Adam Obeng
@Adam_Obeng
Followers
324
Following
13
Media
46
Statuses
1K
Computational Social Scientist. Senior Staff Data Scientist @ Pinterest. I speak on behalf of no-one, no-one speaks for me. Post-irony, post-human, post-posting
NYC
Joined October 2008
.@terryandrob To Sir Pterry! #BananaDaiquiri #GreatATuin #GNUTerryPratchett #fb
http://t.co/5UOUyx2lPa.
0
0
10
RT @jaschasd: Have you ever done a dense grid search over neural network hyperparameters? Like a *really dense* grid search? It looks like….
0
2K
0
I think this is a discovery of comparable importance to the arrow in the FedEx logo.
0
0
1
I wrote a Javascript Autopilot for @SpaceX #CrewDragon @AstroBehnken, @Astro_Doug, in case y'all need to take a break.
1
1
1
Python's implementation of binary logic including None is. surprising
0
0
3
It's a 1 hidden-layer MLP with sigmoid activations, fit with SGD. Usage: set X and Y to numpy arrays of training feature/labels, resulting weights will be in variables w and e.
1
1
1
from numpy import *.r=random.c=hstack.p=dot.N=r.randn.s=lambda x:1/(1+exp(-x)).a=lambda x:s(x)*(1-s(x)).h,j=X.shape.w=N(5,j+1).e=N(1,6).for i in r.randint(0,h,1000*h):o=c((1,X[i]));m=p(w,o);n=c((1,s(m)));d=p(e,n);b=0.02*(Y[i]-s(d));g=b;w+=outer(e[:,1:]*a(d)*a(m),o)*b;e+=n*g*a(d).
1
2
3
RT @seanjtaylor: ICYMI: We awarded three $50K grants for some really exciting applied statistics research proposals.
0
3
0
Some people, when faced with a problem, think "Some people, when faced with a problem, think "Some people, when faced with a problem, think….
0
0
2
Verifying myself: I am adamobeng on Keybase.io. HHW5selJ4VaJ8iW6_Wk3e8h3vPjl1ytL6_g8 /
0
0
0
0
0
3
0
2
0
RT @john_chandler: I've reached that age when I'm shaking my fist at a #PyConUK lightning talk and saying "use ASCII emoticons you pesky ki….
0
2
0