PyTorchKR Papers

최신 논문

총 358편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2206.12126공식 코드 · Papers with Code

Temporal Attention Unit: Towards Efficient Spatiotemporal Predictive Learning

Spatiotemporal predictive learning aims to generate future frames by learning from historical frames. In this paper, we investigate existing methods and present a general framework of spatiotemporal predictive learning, in which the spatial encoder and decoder capture intra-frame features and the middle temporal module catches inter-frame correlations. While the mainstream methods employ recurrent units to capture long-term temporal dependencies, they suffer from low computational efficiency due to their unparallelizable architectures. To parallelize the temporal module, we propose the Temporal Attention Unit (TAU), which decomposes the temporal attention into intra-frame statical attention and inter-frame dynamical attention. Moreover, while the mean squared error loss focuses on intra-frame errors, we introduce a novel differential divergence regularization to take inter-frame variations into account. Extensive experiments demonstrate that the proposed method enables the derived model to achieve competitive performance on various spatiotemporal prediction benchmarks.

arXiv:2206.11894

MaskViT: Masked Visual Pre-Training for Video Prediction

The ability to predict future visual observations conditioned on past observations and motor commands can enable embodied agents to plan solutions to a variety of tasks in complex environments. This work shows that we can create good video prediction models by pre-training transformers via masked visual modeling. Our approach, named MaskViT, is based on two simple design decisions. First, for memory and training efficiency, we use two types of window attention: spatial and spatiotemporal. Second, during training, we mask a variable percentage of tokens instead of a fixed mask ratio. For inference, MaskViT generates all tokens via iterative refinement where we incrementally decrease the masking ratio following a mask scheduling function. On several datasets we demonstrate that MaskViT outperforms prior works in video prediction, is parameter efficient, and can generate high-resolution videos (256x256). Further, we demonstrate the benefits of inference speedup (up to 512x) due to iterative decoding by using MaskViT for planning on a real robot. Our work suggests that we can endow embodied agents with powerful predictive models by leveraging the general framework of masked visual modeling with minimal domain knowledge.

arXiv:2206.07696공식 코드 · Papers with Code

Diffusion Models for Video Prediction and Infilling

Predicting and anticipating future outcomes or reasoning about missing information in a sequence are critical skills for agents to be able to make intelligent decisions. This requires strong, temporally coherent generative capabilities. Diffusion models have shown remarkable success in several generative tasks, but have not been extensively explored in the video domain. We present Random-Mask Video Diffusion (RaMViD), which extends image diffusion models to videos using 3D convolutions, and introduces a new conditioning technique during training. By varying the mask we condition on, the model is able to perform video prediction, infilling, and upsampling. Due to our simple conditioning scheme, we can utilize the same architecture as used for unconditional training, which allows us to train the model in a conditional and unconditional fashion at the same time. We evaluate RaMViD on two benchmark datasets for video prediction, on which we achieve state-of-the-art results, and one for video generation. High-resolution videos are provided at https://sites.google.com/view/video-diffusion-prediction.

arXiv:2206.05099공식 코드 · Papers with Code

SimVP: Simpler yet Better Video Prediction

From CNN, RNN, to ViT, we have witnessed remarkable advancements in video prediction, incorporating auxiliary inputs, elaborate neural architectures, and sophisticated training strategies. We admire these progresses but are confused about the necessity: is there a simple method that can perform comparably well? This paper proposes SimVP, a simple video prediction model that is completely built upon CNN and trained by MSE loss in an end-to-end fashion. Without introducing any additional tricks and complicated strategies, we can achieve state-of-the-art performance on five benchmark datasets. Through extended experiments, we demonstrate that SimVP has strong generalization and extensibility on real-world datasets. The significant reduction of training cost makes it easier to scale to complex scenarios. We believe SimVP can serve as a solid baseline to stimulate the further development of video prediction. The code is available at \href{https://github.com/gaozhangyang/SimVP-Simpler-yet-Better-Video-Prediction}{Github}.

arXiv:2206.04381공식 코드 · Papers with Code

STIP: A SpatioTemporal Information-Preserving and Perception-Augmented Model for High-Resolution Video Prediction

