Download

You find on this page the data for the competition.

For each problem, you can download the train and test set available during the competition. The idea was to use the first set (or both if you thought that helped) to train your algorithm and learn a model. Participants could then assign probabilities to each element of the test set and submit their answers on the compete page. More details are given on the participate page.

The solution file (which contains the probabilities in the target of the elements of the test set) and the model file (with a description of the target machine) are now available.

Competition data

Artificial data

During this phase, the machine used to generate each data set WAS not given. In addition, different values had been given to several parameters while they were fixed in the previous phase of the competition: the size of the samples and of the alphabet varies, the number of states of the target is not the same, the symbol and transition sparsities take different values. More details are given in the paper containing the result of the competition. All these files can be found in this archive.

Problem number	Files		Problem number	Files		Problem number	Files
Problem 1	Train set Test set Solution file Model file		Problem 2	Train set Test set Solution file Model file		Problem 3	Train set Test set Solution file Model file
Problem 4	Train set Test set Solution file Model file		Problem 5	Train set Test set Solution file Model file		Problem 6	Train set Test set Solution file Model file
Problem 7	Train set Test set Solution file Model file		Problem 8	Train set Test set Solution file Model file		Problem 9	Train set Test set Solution file Model file
Problem 10	Train set Test set Solution file Model file		Problem 11	Train set Test set Solution file Model file		Problem 12	Train set Test set Solution file Model file
Problem 13	Train set Test set Solution file Model file		Problem 14	Train set Test set Solution file Model file		Problem 15	Train set Test set Solution file Model file
Problem 16	Train set Test set Solution file Model file		Problem 17	Train set Test set Solution file Model file		Problem 18	Train set Test set Solution file Model file
Problem 19	Train set Test set Solution file Model file		Problem 20	Train set Test set Solution file Model file		Problem 21	Train set Test set Solution file Model file
Problem 22	Train set Test set Solution file Model file		Problem 23	Train set Test set Solution file Model file		Problem 24	Train set Test set Solution file Model file
Problem 25	Train set Test set Solution file Model file		Problem 26	Train set Test set Solution file Model file		Problem 27	Train set Test set Solution file Model file
Problem 28	Train set Test set Solution file Model file		Problem 29	Train set Test set Solution file Model file		Problem 30	Train set Test set Solution file Model file
Problem 31	Train set Test set Solution file Model file		Problem 32	Train set Test set Solution file Model file		Problem 33	Train set Test set Solution file Model file
Problem 34	Train set Test set Solution file Model file		Problem 35	Train set Test set Solution file Model file		Problem 36	Train set Test set Solution file Model file
Problem 37	Train set Test set Solution file Model file		Problem 38	Train set Test set Solution file Model file		Problem 39	Train set Test set Solution file Model file
Problem 40	Train set Test set Solution file Model file		Problem 41	Train set Test set Solution file Model file		Problem 42	Train set Test set Solution file Model file
Problem 43	Train set Test set Solution file Model file		Problem 44	Train set Test set Solution file Model file		Problem 45	Train set Test set Solution file Model file
Problem 46	Train set Test set Solution file Model file		Problem 47	Train set Test set Solution file Model file		Problem 48	Train set Test set Solution file Model file

Real data

The first real data problem corresponds to part of speech tagging. The train and test sets are randomly selected sentences where words have been automatically replaced by POS. The evaluation is done by comparing the submitted probabilities with the ones obtained with the 3-gram baseline trained on the whole corpus (which is 10 times bigger than the available train set).

The second real-data problem comes from sensor information. Note that all strings are of length 20 as they correspond to sliding windows over a discretized sensor signal. It may thus worth to include a stop after 20 steps in your models instead of final probabilities... The evaluation is done in the same way than for the first real data problem.

Train set for real problem 1	Test set for real problem 1
Train set for real problem 2	Test set for real problem 2

Train data

These are the data that were used during the training phase of the conference which ended on May 20th 2012. The files containing the real probabilities of the test sets are now available. They correspond to the probabilities that are attributed to the element of the test sets by the target machine.

Train Set	Test set	Solution		Train Set	Test set	Solution		Train Set	Test set	Solution
Train set 1	Test set 1	Solution 1		Train set 2	Test set 2	Solution 2		Train set 3	Test set 3	Solution 3
Train set 4	Test set 4	Solution 4		Train set 5	Test set 5	Solution 5		Train set 6	Test set 6	Solution 6
Train set 7	Test set 7	Solution 7		Train set 8	Test set 8	Solution 8		Train set 9	Test set 9	Solution 9
Train set 10	Test set 10	Solution 10		Train set 11	Test set 11	Solution 11		Train set 12	Test set 12	Solution 12
Train set 13	Test set 13	Solution 13		Train set 14	Test set 14	Solution 14		Train set 15	Test set 15	Solution 15
Train set 16	Test set 16	Solution 16		Train set 17	Test set 17	Solution 17		Train set 18	Test set 18	Solution 18
Train set 19	Test set 19	Solution 19		Train set 20	Test set 20	Solution 20		Train set 21	Test set 21	Solution 21
Train set 22	Test set 22	Solution 22		Train set 23	Test set 23	Solution 23		Train set 24	Test set 24	Solution 24
Train set 25	Test set 25	Solution 25		Train set 26	Test set 26	Solution 26		Train set 27	Test set 27	Solution 27
Train set 28	Test set 28	Solution 28		Train set 29	Test set 29	Solution 29		Train set 30	Test set 30	Solution 30
Train set 31	Test set 31	Solution 31		Train set 32	Test set 32	Solution 32		Train set 33	Test set 33	Solution 34
Train set 34	Test set 34	Solution 35		Train set 35	Test set 35	Solution 35		Train set 36	Test set 36	Solution 36
Train set 37	Test set 37	Solution 37		Train set 38	Test set 38	Solution 38		Train set 39	Test set 39	Solution 39
Train set 40	Test set 40	Solution 40		Train set 41	Test set 41	Solution 41		Train set 42	Test set 42	Solution 42
Train set 43	Test set 43	Solution 43		Train set 44	Test set 44	Solution 44		Train set 45	Test set 45	Solution 45
Train set 46	Test set 46	Solution 46		Train set 47	Test set 47	Solution 47		Train set 48	Test set 48	Solution 48
Train set 49	Test set 49	Solution 49		Train set 50	Test set 50	Solution 50		Train set 51	Test set 51	Solution 51

Hint: the name of each file (except for the first problem) contains an information about the way it had been generated.