；School of Computer, Data & Mathematical Sciences, Western Sydney University, Australia（西悉尼大学计算机、数据与数学科学学院）
Nanjing University（南京大学）；Tencent AI Lab（腾讯人工智能实验室）；Shanghai AI Lab（上海人工智能实验室）
Pre-training video transformers on extra large-scale datasets is generally required to achieve premier performance on relatively small datasets. In this paper, we show that video masked autoencoders (VideoMAE) are data-efficient learners for self-supervised video pre-training (SSVP). We are inspired by the recent ImageMAE and propose customized video tube masking with an extremely high ratio. This simple design makes video reconstruction a more challenging self-supervision task, thus encouraging extracting more effective video representations during this pre-training process. We obtain three i…
Novelty detection is the task of recognizing samples that do not belong to the distribution of the target class. During training, the novelty class is absent, preventing the use of traditional classification approaches. Deep autoencoders have been widely used as a base of many unsupervised novelty detection methods. In particular, context autoencoders have been successful in the novelty detection task because of the more effective representations they learn by reconstructing original images from randomly masked images. However, a significant drawback of context autoencoders is that random mask…
College of Intelligence and Computing, Tianjin University（天津大学智能与计算学部）
Visual context provides grounding information for multimodal machine translation (MMT). However, previous MMT models and probing studies on visual features suggest that visual information is less explored in MMT as it is often redundant to textual information. In this paper, we propose an object-level visual context modeling framework (OVC) to efficiently capture and explore visual information for multimodal machine translation. With detected objects, the proposed OVC encourages MMT to ground translation on desirable visual objects by masking irrelevant objects in the visual modality. We equip…
Semantics-Guided Multimodal Masked Autoencoder Pretraining for 3D BEV Object Detection
University of Lincoln, Lincoln Centre for Autonomous Systems（林肯大学，林肯自主系统中心）；University of Cambridge, Institute for Manufacturing, Department of Engineering（剑桥大学，制造研究所，工程系）
This paper studies a conceptually simple extension of Masked Autoencoders (MAE) to spatiotemporal representation learning from videos. We randomly mask out spacetime patches in videos and learn an autoencoder to reconstruct them in pixels. Interestingly, we show that our MAE method can learn strong representations with almost no inductive bias on spacetime (only except for patch and positional embeddings), and spacetime-agnostic random masking performs the best. We observe that the optimal masking ratio is as high as 90% (vs. 75% on images), supporting the hypothesis that this ratio is related…
Great Bay University（大湾区大学）；Nanyang Technological University（南洋理工大学）；University of Oulu（奥卢大学）；Lappeenranta-Lahti University of Technology（拉彭兰塔-拉赫蒂理工大学）；South China University of Technology（华南理工大学）
Recently, vision transformer (ViT) based multimodal learning methods have been proposed to improve the robustness of face anti-spoofing (FAS) systems. However, there are still no works to explore the fundamental natures (\textit{e.g.}, modality-aware inputs, suitable multimodal pre-training, and efficient finetuning) in vanilla ViT for multimodal FAS. In this paper, we investigate three key factors (i.e., inputs, pre-training, and finetuning) in ViT for multimodal FAS with RGB, Infrared (IR), and Depth. First, in terms of the ViT inputs, we find that leveraging local feature descriptors benefi…
Beyond Image-Text Matching: Verb Understanding in Multimodal Transformers Using Guided Masking
Faculty of Information Technology, Brno University of Technology（布尔诺理工大学信息技术学院）；Kempelen Institute of Intelligent Technologies（肯佩伦智能技术研究所）；George Mason University（乔治梅森大学）
Point-DAE: Denoising Autoencoders for Self-supervised Point Cloud Learning
Department of Computing, The Hong Kong Polytechnic University（计算系，香港理工大学）；School of Electronic and Information Engineering, South China University of Technology（电子与信息工程学院，华南理工大学）
VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking
State Key Laboratory for Novel Software Technology, Nanjing University（南京大学计算机软件新技术国家重点实验室）；Shanghai AI Lab（上海人工智能实验室）；Shenzhen Institute of Advanced Technology, CAS（中国科学院深圳先进技术研究院）
本文提出双重掩码策略和渐进式训练范式，高效扩展 VideoMAE 至十亿参数视频 ViT 模型，在多个视频基准上取得领先性能。
Building scalable models to learn from diverse, multimodal data remains an open challenge. For vision-language data, the dominant approaches are based on contrastive learning objectives that train a separate encoder for each modality. While effective, contrastive learning approaches introduce sampling bias depending on the data augmentations used, which can degrade performance on downstream tasks. Moreover, these methods are limited to paired image-text data, and cannot leverage widely-available unpaired data. In this paper, we investigate whether a large multimodal model trained purely via ma…
