Convolutional Decision Trees for Feature Learning and Segmentation

2014 ◽  
pp. 95-106 ◽  
Author(s):  
Dmitry Laptev ◽  
Joachim M. Buhmann
Keyword(s):  
Decision Trees ◽  
Feature Learning

scholarly journals Deep Supervised Hashing with Nonlinear Projections

2017 ◽  
Author(s):  
Sen Su ◽  
Gang Chen ◽  
Xiang Cheng ◽  
Rong Bi
Keyword(s):  
Decision Trees ◽  
Network Architecture ◽  
Large Scale ◽  
Feature Learning ◽  
Superior Performance ◽  
Soft Decision ◽  
Deep Hashing ◽  
Nonlinear Projection ◽  
Search Speed ◽  
Projection Functions

Hashing has attracted broad research interests in large scale image retrieval due to its high search speed and efficient storage. Recently, many deep hashing methods have been proposed to perform simultaneous nonlinear feature learning and hash projection learning, which have shown superior performance compared to hand-crafted feature based hashing methods. Nonlinear projection functions have shown their advantages over the linear ones due to their powerful generalization capabilities. To improve the performance of deep hashing methods by generalizing projection functions, we propose the idea of implementing a pure nonlinear deep hashing network architecture. By consolidating the above idea, this paper presents a Deep Supervised Hashing architecture with Nonlinear Projections (DSHNP). In particular, soft decision trees are adopted as the nonlinear projection functions, since they can generate differentiable nonlinear outputs and can be trained with deep neural networks in an end-to-end way. Moreover, to make the hash codes as independent as possible, we design two regularizers imposed on the parameter matrices of the leaves in the soft decision trees. Extensive evaluations on two benchmark image datasets show that the proposed DSHNP outperforms several state-of-the-art hashing methods.

Optimizing Diagnostic Test Sequences: The Probability Modifying Plot

1999 ◽  
Vol 38 (01) ◽  
pp. 50-55 ◽  
Author(s):  
P. F. de Vries Robbé ◽  
A. L. M. Verbeek ◽  
J. L. Severens
Keyword(s):  
Urinary Tract Infection ◽  
Urinary Tract ◽  
Tract Infection ◽  
Decision Trees ◽  
Diagnostic Tests ◽  
Test Sequence ◽  
Optimal Sequence ◽  
Specific Test ◽  
Total Cost ◽  
Modelling Techniques

Abstract:The problem of deciding the optimal sequence of diagnostic tests can be structured in decision trees, but unmanageable bushy decision trees result when the sequence of two or more tests is investigated. Most modelling techniques include tests on the basis of gain in certainty. The aim of this study was to explore a model for optimizing the sequence of diagnostic tests based on efficiency criteria. The probability modifying plot shows, when in a specific test sequence further testing is redundant and which costs are involved. In this way different sequences can be compared. The model is illustrated with data on urinary tract infection. The sequence of diagnostic tests was optimized on the basis of efficiency, which was either defined as the test sequence with the least number of tests or the least total cost for testing. Further research on the model is needed to handle current limitations.

Automated Development of Clinical Strategies Using Multistage Decision Analysis

1986 ◽  
Vol 25 (04) ◽  
pp. 207-214 ◽  
Author(s):  
P. Glasziou
Keyword(s):  
Decision Tree ◽  
Decision Analysis ◽  
Decision Trees ◽  
Optimal Strategy ◽  
Cholestatic Jaundice ◽  
Clinical Strategies ◽  
Simultaneous Study

SummaryThe development of investigative strategies by decision analysis has been achieved by explicitly drawing the decision tree, either by hand or on computer. This paper discusses the feasibility of automatically generating and analysing decision trees from a description of the investigations and the treatment problem. The investigation of cholestatic jaundice is used to illustrate the technique.Methods to decrease the number of calculations required are presented. It is shown that this method makes practical the simultaneous study of at least half a dozen investigations. However, some new problems arise due to the possible complexity of the resulting optimal strategy. If protocol errors and delays due to testing are considered, simpler strategies become desirable. Generation and assessment of these simpler strategies are discussed with examples.

Generation of Correlated Logistic-Normal Random Variates for Medical Decision Trees

1998 ◽  
Vol 37 (03) ◽  
pp. 235-238 ◽  
Author(s):  
M. El-Taha ◽  
D. E. Clark
Keyword(s):  
Monte Carlo ◽  
Decision Trees ◽  
Computer Algebra ◽  
Taylor Series ◽  
Computer Algebra System ◽  
Random Variable ◽  
Monte Carlo Analysis ◽  
Normal Random Variable ◽  
Medical Decision ◽  
Theoretical Results

AbstractA Logistic-Normal random variable (Y) is obtained from a Normal random variable (X) by the relation Y = (ex)/(1 + ex). In Monte-Carlo analysis of decision trees, Logistic-Normal random variates may be used to model the branching probabilities. In some cases, the probabilities to be modeled may not be independent, and a method for generating correlated Logistic-Normal random variates would be useful. A technique for generating correlated Normal random variates has been previously described. Using Taylor Series approximations and the algebraic definitions of variance and covariance, we describe methods for estimating the means, variances, and covariances of Normal random variates which, after translation using the above formula, will result in Logistic-Normal random variates having approximately the desired means, variances, and covariances. Multiple simulations of the method using the Mathematica computer algebra system show satisfactory agreement with the theoretical results.

Context dependent modeling of phones in continuous speech using decision trees

1991 ◽  
Author(s):  
L. R. Bahl ◽  
P. V. de Soutza ◽  
P. S. Gopalakrishnan ◽  
D. Nahamoo ◽  
M. A. Picheny
Keyword(s):  
Decision Trees ◽  
Continuous Speech ◽  
Context Dependent
Sign in / Sign up

Export Citation Format

Share Document