We launched an African NLP dataset challenge over at @AI4Dev.
Why?
Because the African NLP ecosystem is 🔥🔥and we want to see what happens when we add some fuel to the flame.
It's hosted on @ZindiAfrica and there's USD 1000 up for grabs each month for the next 5 months.
Millennials, quit whining. I paid off $150,000 in student loans and own a $400,000 home, because I SAVE. It’s not that hard. I
-Make coffee at home
-Bus instead of Uber
-Shop sales
-Had parents pay off my loans & buy me a house because I’m daddy’s special boy
-Got Hulu with ads
These cups are made of linear mirrors with 24 sides: when put on the saucer the anamorphic design reveals a hidden image. Bonus: some of them also use a moiré effect, to give an even more impressive, animated result https://t.co/QCf4Bo7Zqj
Did you know that for 200 years we've known that spiders can fly but haven't really understood how they did that and we finally worked it out and the answer is "they ride the earth's magnetic fields at will"??? Nature is fucking awesome. https://t.co/sZH7RJfKv7
Remembering the time in 2005 that @yahoo exec told me that @reddit was a "rounding error" compared to them and asked the room why this meeting was even happening. 😆
Fight on, founders. No one will believe in you until you believe in yourself.
Currently training a Q&A model, and it is producing crazy impressive results!
Q: How do you find a good title?
A: See attached
None of the samples can be found in the training set that I used.
😱😱😱😱😱
Picked up a hitchhiker last night. He said thanks how do you know I’m not a serial killer? I replied the chances of two serial killer’s being in the same car are astronomical.
What you're seeing is a myosin protein (or possibly kinesin) dragging an endorphin along a filament to the inner part of the brain's parietal cortex which creates happiness. Happiness.
You're looking at happiness.
❤❤❤
Computer vision model training uses different augmentation of input images (random crops, distortions, etc.). Instead of hand-engineering the operation sequence, one can instead use meta learning to learn effective sequences. Gets new state-of-the-art of 83.54% top1 on ImageNet! https://t.co/NRWfrdARDx
From a paper I'm reading: "we have also created a docker container that reproduces all the experiments and figures in this paper". Very nice! A high bar, but I think eventually this should be the norm in machine learning conferences too.