我们提出了一个新的基准数据集，即Sapsucker Woods 60（SSW60），用于推进视听细颗粒分类的研究。尽管我们的社区在图像上的细粒度视觉分类方面取得了长足的进步，但音频和视频细颗粒分类的对应物相对尚未探索。为了鼓励在这个领域的进步，我们已经仔细构建了SSW60数据集，以使研究人员能够以三种不同的方式对相同的类别进行分类：图像，音频和视频。该数据集涵盖了60种鸟类，由现有数据集以及全新的专家策划音频和视频数据集组成。我们通过使用最先进的变压器方法进行了彻底基准的视听分类性能和模态融合实验。我们的发现表明，视听融合方法的性能要比仅使用基于图像或音频的方法来进行视频分类任务要好。我们还提出了有趣的模态转移实验，这是由SSW60的独特构造所涵盖的三种不同模态所实现的。我们希望SSW60数据集和伴随的基线在这个迷人的地区进行研究。

Humans perceive the world by concurrently processing and fusing high-dimensional inputs from multiple modalities such as vision and audio. Machine perception models, in stark contrast, are typically modality-specific and optimised for unimodal benchmarks, and hence late-stage fusion of final representations or predictions from each modality (`late-fusion') is still a dominant paradigm for multimodal video classification. Instead, we introduce a novel transformer based architecture that uses `fusion bottlenecks' for modality fusion at multiple layers. Compared to traditional pairwise self-attention, our model forces information between different modalities to pass through a small number of bottleneck latents, requiring the model to collate and condense the most relevant information in each modality and only share what is necessary. We find that such a strategy improves fusion performance, at the same time reducing computational cost. We conduct thorough ablation studies, and achieve state-of-the-art results on multiple audio-visual classification benchmarks including Audioset, Epic-Kitchens and VGGSound. All code and models will be released.

为了使AI安全地在医院，学校和工作场所等现实世界中安全部署，它必须能够坚定地理解物理世界。这种推理的基础是物理常识：了解可用对象的物理特性和提供的能力，如何被操纵以及它们如何与其他对象进行交互。物理常识性推理从根本上是一项多感官任务，因为物理特性是通过多种模式表现出来的，其中两个是视觉和声学。我们的论文通过贡献PACS来朝着现实世界中的物理常识推理：第一个用于物理常识属性注释的视听基准。 PACS包含13,400对答案对，涉及1,377个独特的物理常识性问题和1,526个视频。我们的数据集提供了新的机会来通过将音频作为此多模式问题的核心组成部分来推进物理推理的研究领域。使用PACS，我们在我们的新挑战性任务上评估了多种最先进的模型。尽管某些模型显示出令人鼓舞的结果（精度为70％），但它们都没有人类的绩效（精度为95％）。我们通过证明多模式推理的重要性并为未来的研究提供了可能的途径来结束本文。

在本文中，我们考虑了视听同步的问题应用于视频`in-wild'（即，超越语音的一般类）。作为一项新任务，我们识别并策划具有高视听相关性的测试集，即VGG-SOCK SYNC。我们比较了一些专门设计的基于变压器的架构变体，用于模拟任意长度的音频和视觉信号，同时显着降低训练期间的内存要求。我们进一步对策划数据集进行了深入的分析，并定义了开放域视听同步的评估度量。我们在标准唇读语音基准测试中应用我们的方法，LRS2和LRS3，在各个方面的消融。最后，我们在新的VGG-SOCKC SYNC视频数据集中设置了与超过160个不同类别的通用视听同步的第一个基准。在所有情况下，我们所提出的模型通过显着的保证金优于以前的最先进。

我们呈现了一个用于学习视听表示的自我监督的框架。在我们的框架中引入了一种小说概念，其中除了学习模态和标准的“同步的”跨模型关系之外，riscross也会学习“异步”的跨模式关系。我们展示通过放松音频和视觉模态之间的时间同步性，网络了解强劲的时间不变的表示。我们的实验表明，音频和视觉方式的强大增强，可放松交叉模态时间同步优化性能。要预先绘制我们提出的框架，我们使用具有不同大小，动力学，动力学-400和augioset的不同数据集。学习的表示是在许多下游任务中评估的，即行动识别，声音分类和检索。 Crisscross显示了动作识别的最先进的性能（UCF101和HMDB51）和声音分类（ESC50）。将公开可用的代码和预赠品模型。