Although significant achievements have been achieved by recurrent neural network (RNN) based video prediction methods, their performance in datasets with high resolutions is still far from satisfactory because of the information loss problem and the perception-insensitive mean square error (MSE) based loss functions. In this paper, we propose a Spatiotemporal Information-Preserving and Perception-Augmented Model (STIP) to solve the above two problems. To solve the information loss problem, the proposed model aims to preserve the spatiotemporal information for videos during the feature extraction and the state transitions, respectively. Firstly, a Multi-Grained Spatiotemporal Auto-Encoder (MGST-AE) is designed based on the X-Net structure. The proposed MGST-AE can help the decoders recall multi-grained information from the encoders in both the temporal and spatial domains. In this way, more spatiotemporal information can be preserved during the feature extraction for high-resolution videos. Secondly, a Spatiotemporal Gated Recurrent Unit (STGRU) is designed based on the standard Gated Recurrent Unit (GRU) structure, which can efficiently preserve spatiotemporal information during the state transitions. The proposed STGRU can achieve more satisfactory performance with a much lower computation load compared with the popular Long Short-Term (LSTM) based predictive memories. Furthermore, to improve the traditional MSE loss functions, a Learned Perceptual Loss (LP-loss) is further designed based on the Generative Adversarial Networks (GANs), which can help obtain a satisfactory trade-off between the objective quality and the perceptual quality. Experimental results show that the proposed STIP can predict videos with more satisfactory visual quality compared with a variety of state-of-the-art methods. Source code has been available at \url{https://github.com/ZhengChang467/STIPHR}.

arXiv:2206.04003공식 코드 · Papers with Code

Patch-based Object-centric Transformers for Efficient Video Generation

In this work, we present Patch-based Object-centric Video Transformer (POVT), a novel region-based video generation architecture that leverages object-centric information to efficiently model temporal dynamics in videos. We build upon prior work in video prediction via an autoregressive transformer over the discrete latent space of compressed videos, with an added modification to model object-centric information via bounding boxes. Due to better compressibility of object-centric representations, we can improve training efficiency by allowing the model to only access object information for longer horizon temporal information. When evaluated on various difficult object-centric datasets, our method achieves better or equal performance to other video generation models, while remaining computationally more efficient and scalable. In addition, we show that our method is able to perform object-centric controllability through bounding box manipulation, which may aid downstream tasks such as video editing, or visual planning. Samples are available at https://sites.google.com/view/povt-public

arXiv:2206.03010공식 코드 · Papers with Code

MS-RNN: A Flexible Multi-Scale Framework for Spatiotemporal Predictive Learning

Spatiotemporal predictive learning, which predicts future frames through historical prior knowledge with the aid of deep learning, is widely used in many fields. Previous work essentially improves the model performance by widening or deepening the network, but it also brings surging memory overhead, which seriously hinders the development and application of this technology. In order to improve the performance without increasing memory consumption, we focus on scale, which is another dimension to improve model performance but with low memory requirement. The effectiveness has been widely demonstrated in many CNN-based tasks such as image classification and semantic segmentation, but it has not been fully explored in recent RNN models. In this paper, learning from the benefit of multi-scale, we propose a general framework named Multi-Scale RNN (MS-RNN) to boost recent RNN models for spatiotemporal predictive learning. We verify the MS-RNN framework by thorough theoretical analyses and exhaustive experiments, where the theory focuses on memory reduction and performance improvement while the experiments employ eight RNN models (ConvLSTM, TrajGRU, PredRNN, PredRNN++, MIM, MotionRNN, PredRNN-V2, and PrecipLSTM) and four datasets (Moving MNIST, TaxiBJ, KTH, and Germany). The results show the efficiency that RNN models incorporating our framework have much lower memory cost but better performance than before. Our code is released at \url{https://github.com/mazhf/MS-RNN}.

