블로그 이미지
Leeway is... the freedom that someone has to take the action they want to or to change their plans.
maetel

Notice

Recent Post

Recent Comment

Recent Trackback

Archive

calendar

1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31
  • total
  • today
  • yesterday

Category

'artificial neural network'에 해당되는 글 2건

  1. 2010.12.14 Jordan & Bishop "Neural Networks"
  2. 2010.12.11 Michael I. Jordan, "Generic constraints on underspecified target trajectories"
2010. 12. 14. 20:02 Computer Vision
Michael I. Jordan & Christopher M. Bishop, "Neural Networks", In Tucker, A. B. (Ed.) CRC Handbook of Computer Science, Boca Raton, FL: CRC Press, 1997.
download: http://www.cs.berkeley.edu/~jordan/papers/crc.ps







1. Introduction


Neural network methods have had their greatest impact in problems where statistical issues dominate and where data are easily obtained.

"conjunction of graphical algorithms and probability theory":
A neural network is first and foremost a graph with patterns represented in terms of numerical values attached to the nodes of the graph and transformations between patterns achieved via simple message-passing algorithms. Many neural network architectures, however, are also statistical processors, characterized by making particular probabilistic assumptions about data.


Based on a source of training data, the aim is to produce a statistical model of the process from which the data are generated so as to allow the best predictions to be made for new data.

statistical modeling - density estimation (unsupervised learning), classification & regression

density estimation ("unsupervised learning")
: to model the unconditional distribution of data described by some vector
- to train samples and a network model to build a representation of the probability density 
- to label regions for a new input vector
 
classification & regression ("supervised learning")
: to distinguish between input variables and target variables
- to assign each input vector to one of classes and target variables to class labels 
-> estimation of conditional densities from the joint input-target space



2. Representation

2.1 Density estimation

To form an explicit model of the input density

Gaussian mixture distribution







2.2 Linear regression and linear discriminants






posted by maetel
2010. 12. 11. 01:58 Computer Vision
Michael I. Jordan, Generic constraints on underspecified target trajectories, Proceedings of international conference on neural networks, (1989), 217-225
http://dx.doi.org/10.1109/IJCNN.1989.118584


informed by 함교수님

cf.

Michael I. Jordan



Introduction

connectionist networks

feedforward controller

forward model


activation patterns (in a network)

motor learning

interpretation


visible units & hidden units

The output units of such a network are hidden with respect to learning and yet are visible given their direct connection to the environment.

task space 
articulatory space



Forward models of the environment


The forward modeling approach assumes that the solution to this minimization problem is based on the computation of a gradient.


'Computer Vision' 카테고리의 다른 글

Seam Carving  (0) 2011.01.17
Jordan & Bishop "Neural Networks"  (0) 2010.12.14
Donald M. Wiberg [Schaum's Outline of Theory and Problems of State Space and Linear System]  (0) 2010.12.09
Winsock  (0) 2010.11.26
camera Firefly (FFMV-03M2M)  (0) 2010.11.25
posted by maetel