Alex Fang

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2020-93

May 29, 2020

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-93.pdf

Inspired by the success of deep learning, we investigate the connections between neural networks and simple building blocks in kernel space in order to create a kernel that achieves performance closer to state of the art on CIFAR10. Additionally, we evaluate human accuracy on ImageNet with a multi-label accuracy metric in order to better understand the robustness of machine models. By looking at both, we can better understand different failure modes of various machine learning models to improve performance in the field.

Advisors: Jonathan Ragan-Kelley

---

BibTeX citation:

@mastersthesis{Fang:EECS-2020-93,
    Author= {Fang, Alex},
    Title= {Neural Kernels and ImageNet Multi-Label Accuracy},
    School= {EECS Department, University of California, Berkeley},
    Year= {2020},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-93.html},
    Number= {UCB/EECS-2020-93},
    Abstract= {Inspired by the success of deep learning, we investigate the connections between neural networks and simple building blocks in kernel space in order to create a kernel that achieves performance closer to state of the art on CIFAR10. Additionally, we evaluate human accuracy on ImageNet with a multi-label accuracy metric in order to better understand the robustness of machine models. By looking at both, we can better understand different failure modes of various machine learning models to improve performance in the field.},
}

EndNote citation:

%0 Thesis
%A Fang, Alex 
%T Neural Kernels and ImageNet Multi-Label Accuracy
%I EECS Department, University of California, Berkeley
%D 2020
%8 May 29
%@ UCB/EECS-2020-93
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-93.html
%F Fang:EECS-2020-93