Neural networks (or artificial neural networks) learn by experience, generalize from previous experiences to new ones, and can make decisions. A neural network can be thought of as a black box non-parametric classifier.

The CMRM is a non-parametric model for image annotation. CMRM represents an image as a bag of regions and annotates a test image by estimating the joint probability of a word w and its viusal blobs.

Support Vector Machine

SVMs are classifiers that construct a maximum margin hyperplane between positive and negative examples, which is then used to classify unseen examples.

Random Forests

Random Forests are collections of decision trees, in which each individual decision tree is learned in a standard way but with the exception that only a small random subset of the attributes is available for learning. the votes of the individual trees in the forest are then averaged to classify new examples.

Boosted Decision Tree

Boosted decision trees refer to a method in which each decision tree is learned on the entire dataset, but after the first tree is learned, the weights of the instances in the data are adjusted so that incorrectly classified example are given a higher weight. the decision tree algorithm is then run again, but this time with a focus on the “harder” examples. this process is repeated many times. the resulting collection of decision trees classifies new examples in a similar manner to the Random Forests, by averaging the individual predictions, except that the individual trees are weighted.

In 1879, A. A. Michelson made 100 determinations of the velocity of light in air using a modification of a method proposed by the French physicist Foucault. These measurements were grouped into five trials of 20 measurements each. The numbers are in km/sec, and have had 299,000 subtracted from them. The currently accepted “true” velocity of light in vacuum is 299,792.5 km/sec.

Reference: Stigler, Stephen M. (1977). Do Robust Estimators Work with Real Data? The Annals of Statistics 5:4, 1075.