Optimized Fused Floating-Point Many-Term Dot-Product Hardware for Machine Learning Accelerators.
Himanshu KaulMark A. AndersSanu MathewSeongjong KimRam KrishnamurthyPublished in: ARITH (2019)
Keyphrases
- floating point
- dot product
- machine learning
- graphics processing units
- floating point arithmetic
- instruction set
- computing systems
- fixed point
- floating point unit
- kernel function
- similarity function
- low cost
- field programmable gate array
- gaussian kernels
- machine learning algorithms
- feature space
- learning algorithm
- hardware and software
- support vector machine
- decision trees
- hardware implementation
- kernel methods
- embedded systems
- semi supervised learning
- text classification
- learning problems
- processing units
- supervised learning
- feature selection