Learning Object Affordances: From Sensory--Motor Coordination to Imitation

By Luis Montesano, Manuel Lopes, Alexandre Bernardino, José Santos-Victor. 2008

438 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering

How researchers see this paper

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

In Artificial Intelligence

In Control and Systems Engineering

Citations in the corpus

Learning human activities and object affordances from RGB-D videos (2012, background): “, 2012) or through observation of the effects of exploratory behaviors (Montesano et al., 2008; Ridge et al., 2009; Moldovan et al., 2012).”
Data-Driven Grasp Synthesis—A Survey (2013, background): “As in [94], the joint distribution over the set of variables inﬂuenc-ing this choice is modeled as a Bayesian network.”
What can I do here? A Theory of Affordances in Reinforcement Learning (2020, background): “, such as STRIPS (Fikes & Nilsson,1971). In AI, researchers have also studied object-affordances, in order to map actions to objects (Slocum et al.,2000;Fitzpatrick et al.,2003;Lopes et al.,2007).Montesano et al. (2008) presented a developmental approach to learning object affordances. Their app”
Learning about objects with human teachers (2009, neutral): “Finally, we present the impact of social versus nonsocial data sets when training SVM classi.ers.”
Domain Randomization and Generative Models for Robotic Grasping (2017, background): “Others use human labeling or self-supervised learning to measure success [28], [29], [43].”
Learning Generalizable Control Programs (2011, background): “and the MultiSensory Autonomous Cognitive Systems group have provided formalisms that capture the statistical relationships between taking actions on objects and observing the effects [20], [29].”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, background): “other approaches have taken advantage of structure specific to robotic manipulation by directing exploration to discover reusable object affordances [268], or learning exploration strategies that exploit the distribution of problems the agent might face via metalearning [425].”
Active learning of visual descriptors for grasping using non-parametric smoothed beta distributions (2012, methodology): “This scheme has been widely used in many of the latest developed systems such as generic visual descriptors combined with planning [9], 3D object centered representations and reinforcement learning[10], 3D multi-modal features combined with a set of predefined grasp strategies [11], single image vis”
Statistical Relational Learning of Object Affordances for Robotic Manipulation (2011, methodology): “Affordances model: relations between objects, actions, effects [5, 6].”
Supervised learning of hidden and non-hidden 0-order affordances and detection in real scenes (2012, methodology): “In robotics, affordances have been primarily exploited in grasping or action-behaviour learning of objects, where 2D motion or colour cues have been related to object shape, e.g., [2] [3].”
Affordance Prediction via Learned Object Attributes (2011, background): “[12] present a Bayesian network model that implicitly represents affor-dances as mappings from action to effect, which are mediated by the visual features of objects.”
Learning Visual Object Categories for Robot Affordance Prediction (2010, background): “Montesano et al. (2008) present a Bayesian network model that implicitly represents affordances as mappings from action to effect, which are mediated by the visual features of objects.”
Learning visual representations for perception-action systems (2011, background): “On the other hand, reactive approaches attempt to learn the empirical manipulative possibilities an object offers with respect to a manipulator (i.e., affordances, Gibson 1979; Montesano et al. 2008), or learn visual aspects that predict graspability (Saxena et al. 2008).”
Abstraction Levels for Robotic Imitation: Overview and Computational Approaches (2010, background): “Repeated interaction with the world allows the robot to understand how the environment behaves under its actions [46,94,95,122].”
Tell me Dave: Context-sensitive grounding of natural language to manipulation instructions (2014, methodology): “Our set of object affordance is inspired by previous work (Kjellström et al., 2011; Koppula et al., 2013; Montesano et al., 2008).”
SceneGrok (2014, background): “…to improve pose estimation, object recognition, action classification, and other related problem in different communities [Fritz et al. 2006; Montesano et al. 2008; Stark et al. 2008; Sun et al. 2010; Hermans et al. 2011; Goldfeder and Allen 2011; Bohg et al. 2013; Koppula and Saxena 2013;…”