Author: Kshirsagar, Meghana; Carbonell, Jaime; Klein-Seetharaman, Judith
Title: Multitask learning for host–pathogen protein interactions Document date: 2013_7_1
ID: sdgt2ms5_74
Snippet: To get an understanding of inter-task model similarity, we compared the parameter vectors 'w' of all tasks with each other (each w was learned on the entire training data). Because the number of features is large, we computed the cosine similarity between them. Note that we only use features that are common across tasks for this comparison. Gene expression features for instance were not used as they vary with regard to the number of expression ti.....
Document: To get an understanding of inter-task model similarity, we compared the parameter vectors 'w' of all tasks with each other (each w was learned on the entire training data). Because the number of features is large, we computed the cosine similarity between them. Note that we only use features that are common across tasks for this comparison. Gene expression features for instance were not used as they vary with regard to the number of expression time points, the experiment protocol, etc.
Search related documents:
Co phrase search for related documents- expression time and parameter vector: 1
- expression time point and gene expression: 1, 2, 3, 4, 5, 6
- feature number and gene expression: 1
- gene expression and model similarity: 1, 2
- gene expression and parameter vector: 1
Co phrase search for related documents, hyperlinks ordered by date