High Performance Parallel Stochastic Gradient Descent in Shared Memory.
Scott SallinenNadathur SatishMikhail SmelyanskiySamantika S. SuryChristopher RéPublished in: IPDPS (2016)
Keyphrases
- shared memory
- distributed memory
- stochastic gradient descent
- parallel computers
- message passing
- parallel algorithm
- parallel computing
- multi processor
- loss function
- parallel programming
- least squares
- matrix factorization
- step size
- parallel architecture
- random forests
- parallel machines
- parallel execution
- support vector machine
- online algorithms
- parallel implementation
- parallel architectures
- belief propagation
- weight vector
- cost function
- shared memory multiprocessor
- similarity measure
- shared memory multiprocessors
- graphical models
- convergence speed
- random forest
- linear combination
- online learning
- higher order
- upper bound
- multi core systems
- evolutionary algorithm