Active learning of visual descriptors for grasping using non-parametric smoothed beta distributions

By Luis Montesano, Manuel Lopes. 2012

47 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

Supersizing self-supervision: Learning to grasp from 50K tries and 700 robot hours (2015, methodology): “Also, in order to gather data as close to real world test conditions, we perform trial and error grasping experiments in cluttered environment.”
Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics (2017, methodology): “The time cost of generating samples on a physical robot led to the development active methods for acquiring grasping experiences such as Multi-Armed Bandits using Correlated Beta Processes [37] or Prior Confidence Bounds [39].”
Data-Driven Grasp Synthesis—A Survey (2013, background): “Instead of assuming the availability of a labeled data set, Montesano and Lopes [79] allow the robot to autonomously explore which features encode graspability.”
Dex-Net 1.0: A cloud-based network of 3D objects for robust grasp planning using a Multi-Armed Bandit model with correlated rewards (2016, methodology): “Montesano and Lopes [31] used Continuous Correlated Beta Processes [11] to actively acquire grasp executions on a physical robot using image filters to measure similarity.”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, methodology): “As these parameters are often lower dimensional than full skills, active learning and multi-armed bandit methods can be used to select suitable argument values [267, 255, 221].”
The strategic student approach for life-long exploration and learning (2012, background): “A third family of approaches have considered internal mechanisms for active choice of either training examples [8]– [10], tasks to be explored [11], learning methods to be used [12], [13] or questions to pose to a teacher [14].”
Learning Grasping Interaction with Geometry-aware 3D Representations (2017, methodology): “A common approach for robotic grasping is to detect the optimal grasping location from 2D visual input (RGB or RGBD images) Saxena et al. [2008], Montesano and Lopes [2012], Lenz et al. [2015], Pinto and Gupta [2016].”