Object Recognition with Deep Boltzmann Machines

From AIRWiki
Revision as of 00:48, 28 February 2014 by FrancescoVisin (Talk | contribs) (FrancescoVisin moved page Projects - HOWTO to Object Recognition with Deep Belief Networks: Very misleading title...)

Jump to: navigation, search


This project aims to exploit the ability of Deep Belief Network (DBN) to classify (and generate) multiclass objects. Initially, they will be used to classify road signs with the purpose to help the navigation of an hypotetical autonomous car. Since the initial results on the classification tasks of road signs using DBN are very good, we are pretty confident about the quality of the project. After that, we will consider the possibility to extend the task to other domains.

Short Description: DBN for classification tasks
Coordinator: MatteoMatteucci (matteo.matteucci@polimi.it)
Tutor: MatteoMatteucci (matteo.matteucci@polimi.it), FrancescoVisin (visin@elet.polimi.it)
Students: CarloDEramo (carlo.deramo@polimi.it)
Research Area: Robotics
Research Topic: Machine Learning
Start: 2014/03/1
Status: Active
Level: Ms

Deep learning and Deep Belief Network

Deep learning is a set of algorithms in machine learning that attempt to model high-level abstractions in data by using architectures composed of multiple non-linear transformations. A deep belief network (DBN) is used for deep learning tasks, it is a generative graphical model, or alternatively a type of deep neural network, composed of multiple layers of latent variables ("hidden units"), with connections between the layers but not between units within each layer. More precisely, each layer acts as a feature detector on inputs and serves as the visible layer for the next. In this way it is possible to have a fast unsupervised training procedure which will be very useful for the purpose of our project due to the complexity of the domain. We will use Pylearn2 framework which offers training procedures of deep neural networks. During the project we will go deep in the analysis of this kind of network studying their potentiality and flexibility; we'll try different configuration of network in order to obtain the good performances we're looking for.