我们建议探索一个称为视听分割（AVS）的新问题，其中的目标是输出在图像帧时产生声音的对象的像素级映射。为了促进这项研究，我们构建了第一个视频分割基准（AVSBENCH），为声音视频中的声音对象提供像素的注释。使用此基准测试了两个设置：1）具有单个声源的半监督音频分割和2）完全监督的音频段段，并带有多个声源。为了解决AVS问题，我们提出了一种新颖的方法，该方法使用时间像素的视听相互作用模块注入音频语义作为视觉分割过程的指导。我们还设计正规化损失，以鼓励训练期间的视听映射。 AVSBench上的定量和定性实验将我们的方法与相关任务中的几种现有方法进行了比较，这表明所提出的方法有望在音频和像素视觉语义之间建立桥梁。代码可从https://github.com/opennlplab/avsbench获得。

We consider the question: what can be learnt by looking at and listening to a large number of unlabelled videos? There is a valuable, but so far untapped, source of information contained in the video itself -the correspondence between the visual and the audio streams, and we introduce a novel "Audio-Visual Correspondence" learning task that makes use of this. Training visual and audio networks from scratch, without any additional supervision other than the raw unconstrained videos themselves, is shown to successfully solve this task, and, more interestingly, result in good visual and audio representations. These features set the new state-of-the-art on two sound classification benchmarks, and perform on par with the state-of-the-art selfsupervised approaches on ImageNet classification. We also demonstrate that the network is able to localize objects in both modalities, as well as perform fine-grained recognition tasks.

Videos are a rich source of multi-modal supervision. In this work, we learn representations using self-supervision by leveraging three modalities naturally present in videos: visual, audio and language streams. To this end, we introduce the notion of a multimodal versatile network -a network that can ingest multiple modalities and whose representations enable downstream tasks in multiple modalities. In particular, we explore how best to combine the modalities, such that fine-grained representations of the visual and audio modalities can be maintained, whilst also integrating text into a common embedding. Driven by versatility, we also introduce a novel process of deflation, so that the networks can be effortlessly applied to the visual data in the form of video or a static image. We demonstrate how such networks trained on large collections of unlabelled video data can be applied on video, video-text, image and audio tasks. Equipped with these representations, we obtain state-of-the-art performance on multiple challenging benchmarks including UCF101, HMDB51, Kinetics600, Audioset and ESC-50 when compared to previous self-supervised work. Our models are publicly available [1, 2, 3]. * Equal contribution. † Work done during an internship at DeepMind. 34th Conference on Neural Information Processing Systems (NeurIPS 2020),

There is a natural correlation between the visual and auditive elements of a video. In this work we leverage this connection to learn general and effective models for both audio and video analysis from self-supervised temporal synchronization. We demonstrate that a calibrated curriculum learning scheme, a careful choice of negative examples, and the use of a contrastive loss are critical ingredients to obtain powerful multi-sensory representations from models optimized to discern temporal synchronization of audio-video pairs. Without further finetuning, the resulting audio features achieve performance superior or comparable to the state-of-the-art on established audio classification benchmarks (DCASE2014 and ESC-50). At the same time, our visual subnet provides a very effective initialization to improve the accuracy of video-based action recognition models: compared to learning from scratch, our self-supervised pretraining yields a remarkable gain of +19.9% in action recognition accuracy on UCF101 and a boost of +17.7% on HMDB51.

作为人类，我们通过我们所有的感官来驾驭世界，使用每个人从每个人纠正其他人。我们介绍了Merlot Reserve，一个模型，该模型是联合随着时间的推移而表示视频的模型 - 通过从音频，字幕和视频帧学习的新培训目标。给出了一个视频，我们用掩模令牌替换文本和音频的片段;该模型通过选择正确的蒙版片段来学习。我们的目标比替代方面更快地学习，并在规模上表现良好：我们预先逼近2000万YouTube视频。经验结果表明，Merlot Reserve学会通过所有组成模式的视频的强烈陈述。在FineTuned时，它在VCR和TVQA上为VCR和TVQA进行了新的最先进，优先于前勤工作分别为5％和7％。消融表明，两个任务都受益于音频预制 - 甚至录像机，围绕图像中心的QA任务（没有声音）。此外，我们的客观使开箱即用的预测，揭示了强大的多式联合致辞理解。在一个完全零拍摄的环境中，我们的模型在四个视频理解任务中获得竞争结果，甚至优于最近提出的定位推理（星）基准的监督方法。我们分析为什么包含音频导致更好的视觉语言表示，这表明未来研究的重要机会。我们通过讨论多式联运预测的道德和社会影响来得出结论。

