Time-Contrastive Networks: Self-Supervised Learning from Video

Sermanet, Pierre; Lynch, Corey; Chebotar, Yevgen; Hsu, Jasmine; Schaal, Stefan; Levine, Sergey

doi:10.48550/arxiv.1704.06888

Cited by 40 publications

(70 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A third class of techniques that is able to perform imitation learning without requiring knowledge of actions includes those that first focus on learning a representation of the task and then use an RL method with a predefined surrogate reward over that representation. For example, Gupta et al (2017) have proposed an invariant feature space to transfer skills between agents with different embodiments, Liu et al (2017) have presented a network architecture which is capable of handling differences in viewpoints and contexts between the imitator and the demonstrator, and Sermanet et al (2017) have proposed a time-contrastive network which is invariant to both different embodiments and viewpoints. While these techniques represent significant advances in representation learning, each of them uses the same surrogate reward function, i.e., the proximity of the imitator's and demonstrator's encoded representation at each time step.…”

Section: Related Workmentioning

confidence: 99%

“…• Time Contrastive Networks (TCN) (Sermanet et al, 2017): TCNs use a triplet loss to train a neural network to learn an encoded form of the task at each time step. This loss function brings the states that occur in a small time-window closer together in the embedding space and pushes the ones from distant time-steps far apart.…”

Section: Experimental Setup and Implementation Detailsmentioning

confidence: 99%

“…Broadly speaking, the IfO problem consists of two major subproblems: (1) perception of the demonstrations, i.e., extracting useful features from raw visual data, and (2) learning a control policy using the extracted features. Most IfO work thus far (Liu et al, 2017;Sermanet et al, 2017) has focused on perception and not on control. While powerful methods for perceiving the demonstrations have been developed, the control problem is solved via relatively simple means, i.e., reinforcement learning over a pre-defined reward function.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Generative Adversarial Imitation from Observation

Torabi,

Warnell,

Stone

2018

Preprint

View full text Add to dashboard Cite

Imitation from observation (IfO) is the problem of learning directly from state-only demonstrations without having access to the demonstrator's actions. The lack of action information both distinguishes IfO from most of the literature in imitation learning, and also sets it apart as a method that may enable agents to learn from a large set of previously inapplicable resources such as internet videos. In this paper, we propose both a general framework for IfO approaches and also a new IfO approach based on generative adversarial networks called generative adversarial imitation from observation (GAIfO). We conduct experiments in two different settings: (1) when demonstrations consist of low-dimensional, manuallydefined state features, and (2) when demonstrations consist of high-dimensional, raw visual data. We demonstrate that our approach performs comparably to classical imitation learning approaches (which have access to the demonstrator's actions) and significantly outperforms existing imitation from observation methods in high-dimensional simulation environments.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Experimental Setup and Implementation Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Generative Adversarial Imitation from Observation

Torabi,

Warnell,

Stone

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Self-supervised learning from sequences. Previous work in contrastive learning for sequential data often leverages a slowness assumption to use nearby samples as positive examples and farther samples as negative examples (Oord et al, 2018;Sermanet et al, 2018;Dwibedi et al, 2019;Le-Khac et al, 2020;Banville et al, 2020). Contrastive predictive coding (CPC) (Oord et al, 2018) builds upon the idea of temporal contrastive learning by building an AR-model that predicts future points given previous observed timesteps.…”

Section: Related Workmentioning

confidence: 99%

“…Temporal shift: As in previous work in temporal contrastive learning (Oord et al, 2018;Sermanet et al, 2018;Dwibedi et al, 2019;Le-Khac et al, 2020;Banville et al, 2020), we can use nearby samples as positive examples for one another.…”

Section: B3 Augmentations For Neural Datamentioning

confidence: 99%

Mine Your Own vieW: Self-Supervised Learning Through Across-Sample Prediction

Azabou¹,

Azar²,

Liu³

et al. 2021

Preprint

View full text Add to dashboard Cite

State-of-the-art methods for self-supervised learning (SSL) build representations by maximizing the similarity between different augmented "views" of a sample. Because these approaches try to match views of the same sample, they can be too myopic and fail to produce meaningful results when augmentations are not sufficiently rich. This motivates the use of the dataset itself to find similar, yet distinct, samples to serve as views for one another. In this paper, we introduce Mine Your Own vieW (MYOW), a new approach for building across-sample prediction into SSL. The idea behind our approach is to actively mine views, finding samples that are close in the representation space of the network, and then predict, from one sample's latent representation, the representation of a nearby sample. In addition to showing the promise of MYOW on standard datasets used in computer vision, we highlight the power of this idea in a novel application in neuroscience where rich augmentations are not already established. When applied to neural datasets, MYOW outperforms other self-supervised approaches in all examples (in some cases by more than 10%), and surpasses the supervised baseline for most datasets. By learning to predict the latent representation of similar samples, we show that it is possible to learn good representations in new domains where augmentations are still limited.

show abstract

MT-VAE: Learning Motion Transformations to Generate Multimodal Human Dynamics

Yan

Rastogi

Villegas

et al. 2018

Lecture Notes in Computer Science

124

View full text Add to dashboard Cite

Long-term human motion can be represented as a series of motion modes-motion sequences that capture short-term temporal dynamics-with transitions between them. We leverage this structure and present a novel Motion Transformation Variational Auto-Encoders (MT-VAE) for learning motion sequence generation. Our model jointly learns a feature embedding for motion modes (that the motion sequence can be reconstructed from) and a feature transformation that represents the transition of one motion mode to the next motion mode. Our model is able to generate multiple diverse and plausible motion sequences in the future from the same input. We apply our approach to both facial and full body motion, and demonstrate applications like analogy-based motion transfer and video synthesis. * Work partially done during internship with Adobe Research.

show abstract

Time-Contrastive Networks: Self-Supervised Learning from Video

Cited by 40 publications

References 0 publications

Generative Adversarial Imitation from Observation

Generative Adversarial Imitation from Observation

Mine Your Own vieW: Self-Supervised Learning Through Across-Sample Prediction

MT-VAE: Learning Motion Transformations to Generate Multimodal Human Dynamics

Contact Info

Product

Resources

About