Global analytic solution of fully-observed variational Bayesian matrix factorization

By Shinichi Nakajima, Masashi Sugiyama, S. D. Babacan, Ryota Tomioka. 2013

158 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.

1. Markus Nagel
2. Tijmen Blankevoort
3. Andrey Kuzmin
4. Jaemin Yoo
5. Saurabh Pitre
6. Eunhyeok Park
7. Minyong Cho
8. Sandeep Pendyam

Citations in the corpus

• Compression of Deep Convolutional Neural Networks for Fast and Low Power Mobile Applications (2015, methodology): “• Each step of our scheme can be easily implemented using publicly available tools, (Naka-jima, 2015) for VBMF, (Bader et al., 2015) for Tucker decomposition, and Caffe for fine-tuning.”
• Knowledge Extraction with No Observable Data (2019, methodology): “We initialize the student networks by compressing the weight tensors using Tucker-2 decomposition [13] with VBMF [24]; we compress only the convolutional layers except the dense layers as [13].”
• Taxonomy and Evaluation of Structured Compression of Convolutional Neural Networks (2019, methodology): “The original work by Kim et al. (2015) suggests using variational Bayesian matrix factorization (Nakajima et al. (2013)) for rank selection.”