arXiv:2205.15821공식 코드 · Papers with Code

Unsupervised Image Representation Learning with Deep Latent Particles

We propose a new representation of visual data that disentangles object position from appearance. Our method, termed Deep Latent Particles (DLP), decomposes the visual input into low-dimensional latent ``particles'', where each particle is described by its spatial location and features of its surrounding region. To drive learning of such representations, we follow a VAE-based approach and introduce a prior for particle positions based on a spatial-softmax architecture, and a modification of the evidence lower bound loss inspired by the Chamfer distance between particles. We demonstrate that our DLP representations are useful for downstream tasks such as unsupervised keypoint (KP) detection, image manipulation, and video prediction for scenes composed of multiple dynamic objects. In addition, we show that our probabilistic interpretation of the problem naturally provides uncertainty estimates for particle locations, which can be used for model selection, among other tasks. Videos and code are available: https://taldatech.github.io/deep-latent-particles-web/

arXiv:2205.09853공식 코드 · Papers with Code

MCVD: Masked Conditional Video Diffusion for Prediction, Generation, and Interpolation

Video prediction is a challenging task. The quality of video frames from current state-of-the-art (SOTA) generative models tends to be poor and generalization beyond the training data is difficult. Furthermore, existing prediction frameworks are typically not capable of simultaneously handling other video-related tasks such as unconditional generation or interpolation. In this work, we devise a general-purpose framework called Masked Conditional Video Diffusion (MCVD) for all of these video synthesis tasks using a probabilistic conditional score-based denoising diffusion model, conditioned on past and/or future frames. We train the model in a manner where we randomly and independently mask all the past frames or all the future frames. This novel but straightforward setup allows us to train a single model that is capable of executing a broad range of video tasks, specifically: future/past prediction -- when only future/past frames are masked; unconditional generation -- when both past and future frames are masked; and interpolation -- when neither past nor future frames are masked. Our experiments show that this approach can generate high-quality frames for diverse types of videos. Our MCVD models are built from simple non-recurrent 2D-convolutional architectures, conditioning on blocks of frames and generating blocks of frames. We generate videos of arbitrary lengths autoregressively in a block-wise manner. Our approach yields SOTA results across standard video prediction and interpolation benchmarks, with computation times for training models measured in 1-12 days using $\le$ 4 GPUs. Project page: https://mask-cond-video-diffusion.github.io ; Code : https://github.com/voletiv/mcvd-pytorch

arXiv:2205.09430공식 코드 · Papers with Code

Action Conditioned Tactile Prediction: case study on slip prediction

Tactile predictive models can be useful across several robotic manipulation tasks, e.g. robotic pushing, robotic grasping, slip avoidance, and in-hand manipulation. However, available tactile prediction models are mostly studied for image-based tactile sensors and there is no comparison study indicating the best performing models. In this paper, we presented two novel data-driven action-conditioned models for predicting tactile signals during real-world physical robot interaction tasks (1) action condition tactile prediction and (2) action conditioned tactile-video prediction models. We use a magnetic-based tactile sensor that is challenging to analyse and test state-of-the-art predictive models and the only existing bespoke tactile prediction model. We compare the performance of these models with those of our proposed models. We perform the comparison study using our novel tactile-enabled dataset containing 51,000 tactile frames of a real-world robotic manipulation task with 11 flat-surfaced household objects. Our experimental results demonstrate the superiority of our proposed tactile prediction models in terms of qualitative, quantitative and slip prediction scores.

arXiv:2205.03571

Deep learning for spatio-temporal forecasting -- application to solar energy

This thesis tackles the subject of spatio-temporal forecasting with deep learning. The motivating application at Electricity de France (EDF) is short-term solar energy forecasting with fisheye images. We explore two main research directions for improving deep forecasting methods by injecting external physical knowledge. The first direction concerns the role of the training loss function. We show that differentiable shape and temporal criteria can be leveraged to improve the performances of existing models. We address both the deterministic context with the proposed DILATE loss function and the probabilistic context with the STRIPE model. Our second direction is to augment incomplete physical models with deep data-driven networks for accurate forecasting. For video prediction, we introduce the PhyDNet model that disentangles physical dynamics from residual information necessary for prediction, such as texture or details. We further propose a learning framework (APHYNITY) that ensures a principled and unique linear decomposition between physical and data-driven components under mild assumptions, leading to better forecasting performances and parameter identification.

