Explore >> Select a destination
|
You are here 
|
ilya.grigorik.com | ||
| | | | |
jimmytechsf.com
|
|
| | | | | Do you worry about family members with whom you're otherwise happy to share your iPhone passcode reading your private diary in a journaling app? Or maybe the existence of a particular app icon could be embarrassing. Or perhaps you want to keep your child out of apps where they could cause mischief. A new feature... | |
| | | | |
www.drorpoleg.com
|
|
| | | | | Energy Is Too Valuable to Waste on Tasks Humans Can Do. | |
| | | | |
www.scottsantens.com
|
|
| | | | | There's something about unpaid work that I've never actually seen discussed, and that's the cost of the work that's paid... Take for example the amount of unpaid care work in the US that's estimated as being around $700 billion per year [http://www.washington.edu/news/2015/06/16/study-reveals-surprising-truths-about-caregivers/ | |
| | | | |
www.paepper.com
|
|
| | | Today's paper: Rethinking 'Batch' in BatchNorm by Wu & Johnson BatchNorm is a critical building block in modern convolutional neural networks. Its unique property of operating on "batches" instead of individual samples introduces significantly different behaviors from most other operations in deep learning. As a result, it leads to many hidden caveats that can negatively impact model's performance in subtle ways. This is a citation from the paper's abstract and the emphasis is mine which caught my attention. Let's explore these subtle ways which can negatively impact your model's performance! The paper of Wu & Johnson can be found on arxiv. | ||