Aalborg University（奥尔堡大学）；University Politehnica of Bucharest（布加勒斯特理工大学）；University of Bucharest（布加勒斯特大学）；Milestone Systems（迈尔斯通系统公司）；SecurifAI（SecurifAI公司）
本文提出 CL-MAE，利用可学习掩码模块在 MAE 自监督训练中生成由易到难的重建课程，并在 ImageNet 及五个下游任务上验证其优于 MAE 的表示学习效果。
UiT The Arctic University of Norway（北极挪威特罗姆瑟大学）；NORCE Norwegian Research Center（NORCE挪威研究中心）；University of Genoa（热那亚大学）
Image translation with convolutional autoencoders has recently been used as an approach to multimodal change detection in bitemporal satellite images. A main challenge is the alignment of the code spaces by reducing the contribution of change pixels to the learning of the translation function. Many existing approaches train the networks by exploiting supervised information of the change areas, which, however, is not always available. We propose to extract relational pixel information captured by domain-specific affinity matrices at the input and use this to enforce alignment of the code spaces…
Social-MAE: A Transformer-Based Multimodal Autoencoder for Face and Voice
Numediart Institute, ISIA Lab, University of Mons（Numediart研究院、ISIA实验室、蒙斯大学）；Institute for Creative Technologies, University of Southern California（创意技术研究所、南加州大学）
Structure-Guided Masked Autoencoders for Ultra-High Resolution Scientific Image Understanding
Hokkaido University（北海道大学）；RIKEN Center for Computational Science (R-CCS)（理化学研究所计算科学研究中心（R-CCS））；Oak Ridge National Laboratory（橡树岭国家实验室）；A*STAR（新加坡科技研究局）；Institute of Science Tokyo（东京科学大学）；Japan Synchrotron Radiation Research Institute (JASRI)（日本同步辐射研究所（JASRI））；RIKEN SPring-8 Center（理化学研究所SPring-8中心）
A$^{2}$-MAE: A spatial-temporal-spectral unified remote sensing pre-training method based on anchor-aware masked autoencoder
National Supercomputing Center in Shenzhen（深圳国家超级计算中心）；Tsinghua University（清华大学）；The University of Hong Kong（香港大学）；Sun Yat-Sen University（中山大学）；SenseTime Group（商汤集团）
Vision Transformer (ViT) suffers from data scarcity in semi-supervised learning (SSL). To alleviate this issue, inspired by masked autoencoder (MAE), which is a data-efficient self-supervised learner, we propose Semi-MAE, a pure ViT-based SSL framework consisting of a parallel MAE branch to assist the visual representation learning and make the pseudo labels more accurate. The MAE branch is designed as an asymmetric architecture consisting of a lightweight decoder and a shared-weights encoder. We feed the weakly-augmented unlabeled data with a high masking ratio to the MAE branch and reconstru…
In this paper, we propose a multimodal transformer using parallel concatenated architecture. Instead of using patches, we use column stripes for images in R, G, B channels as the transformer input. The column stripes keep the spatial relations of original image. We incorporate the multimodal transformer with variational autoencoder for synthetic cross-modal data generation. The multimodal transformer is designed using multiple compression matrices, and it serves as encoders for Parallel Concatenated Variational AutoEncoders (PC-VAE). The PC-VAE consists of multiple encoders, one latent space, …
Cooperative Medianet Innovation Center（未来媒体网络协同创新中心）；Shanghai Jiao Tong University（上海交通大学）；Shanghai AI Laboratory（上海人工智能实验室）
Recently, anomaly detection and localization in multimedia data have received significant attention among the machine learning community. In real-world applications such as medical diagnosis and industrial defect detection, anomalies only present in a fraction of the images. To extend the reconstruction-based anomaly detection architecture to the localized anomalies, we propose a self-supervised learning approach through random masking and then restoring, named Self-Supervised Masking (SSM) for unsupervised anomaly detection and localization. SSM not only enhances the training of the inpaintin…
CREATIS（CREATIS实验室）；Université Lyon 1（里昂第一大学）；INSA-Lyon（里昂国立应用科学学院）；Hospices Civils de Lyon（里昂民用医院）；CarMeN（CarMeN实验室）；LARIS（LARIS实验室）；Universite d’Angers（昂热大学）
Patient outcome prediction is critical in management of ischemic stroke. In this paper, a novel machine learning model is proposed for stroke outcome prediction using multimodal Magnetic Resonance Imaging (MRI). The proposed model consists of two serial levels of Autoencoders (AEs), where different AEs at level 1 are used for learning unimodal features from different MRI modalities and a AE at level 2 is used to combine the unimodal features into compressed multimodal features. The sequences of multimodal features of a given patient are then used by an LSTM network for predicting outcome score…
Joint Low-level and High-level Textual Representation Learning with Multiple Masking Strategies
Wenzhou University Artificial Intelligence and Advanced Manufacturing Institute (AIAMI)（温州大学人工智能与先进制造研究院）；Graduate School of Engineering, Tohoku University（东京东京大学工学研究科）
