Adaptive Quantization of Neural Networks

By Soroosh Khoram, J. Li. 2018

59 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Similarity-Preserving Knowledge Distillation (2019, background): “For example, unstructured pruning requires optimized sparse matrix multiplication routines to realize practical acceleration [26], platform constraint-aware compression [2, 36, 37] requires hardware simulators or empirical measurements, and arbitrarybit quantization [9, 17] requires specialized hard”
Layer-compensated Pruning for Resource-constrained Convolutional Neural Networks (2018, background): “While there is emerging work toward network quantization (Khoram & Li, 2018) in a resource-constrained setting, structural pruning is often adopted as the solution for the problem (Gordon et al., 2018; Luo et al., 2017; He et al., 2017; Molchanov et al., 2017; He et al., 2018b; Yang et al., 2018a;…”
A Survey on Methods and Theories of Quantized Neural Networks (2018, methodology): “Later in the work of [Soroosh Khoram, 2018] the authors proposed an adaptive quantization method which incorporates the loss function into an optimization problem to consider the importance of different connections.”
HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks (2019, background): “Quantization [1, 6, 15, 36, 34, 17, 35, 8, 32, 33, 29, 9, 10, 5, 7] is a very promising approach to address this problem by reducing the memory bottleneck, thus allowing the use of lower precision computational units in hardware.”