This article reminds me a lot of playing around the the tensor flow playground, when I first start learning about neural networks. Although it could use a little more explanatory umphf.
I am really curious and excited to see where this serious goes. Especially with a line like "observing them in the wild." I love to understand how ambiguous system work and make that information accessible, to help build trust and understanding
I hope you will like the next posts. I plan to describe existing visualization techniques for three different cases: model input-output behavior, explanations and extrapolations, and model components/internals. Thanks for reading and commenting. Let me know how it goes!
It’s one of the biggest issues. I don’t see a lot of adoption. We need to study more real-world problems people have in this space. A lot of the existing work targets very small and specialized niches of people. My hope is that my work here is going to guide people in selecting relevant problems to work on. In a way I am doing it for myself too! The effort necessary to write the whole series is going to force me to clarify a lot of issues. P.s. I love nursing! My wife is a nurse too.
This article reminds me a lot of playing around the the tensor flow playground, when I first start learning about neural networks. Although it could use a little more explanatory umphf.
I am really curious and excited to see where this serious goes. Especially with a line like "observing them in the wild." I love to understand how ambiguous system work and make that information accessible, to help build trust and understanding
I hope you will like the next posts. I plan to describe existing visualization techniques for three different cases: model input-output behavior, explanations and extrapolations, and model components/internals. Thanks for reading and commenting. Let me know how it goes!
I come from a nursing background and intrigued by bench to bedside translation. Do you see a gap between ML viz research translation to the "desktop"
It’s one of the biggest issues. I don’t see a lot of adoption. We need to study more real-world problems people have in this space. A lot of the existing work targets very small and specialized niches of people. My hope is that my work here is going to guide people in selecting relevant problems to work on. In a way I am doing it for myself too! The effort necessary to write the whole series is going to force me to clarify a lot of issues. P.s. I love nursing! My wife is a nurse too.