Bagging, Boosting and Random Forest

For more info about our services contact : help@bestpfe.com

Table of contents

Amorcer la perception écologique d’un robot par exploration et interactions
Résumé
Introduction
Le contexte
Travaux et domaines proches
La carte de pertinence
Les modèles de mélange collaborateurs
La carte d’affordances
Conclusion
1 Introduction
2 Context
2.1 Introduction
2.2 Developmental Robotics
2.3 Classification problem and learning methods
2.3.1 Classification Problem
2.3.2 Semi-supervised and Active learning
2.4 Related Works
2.4.1 Related Domains
2.4.2 Putting it all together
2.5 Conclusion
3 Background
3.1 Introduction
3.2 Gaussian Mixture Models
3.2.1 Classical GMM
3.2.2 Geometrical analysis of Multivariate Normal Distribution
3.3 Image Processing
3.3.1 Supervoxels Segmentation
3.3.2 Visual Features and descriptors extraction
3.4 Conclusion
4 Collaborative Mixture Models
4.1 Introduction
4.2 Online Learning
4.2.1 Support Vector Machines
4.2.2 Bagging, Boosting and Random Forest
4.2.3 Mixture Models
4.3 Gaussian Mixture Models with an unknown number of components .
4.4 Query Strategies in Active Learning
4.4.1 Uncertainty Sampling
4.4.2 Other Query Strategies
4.5 Definition of the classifier
4.6 Algorithm
4.6.1 Split and Merge operation
4.6.2 Query Strategy
4.7 Conclusion
5 Relevance Map
5.1 Introduction
5.2 Interactive Perception
5.2.1 Object Segmentation by Interactive Perception
5.2.2 Discussion
5.3 Saliency Map
5.3.1 Salient Object Detection
5.3.2 Discussion
5.4 Method
5.4.1 Overview
5.4.2 Features Extraction
5.4.3 Building the Relevance Map
5.4.4 Query Strategy
5.4.5 Push Primitive
5.4.6 Change Detection
5.5 Experiments
5.5.1 Protocol
5.5.2 Classification Quality Measures
5.6 Results
5.6.1 Simplified Setups
5.6.2 Real World Experiments
5.7 Discussion and Future work
5.8 Conclusion
6 Exstensive study of CMMs
6.1 Introduction
6.2 Splitting and Merging
6.2.1 Protocol
6.2.2 Results
6.3 Query strategy
6.4 Supervoxel features
6.4.1 Protocol
6.4.2 Results
6.5 Discussion and Future works
6.6 Conclusion
7 Affordances Map
7.1 Introduction
7.2 Affordances
7.2.1 Foundation and Definition(s)
7.2.2 Affordances in Robotic
7.2.3 Learning affordances from local features
7.3 Method
7.3.1 Affordances Formalisation
7.3.2 Classifier
7.3.3 Primitives and Effects Detection
7.4 Experiments
7.5 Results
7.6 Discussion and Future Works
7.7 Conclusion
8 Conclusion and Discussions
8.1 Summary of the contributions
8.2 Discussion and Limitations
8.2.1 CMMs Limitations
8.2.2 Supervoxels
8.2.3 Learning from local features
8.3 Future Works
8.3.1 Possible Improvements
8.3.2 Next Developmental Steps
Bibliography
A Singular Value Decomposition (SVD)