Timezone: »
Contemporary sensorimotor learning approaches typically start with an existing complex agent (e.g., a robotic arm), which they learn to control. In contrast, this paper investigates a modular co-evolution strategy: a collection of primitive agents learns to dynamically self-assemble into composite bodies while also learning to coordinate their behavior to control these bodies. Each primitive agent consists of a limb with a motor attached at one end. Limbs may choose to link up to form collectives. When a limb initiates a link-up action and there is another limb nearby, the latter is magnetically connected to the 'parent' limb's motor. This forms a new single agent, which may further link with other agents. In this way, complex morphologies can emerge, controlled by a policy whose architecture is in explicit correspondence with the morphology. We evaluate the performance of these dynamic and modular agents in simulated environments. We demonstrate better generalization to test-time changes both in the environment, as well as in the structure of the agent, compared to static and monolithic baselines. Project videos and source code are provided in the supplementary material.
Author Information
Deepak Pathak (UC Berkeley, FAIR, CMU)
Chris Lu (UC Berkeley and Covariant.ai)
Trevor Darrell (UC Berkeley)
Phillip Isola (Massachusetts Institute of Technology)
Alexei Efros (UC Berkeley)
Related Events (a corresponding poster, oral, or spotlight)
-
2019 Poster: Learning to Control Self-Assembling Morphologies: A Study of Generalization via Modularity »
Wed Dec 11th 01:30 -- 03:30 AM Room East Exhibition Hall B + C
More from the Same Authors
-
2020 Poster: Supervised Contrastive Learning »
Prannay Khosla · Piotr Teterwak · Chen Wang · Aaron Sarna · Yonglong Tian · Phillip Isola · Aaron Maschinot · Ce Liu · Dilip Krishnan -
2020 Poster: What Makes for Good Views for Contrastive Learning? »
Yonglong Tian · Chen Sun · Ben Poole · Dilip Krishnan · Cordelia Schmid · Phillip Isola -
2020 Poster: Auxiliary Task Reweighting for Minimum-data Learning »
Baifeng Shi · Judy Hoffman · Kate Saenko · Trevor Darrell · Huijuan Xu -
2020 Poster: Space-Time Correspondence as a Contrastive Random Walk »
Allan Jabri · Andrew Owens · Alexei Efros -
2020 Oral: Space-Time Correspondence as a Contrastive Random Walk »
Allan Jabri · Andrew Owens · Alexei Efros -
2020 Poster: Swapping Autoencoder for Deep Image Manipulation »
Taesung Park · Jun-Yan Zhu · Oliver Wang · Jingwan Lu · Eli Shechtman · Alexei Efros · Richard Zhang -
2020 Poster: Fighting Copycat Agents in Behavioral Cloning from Observation Histories »
Chuan Wen · Jierui Lin · Trevor Darrell · Dinesh Jayaraman · Yang Gao -
2019 Workshop: AI for Humanitarian Assistance and Disaster Response »
Ritwik Gupta · Robin Murphy · Trevor Darrell · Eric Heim · Zhangyang Wang · Bryce Goodman · Piotr BiliĆski -
2019 Poster: Third-Person Visual Imitation Learning via Decoupled Hierarchical Controller »
Pratyusha Sharma · Deepak Pathak · Abhinav Gupta -
2019 Poster: Compositional Plan Vectors »
Coline Devin · Daniel Geng · Pieter Abbeel · Trevor Darrell · Sergey Levine -
2018 Poster: Speaker-Follower Models for Vision-and-Language Navigation »
Daniel Fried · Ronghang Hu · Volkan Cirik · Anna Rohrbach · Jacob Andreas · Louis-Philippe Morency · Taylor Berg-Kirkpatrick · Kate Saenko · Dan Klein · Trevor Darrell -
2017 Poster: Toward Multimodal Image-to-Image Translation »
Jun-Yan Zhu · Richard Zhang · Deepak Pathak · Trevor Darrell · Alexei Efros · Oliver Wang · Eli Shechtman -
2016 Workshop: Machine Learning for Intelligent Transportation Systems »
Li Erran Li · Trevor Darrell -
2014 Poster: Do Convnets Learn Correspondence? »
Jonathan L Long · Ning Zhang · Trevor Darrell -
2014 Poster: LSDA: Large Scale Detection through Adaptation »
Judy Hoffman · Sergio Guadarrama · Eric Tzeng · Ronghang Hu · Jeff Donahue · Ross Girshick · Trevor Darrell · Kate Saenko -
2014 Poster: Weakly-supervised Discovery of Visual Pattern Configurations »
Hyun Oh Song · Yong Jae Lee · Stefanie Jegelka · Trevor Darrell -
2013 Poster: Visual Concept Learning: Combining Machine Vision and Bayesian Generalization on Concept Hierarchies »
Yangqing Jia · Joshua T Abbott · Joseph L Austerweil · Tom Griffiths · Trevor Darrell -
2012 Poster: Learning with Recursive Perceptual Representations »
Oriol Vinyals · Yangqing Jia · Li Deng · Trevor Darrell -
2012 Poster: Timely Object Recognition »
Sergey K Karayev · Tobi Baumgartner · Mario Fritz · Trevor Darrell -
2011 Workshop: Integrating Language and Vision »
Raymond Mooney · Trevor Darrell · Kate Saenko -
2011 Poster: Heavy-tailed Distances for Gradient Based Image Descriptors »
Yangqing Jia · Trevor Darrell -
2011 Poster: Understanding the Intrinsic Memorability of Images »
Phillip Isola · Devi Parikh · Antonio Torralba · Aude Oliva -
2010 Poster: Factorized Latent Spaces with Structured Sparsity »
Yangqing Jia · Mathieu Salzmann · Trevor Darrell -
2010 Poster: Size Matters: Metric Visual Search Constraints from Monocular Metadata »
Mario J Fritz · Kate Saenko · Trevor Darrell -
2009 Poster: Learning to Hash with Binary Reconstructive Embeddings »
Brian Kulis · Trevor Darrell -
2009 Spotlight: Learning to Hash with Binary Reconstructive Embeddings »
Brian Kulis · Trevor Darrell -
2009 Poster: An Additive Latent Feature Model for Transparent Object Recognition »
Mario J Fritz · Michael J Black · Gary R Bradski · Trevor Darrell -
2009 Poster: Filtering Abstract Senses From Image Search Results »
Kate Saenko · Trevor Darrell -
2009 Oral: An Additive Latent Feature Model for Transparent Object Recognition »
Mario J Fritz · Michael J Black · Gary R Bradski · Trevor Darrell -
2008 Poster: Unsupervised Learning of Visual Sense Models for Polysemous Words »
Kate Saenko · Trevor Darrell -
2008 Spotlight: Unsupervised Learning of Visual Sense Models for Polysemous Words »
Kate Saenko · Trevor Darrell -
2006 Poster: Approximate Correspondences in High Dimensions »
Kristen Grauman · Trevor Darrell -
2006 Spotlight: Approximate Correspondences in High Dimensions »
Kristen Grauman · Trevor Darrell