arXiv:2205.01924

Zero-Episode Few-Shot Contrastive Predictive Coding: Solving intelligence tests without prior training

Video prediction models often combine three components: an encoder from pixel space to a small latent space, a latent space prediction model, and a generative model back to pixel space. However, the large and unpredictable pixel space makes training such models difficult, requiring many training examples. We argue that finding a predictive latent variable and using it to evaluate the consistency of a future image enables data-efficient predictions because it precludes the necessity of a generative model training. To demonstrate it, we created sequence completion intelligence tests in which the task is to identify a predictably changing feature in a sequence of images and use this prediction to select the subsequent image. We show that a one-dimensional Markov Contrastive Predictive Coding (M-CPC_1D) model solves these tests efficiently, with only five examples. Finally, we demonstrate the usefulness of M-CPC_1D in solving two tasks without prior training: anomaly detection and stochastic movement video prediction.

arXiv:2205.02084

Video Extrapolation in Space and Time

Novel view synthesis (NVS) and video prediction (VP) are typically considered disjoint tasks in computer vision. However, they can both be seen as ways to observe the spatial-temporal world: NVS aims to synthesize a scene from a new point of view, while VP aims to see a scene from a new point of time. These two tasks provide complementary signals to obtain a scene representation, as viewpoint changes from spatial observations inform depth, and temporal observations inform the motion of cameras and individual objects. Inspired by these observations, we propose to study the problem of Video Extrapolation in Space and Time (VEST). We propose a model that leverages the self-supervision and the complementary cues from both tasks, while existing methods can only solve one of them. Experiments show that our method achieves performance better than or comparable to several state-of-the-art NVS and VP methods on indoor and outdoor real-world datasets.

arXiv:2204.09456

STAU: A SpatioTemporal-Aware Unit for Video Prediction and Beyond

Video prediction aims to predict future frames by modeling the complex spatiotemporal dynamics in videos. However, most of the existing methods only model the temporal information and the spatial information for videos in an independent manner but haven't fully explored the correlations between both terms. In this paper, we propose a SpatioTemporal-Aware Unit (STAU) for video prediction and beyond by exploring the significant spatiotemporal correlations in videos. On the one hand, the motion-aware attention weights are learned from the spatial states to help aggregate the temporal states in the temporal domain. On the other hand, the appearance-aware attention weights are learned from the temporal states to help aggregate the spatial states in the spatial domain. In this way, the temporal information and the spatial information can be greatly aware of each other in both domains, during which, the spatiotemporal receptive field can also be greatly broadened for more reliable spatiotemporal modeling. Experiments are not only conducted on traditional video prediction tasks but also other tasks beyond video prediction, including the early action recognition and object detection tasks. Experimental results show that our STAU can outperform other methods on all tasks in terms of performance and computation efficiency.

arXiv:2204.05624공식 코드 · Papers with Code

Continual Predictive Learning from Videos

Predictive learning ideally builds the world model of physical processes in one or more given environments. Typical setups assume that we can collect data from all environments at all times. In practice, however, different prediction tasks may arrive sequentially so that the environments may change persistently throughout the training procedure. Can we develop predictive learning algorithms that can deal with more realistic, non-stationary physical environments? In this paper, we study a new continual learning problem in the context of video prediction, and observe that most existing methods suffer from severe catastrophic forgetting in this setup. To tackle this problem, we propose the continual predictive learning (CPL) approach, which learns a mixture world model via predictive experience replay and performs test-time adaptation with non-parametric task inference. We construct two new benchmarks based on RoboNet and KTH, in which different tasks correspond to different physical robotic environments or human actions. Our approach is shown to effectively mitigate forgetting and remarkably outperform the na\"ive combinations of previous art in video prediction and continual learning.

arXiv:2204.03458

Video Diffusion Models

Generating temporally coherent high fidelity video is an important milestone in generative modeling research. We make progress towards this milestone by proposing a diffusion model for video generation that shows very promising initial results. Our model is a natural extension of the standard image diffusion architecture, and it enables jointly training from image and video data, which we find to reduce the variance of minibatch gradients and speed up optimization. To generate long and higher resolution videos we introduce a new conditional sampling technique for spatial and temporal video extension that performs better than previously proposed methods. We present the first results on a large text-conditioned video generation task, as well as state-of-the-art results on established benchmarks for video prediction and unconditional video generation. Supplementary material is available at https://video-diffusion.github.io/

arXiv:2203.16084

STRPM: A Spatiotemporal Residual Predictive Model for High-Resolution Video Prediction

