VideoFlow: A Flow-Based Generative Model for Video

By Manoj Kumar, Mohammad Babaeizadeh, Dumitru Erhan, Chelsea Finn, Sergey Levine, Laurent Dinh, Durk Kingma. 2019

135 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Generative Models

Citations in the corpus

Scaling Autoregressive Video Models (2019, methodology): “From Table 1a, we see that the small models already reduce the perplexity in terms of bits/dim by almost 20% compared to the recently proposed VideoFlow model (Kumar et al., 2019) with our large model (L) reducing perplexity even further to a 25% improvement.”
Flow Contrastive Estimation of Energy-Based Models (2019, methodology): “The following are the two scenarios for estimating qα: (1) Generative modeling byMLE [7, 8, 32, 17, 3, 37, 68], based on minα KL(pdata‖qα), where again Epdata can be approximated by average over observed examples.”
Learning Energy-Based Spatial-Temporal Generative ConvNets for Dynamic Patterns (2019, methodology): “Beside the energy-based models and the generator models, there are also flow-based models [66], where the loglikelihoods can be computed in closed form.”
Generative Models for Effective ML on Private, Decentralized Datasets (2019, background): “Deep generative models of various forms (Kingma & Welling, 2013; Goodfellow et al., 2014; Kumar et al., 2019; Radford et al., 2019) have garnered a considerable amount of research interest, particularly for high dimensional spaces where explicit modeling is intractable.”
Neural Spline Flows (2019, background): “In recent years, normalizing flows have received widespread attention in the machine-learning literature, seeing successful use in density estimation [9, 41], variational inference [29, 35, 44, 55], image, audio and video generation [25, 27, 31, 43], likelihood-free inference [42], and learning maxi”
Normalizing Flows for Probabilistic Modeling and Inference (2019, methodology): “Finally, Kumar et al. (2019) propose a normalizing flow for modeling video data, adapting the Glow architecture (Kingma and Dhariwal, 2018) to synthesize raw RGB frames.”
Actionable Models: Unsupervised Offline Reinforcement Learning of Robotic Skills (2021, methodology): “The problem of learning dynamics models is a special case of a more general problem of auto-regressive time-series prediction (e.g., video prediction (Mathieu et al., 2015; Denton & Fergus, 2018; Kumar et al., 2019b)).”
Compression with Flows via Local Bits-Back Coding (2019, methodology): “The cost can be made negligible by coding long sequences, such as entire test sets or audio or video with large numbers of frames [35, 28].”
Moser Flow: Divergence-based Generative Modeling on Manifolds (2021, background): “The major successes of deep generative models in recent years are primarily in domains involving Euclidean data, such as images (Dhariwal and Nichol, 2021), text (Brown et al., 2020), and video (Kumar et al., 2019).”
Adversarial Video Generation on Complex Datasets (2019, background): “…Lee et al., 2018; Hsieh et al., 2018), auto-regressive models (Ranzato et al., 2014; Srivastava et al., 2015; Kalchbrenner et al., 2017; Weissenborn et al., 2019), normalizing ﬂows (Kumar et al., 2019), and GANs (Mathieu et al., 2015; Vondrick et al., 2016; Saito et al., 2017; Saito & Saito, 2018).”
Improved Conditional VRNNs for Video Prediction (2019, background): “Concurrent work [20] has proposed a fully invertible model for video.”
PointFlow: 3D Point Cloud Generation With Continuous Normalizing Flows (2019, background): “image generation [24, 9, 8], video generation [27], and voice synthesis [35].”
Integer Discrete Flows and Lossless Compression (2019, background): “Deep generative models permit learning these complicated distributions from data, and have demonstrated their effectiveness in image, video and audio modeling [19, 20, 23].”
Hierarchical Relational Inference (2020, methodology): “using a variational approach (Babaeizadeh et al. 2018; Denton and Fergus 2018; Kumar et al. 2020), or (additionally) require adversarial training (Lee et al.”
IDF++: Analyzing and Improving Integer Discrete Flows for Lossless Compression (2020, background): “Recently these models have demonstrated promising performance in image [17, 27], audio [24] and video [29] modelling.”
Unsupervised Disentanglement of Pose, Appearance and Background from Images and Videos (2020, methodology): “Several methods have been proposed for video prediction problem, which in high level can be categorized as flowbased methods where methods model motion transformation via flows [22], [29], [37], [39] and latent-representation based methods [11], [12], [19], [31], [42].”