Adapting Deep Visuomotor Representations with Weak Pairwise Constraints

Tzeng, Eric; Devin, Coline; Hoffman, Judy; Finn, Chelsea; Abbeel, Pieter; Levine, Sergey; Saenko, Kate; Darrell, Trevor

doi:10.1007/978-3-030-43089-4_44

Cited by 70 publications

(70 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To reduce the number of real-world images required for learning visuo-motor policies, a method of adapting visual representations from simulated to real environments was proposed, achieving a success rate of 79.2% in a ''hook loop'' task, with 10 times fewer real-world images (Tzeng et al, 2016). Another example of vision-based policy transfer is progressive neural networks, which are proposed to improve transfer and avoid catastrophic forgetting when learning complex sequences of tasks (Rusu et al, 2016).…”

Section: Transfer Learningmentioning

confidence: 99%

“…Therefore, some methods were proposed to reduce the cost of collecting a large amount of real-world data by using simulated or synthetic data (Bateux et al, 2018; D'Innocente et al, 2017; James et al, 2017; Tobin et al, 2017). Some others tried to make use of both simulated and real data for a more balanced solution (Fitzgerald et al, 2015; Tzeng et al, 2016). A particular approach is modular deep Q-networks for learning a planar reaching task in simulation and then transferring to real environments with a small number of labeled real-world images (Zhang et al, 2017a,b).…”

Section: Introductionmentioning

confidence: 99%

“…Similar approaches were also proposed for other classification tasks, such as adversarial discriminative domain adaptation (ADDA) for handwritten digit recognition (Tzeng et al, 2017) and incremental adversarial domain adaptation for drivable-path segmentation (Wulfmeier et al, 2018). Some works also tried to use adversarial approaches for regression transfer such as domain confusion with weak pairwise constraints for deep visuo-motor representation adaptation (Tzeng et al, 2016).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Adversarial discriminative sim-to-real transfer of visuo-motor policies

Zhang

Leitner

et al. 2019

The International Journal of Robotics Research

View full text Add to dashboard Cite

Various approaches have been proposed to learn visuo-motor policies for real-world robotic applications. One solution is first learning in simulation then transferring to the real world. In the transfer, most existing approaches need real-world images with labels. However, the labelling process is often expensive or even impractical in many robotic applications. In this paper, we propose an adversarial discriminative sim-to-real transfer approach to reduce the cost of labelling real data. The effectiveness of the approach is demonstrated with modular networks in a table-top object reaching task where a 7 DoF arm is controlled in velocity mode to reach a blue cuboid in clutter through visual observations. The adversarial transfer approach reduced the labelled real data requirement by 50%. Policies can be transferred to real environments with only 93 labelled and 186 unlabelled real images. The transferred visuo-motor policies are robust to novel (not seen in training) objects in clutter and even a moving target, achieving a 97.8% success rate and 1.8 cm control accuracy.Some others tried to make use of both simulated and real data for a more balanced solution (Fitzgerald et al. 2015;

show abstract

Section: Transfer Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adversarial discriminative sim-to-real transfer of visuo-motor policies

Zhang

Leitner

et al. 2019

The International Journal of Robotics Research

View full text Add to dashboard Cite

show abstract

“…In this sort of methods, the difference between source and target domains is usually reduced by optimizing the MMD in Reproducing Kernel Hilbert Space (RKHS), and the feature representation with domain invariance needs to be learned. Adversarial-based methods [ 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 ] use a discriminator to distinguish whether the data comes from the source domain or the target domain, so as to increase the domain confusion and minimize the distance between the source and the target domain distribution.…”

Section: Introductionmentioning

confidence: 99%

C2DAN: An Improved Deep Adaptation Network with Domain Confusion and Classifier Adaptation

Sun

Chen

Liang

et al. 2020

Sensors

View full text Add to dashboard Cite

Deep neural networks have been successfully applied in domain adaptation which uses the labeled data of source domain to supplement useful information for target domain. Deep Adaptation Network (DAN) is one of these efficient frameworks, it utilizes Multi-Kernel Maximum Mean Discrepancy (MK-MMD) to align the feature distribution in a reproducing kernel Hilbert space. However, DAN does not perform very well in feature level transfer, and the assumption that source and target domain share classifiers is too strict in different adaptation scenarios. In this paper, we further improve the adaptability of DAN by incorporating Domain Confusion (DC) and Classifier Adaptation (CA). To achieve this, we propose a novel domain adaptation method named C2DAN. Our approach first enables Domain Confusion (DC) by using a domain discriminator for adversarial training. For Classifier Adaptation (CA), a residual block is added to the source domain classifier in order to learn the difference between source classifier and target classifier. Beyond validating our framework on the standard domain adaptation dataset office-31, we also introduce and evaluate on the Comprehensive Cars (CompCars) dataset, and the experiment results demonstrate the effectiveness of the proposed framework C2DAN.

show abstract

“…In effect, gap models were applied as generative models to produce data for difficult‐to‐measure dynamics needed in developing approximations. This is conceptually similar to the sim2real paradigm currently at the forefront of artificial intelligence and robotics research (Tzeng et al ). Using gap models as data generators was necessary due to a lack of detailed long‐term observational data.…”

Section: Introductionmentioning

confidence: 99%

Implementation of the perfect plasticity approximation with biogeochemical compartments in R

Erickson

Strigul

2020

Ecography

View full text Add to dashboard Cite

Modeling forest ecosystems is a landmark challenge in science, due to the complexity of the processes involved and their importance in predicting future planetary conditions. While there are a number of open-source forest biogeochemistry models, few papers exist detailing the software development approach used to develop these models. This has left many forest biogeochemistry models large, opaque and/or difficult to use, typically implemented in compiled languages for speed. Here, we present a forest biogeochemistry model from the SORTIE-PPA class of models, PPA-SiBGC. Our model is based on the perfect plasticity approximation with simple biogeochemistry compartments and uses empirical vegetation dynamics rather than detailed prognostic processes to drive the estimation of carbon and nitrogen fluxes. This allows our model to be used with traditional forest inventory data, making it widely applicable and simple to parameterize. We detail the conceptual design of the model as well as the software implementation in the R language for statistical computing. Our aim is to provide a useful tool for the biogeochemistry modeling community that demonstrates the importance of vegetation dynamics in biogeochemical models.

show abstract

Adapting Deep Visuomotor Representations with Weak Pairwise Constraints

Cited by 70 publications

References 17 publications

Adversarial discriminative sim-to-real transfer of visuo-motor policies

Adversarial discriminative sim-to-real transfer of visuo-motor policies

C2DAN: An Improved Deep Adaptation Network with Domain Confusion and Classifier Adaptation

Implementation of the perfect plasticity approximation with biogeochemical compartments in R

Contact Info

Product

Resources

About