来自视频数据的多模态学习最近看过，因为它允许在没有人为注释的情况下培训语义有意义的嵌入，从而使得零射击检索和分类等任务。在这项工作中，我们提出了一种多模态，模态无政府主义融合变压器方法，它学会在多个模态之间交换信息，例如视频，音频和文本，并将它们集成到加入的多模态表示中，以获取聚合的嵌入多模态时间信息。我们建议培训系统的组合丢失，单个模态以及成对的方式，明确地留出任何附加组件，如位置或模态编码。在测试时间时，产生的模型可以处理和融合任意数量的输入模态。此外，变压器的隐式属性允许处理不同长度的输入。为了评估所提出的方法，我们在大规模HOWASET上培训模型，并评估四个具有挑战性的基准数据集上产生的嵌入空间获得最先进的视频检索和零射击视频动作定位。

The remarkable success of deep learning in various domains relies on the availability of large-scale annotated datasets. However, obtaining annotations is expensive and requires great effort, which is especially challenging for videos. Moreover, the use of human-generated annotations leads to models with biased learning and poor domain generalization and robustness. As an alternative, self-supervised learning provides a way for representation learning which does not require annotations and has shown promise in both image and video domains. Different from the image domain, learning video representations are more challenging due to the temporal dimension, bringing in motion and other environmental dynamics. This also provides opportunities for video-exclusive ideas that advance self-supervised learning in the video and multimodal domain. In this survey, we provide a review of existing approaches on self-supervised learning focusing on the video domain. We summarize these methods into four different categories based on their learning objectives: 1) pretext tasks, 2) generative learning, 3) contrastive learning, and 4) cross-modal agreement. We further introduce the commonly used datasets, downstream evaluation tasks, insights into the limitations of existing works, and the potential future directions in this area.

Video synthesis methods rapidly improved in recent years, allowing easy creation of synthetic humans. This poses a problem, especially in the era of social media, as synthetic videos of speaking humans can be used to spread misinformation in a convincing manner. Thus, there is a pressing need for accurate and robust deepfake detection methods, that can detect forgery techniques not seen during training. In this work, we explore whether this can be done by leveraging a multi-modal, out-of-domain backbone trained in a self-supervised manner, adapted to the video deepfake domain. We propose FakeOut; a novel approach that relies on multi-modal data throughout both the pre-training phase and the adaption phase. We demonstrate the efficacy and robustness of FakeOut in detecting various types of deepfakes, especially manipulations which were not seen during training. Our method achieves state-of-the-art results in cross-manipulation and cross-dataset generalization. This study shows that, perhaps surprisingly, training on out-of-domain videos (i.e., videos with no speaking humans), can lead to better deepfake detection systems. Code is available on GitHub.

Visual and audio modalities are highly correlated, yet they contain different information. Their strong correlation makes it possible to predict the semantics of one from the other with good accuracy. Their intrinsic differences make cross-modal prediction a potentially more rewarding pretext task for self-supervised learning of video and audio representations compared to within-modality learning. Based on this intuition, we propose Cross-Modal Deep Clustering (XDC), a novel selfsupervised method that leverages unsupervised clustering in one modality (e.g., audio) as a supervisory signal for the other modality (e.g., video). This cross-modal supervision helps XDC utilize the semantic correlation and the differences between the two modalities. Our experiments show that XDC outperforms single-modality clustering and other multi-modal variants. XDC achieves state-of-the-art accuracy among self-supervised methods on multiple video and audio benchmarks. Most importantly, our video model pretrained on large-scale unlabeled data significantly outperforms the same model pretrained with full-supervision on ImageNet and Kinetics for action recognition on HMDB51 and UCF101. To the best of our knowledge, XDC is the first self-supervised learning method that outperforms large-scale fully-supervised pretraining for action recognition on the same architecture.

利用在大规模图像文本对中预先训练的视觉和语言模型（VLM）成为开放式视觉识别的有希望的范式。在这项工作中，我们通过利用视频中自然存在的运动和音频来扩展这种范式。我们提出\ textbf {mov}，这是\ textbf {m} ult-imodal \ textbf {o} pen- \ textbf {v} ocabulary视频分类的简单而有效的方法。在MOV中，我们直接使用具有最小修改的预训练VLM的视觉编码器来编码视频，光流和音频频谱图。我们设计一种跨模式融合机制来汇总免费的多模式信息。 Kinetics-700和VGGSOUND的实验表明，引入流量或音频模态会带来预先训练的VLM和现有方法的大量性能增长。具体而言，MOV极大地提高了基础类别的准确性，而在新颖的课程上则更好地概括了。 MOV在UCF和HMDB零摄像视频分类基准上实现了最新结果，从而极大地超过了基于VLMS的传统零摄像方法和最新方法。代码和模型将发布。