Although many video prediction methods have obtained good performance in low-resolution (64$\sim$128) videos, predictive models for high-resolution (512$\sim$4K) videos have not been fully explored yet, which are more meaningful due to the increasing demand for high-quality videos. Compared with low-resolution videos, high-resolution videos contain richer appearance (spatial) information and more complex motion (temporal) information. In this paper, we propose a Spatiotemporal Residual Predictive Model (STRPM) for high-resolution video prediction. On the one hand, we propose a Spatiotemporal Encoding-Decoding Scheme to preserve more spatiotemporal information for high-resolution videos. In this way, the appearance details for each frame can be greatly preserved. On the other hand, we design a Residual Predictive Memory (RPM) which focuses on modeling the spatiotemporal residual features (STRF) between previous and future frames instead of the whole frame, which can greatly help capture the complex motion information in high-resolution videos. In addition, the proposed RPM can supervise the spatial encoder and temporal encoder to extract different features in the spatial domain and the temporal domain, respectively. Moreover, the proposed model is trained using generative adversarial networks (GANs) with a learned perceptual loss (LP-loss) to improve the perceptual quality of the predictions. Experimental results show that STRPM can generate more satisfactory results compared with various existing methods.

arXiv:2203.15836공식 코드 · Papers with Code

VPTR: Efficient Transformers for Video Prediction

In this paper, we propose a new Transformer block for video future frames prediction based on an efficient local spatial-temporal separation attention mechanism. Based on this new Transformer block, a fully autoregressive video future frames prediction Transformer is proposed. In addition, a non-autoregressive video prediction Transformer is also proposed to increase the inference speed and reduce the accumulated inference errors of its autoregressive counterpart. In order to avoid the prediction of very similar future frames, a contrastive feature loss is applied to maximize the mutual information between predicted and ground-truth future frame features. This work is the first that makes a formal comparison of the two types of attention-based video future frames prediction models over different scenarios. The proposed models reach a performance competitive with more complex state-of-the-art models. The source code is available at \emph{https://github.com/XiYe20/VPTR}.

arXiv:2203.13880공식 코드 · Papers with Code

Reinforcement Learning with Action-Free Pre-Training from Videos

Recent unsupervised pre-training methods have shown to be effective on language and vision domains by learning useful representations for multiple downstream tasks. In this paper, we investigate if such unsupervised pre-training methods can also be effective for vision-based reinforcement learning (RL). To this end, we introduce a framework that learns representations useful for understanding the dynamics via generative pre-training on videos. Our framework consists of two phases: we pre-train an action-free latent video prediction model, and then utilize the pre-trained representations for efficiently learning action-conditional world models on unseen environments. To incorporate additional action inputs during fine-tuning, we introduce a new architecture that stacks an action-conditional latent prediction model on top of the pre-trained action-free prediction model. Moreover, for better exploration, we propose a video-based intrinsic bonus that leverages pre-trained representations. We demonstrate that our framework significantly improves both final performances and sample-efficiency of vision-based RL in a variety of manipulation and locomotion tasks. Code is available at https://github.com/younggyoseo/apv.

논문 ID: document/8784821

Z-Order Recurrent Neural Networks for Video Prediction

We present a Z-Order RNN (Znet) for predicting future video frames given historical observations. There are two main contributions respectively in deterministic and stochastic modeling perspective. First, we propose a new RNN architecture for modeling the deterministic dynamics, which updates hidden states along a z-order curve to enhance the consistency of the features of mirrored layers. Second, we introduce an adversarial training approach to two-stream Znet for modeling the stochastic variations, which forces the Znet-Predictor to imitate the behavior of the Znet-Probe. This two-stream architecture enables the adversarial training to be conducted in the feature space instead of the image space. Our model achieves the state-of-the-art prediction accuracy on two video datasets.