Accelerating Learning to Rank via SVM with OpenCL and OpenMP on Heterogeneous Platforms.
Huming ZhuZheng LuoYanfei WuPei LiPeng ZhangShuiping GouLicheng JiaoPublished in: ICPADS (2016)
Keyphrases
- heterogeneous platforms
- shared memory
- learning to rank
- ranking svm
- message passing
- ranking functions
- support vector
- loss function
- parallel algorithm
- information retrieval
- parallel computing
- parallel programming
- information retrieval and machine learning
- learning to rank algorithms
- evaluation measures
- document retrieval
- multi class
- support vector machine
- direct optimization
- parallel execution
- test collection
- feature selection
- retrieval systems
- data sets
- query dependent
- knn
- binary classification
- ordinal regression
- cost sensitive
- collaborative filtering
- feature space
- classification accuracy
- user behavior