室内场景识别是一种不断增长的领域，具有巨大的行为理解，机器人本地化和老年人监测等。在这项研究中，我们使用从社交媒体收集的多模态学习和视频数据来从新的角度来看场景识别的任务。社交媒体视频的可访问性和各种可以为现代场景识别技术和应用提供现实数据。我们提出了一种基于转录语音的融合到文本和视觉功能的模型，用于在名为Instaindoor的室内场景的社交媒体视频的新型数据集上进行分类。我们的模型可实现高达70％的精度和0.7 F1分数。此外，我们通过在室内场景的YouTube-8M子集上基准测试，我们突出了我们的方法的潜力，在那里它达到了74％的精度和0.74f1分数。我们希望这项工作的贡献铺平了在挑战领域的室内场景认可领域的新型研究。

我们介绍了一种视听方法，用于远程文本到视频检索。与以前专为简短视频检索设计的方法（例如，持续时间为5-15秒）不同，我们的方法旨在检索捕获复杂人类动作的长时间视频。仅标准视频方法的一个挑战是与从这样的长视频中处理数百个密集提取的帧相关的大量计算成本。为了解决这个问题，我们建议用紧凑的音频提示替换视频的部分，这些线索简洁地汇总了动态音频事件，并且处理便宜。我们的方法称为Eclipse（带有声音编码的有效剪辑），通过添加一个统一的视听变压器块，将流行的剪辑模型调整为视听视频设置，该块从视频和音频流中捕获互补的提示。除了比仅长期视频的方法快2.92倍和2.34倍的内存效率外，我们的方法还可以在几个不同的远程视频数据集上，例如ActivityNet，QVHighighlights，Youcoook2，Youcoook2，Youcook2，Youcook2，Youcook2，Youcook2，Youcook2，Youcook2， Didemo和Charades。

In this paper our objectives are, first, networks that can embed audio and visual inputs into a common space that is suitable for cross-modal retrieval; and second, a network that can localize the object that sounds in an image, given the audio signal. We achieve both these objectives by training from unlabelled video using only audio-visual correspondence (AVC) as the objective function. This is a form of crossmodal self-supervision from video. To this end, we design new network architectures that can be trained for cross-modal retrieval and localizing the sound source in an image, by using the AVC task. We make the following contributions: (i) show that audio and visual embeddings can be learnt that enable both within-mode (e.g. audio-to-audio) and between-mode retrieval; (ii) explore various architectures for the AVC task, including those for the visual stream that ingest a single image, or multiple images, or a single image and multi-frame optical flow; (iii) show that the semantic object that sounds within an image can be localized (using only the sound, no motion or flow information); and (iv) give a cautionary tale on how to avoid undesirable shortcuts in the data preparation.

We present Masked Audio-Video Learners (MAViL) to train audio-visual representations. Our approach learns with three complementary forms of self-supervision: (1) reconstruction of masked audio and video input data, (2) intra- and inter-modal contrastive learning with masking, and (3) self-training by reconstructing joint audio-video contextualized features learned from the first two objectives. Pre-training with MAViL not only enables the model to perform well in audio-visual classification and retrieval tasks but also improves representations of each modality in isolation, without using information from the other modality for fine-tuning or inference. Empirically, MAViL sets a new state-of-the-art on AudioSet (53.1 mAP) and VGGSound (67.1% accuracy). For the first time, a self-supervised audio-visual model outperforms ones that use external supervision on these benchmarks. Code will be available soon.

State-of-the-art computer vision systems are trained to predict a fixed set of predetermined object categories. This restricted form of supervision limits their generality and usability since additional labeled data is needed to specify any other visual concept. Learning directly from raw text about images is a promising alternative which leverages a much broader source of supervision. We demonstrate that the simple pre-training task of predicting which caption goes with which image is an efficient and scalable way to learn SOTA image representations from scratch on a dataset of 400 million (image, text) pairs collected from the internet. After pre-training, natural language is used to reference learned visual concepts (or describe new ones) enabling zero-shot transfer of the model to downstream tasks. We study the performance of this approach by benchmarking on over 30 different existing computer vision datasets, spanning tasks such as OCR, action recognition in videos, geo-localization, and many types of fine-grained object classification. The model transfers non-trivially to most tasks and is often competitive with a fully supervised baseline without the need for any dataset specific training. For instance, we match the accuracy of the original ResNet-50 on ImageNet zero-shot without needing to use any of the 1.28 million training examples it was trained on. We release our code and pre-trained model weights at https://github.com/OpenAI/CLIP.