Join us this Saturday at ICML room 309/online for a stellar lineup of speakers covering a wide range of AI for science topics, along with 52 amazing posters and 4 contributed talks!
Schedule: https://t.co/BMODO5U8Ur
"I'm declaring an interface because one day I may have multiple implementations"
Sure. So when that day comes, do an Extract Interface refactoring. Most IDEs do it for you in a single click. Until then, drop the interface.
SLAC scientists have completed work on a low-cost emergency ventilator that could help save lives of patients suffering from COVID-19. Licenses of the technology are now being made available royalty-free through @Stanford. https://t.co/WuFJaeAS9g
Since 2012, the amount of compute for training to AlexNet-level performance on ImageNet has been decreasing exponentially — halving every 16 months, in total a 44x improvement.
By contrast, Moore's Law would only have yielded an 11x cost improvement: https://t.co/nQnZ4uPXQf
Today's covid19 update
Notable: slight dip in the TED spread (good!), & Trump's approval (still high, for him). Reduced daily positive & fatality count still, but I remain concerned it's a reporting anomaly.
@qrsupport Please reply my DM people need to know their option for changing or refunding flights when countries are closing boarder with little time to react. You need to be quicker in your response in this crunch time.
@qrsupport Your link(https://t.co/UM5Z96mMLx) for entry restriction to destination is not updated with current restrictions. I don't know for every country but its true for India. Your link says the restriction does not apply to OCI card holders which is not true. https://t.co/Joy5rUvab7
A surprising deep learning mystery:
Contrary to conventional wisdom, performance of unregularized CNNs, ResNets, and transformers is non-monotonic: improves, then gets worse, then improves again with increasing model size, data size, or training time.
https://t.co/3PwNElyNC8