Contenu connexe Similaire à News From Mahout (20) Plus de MapR Technologies (20) News From Mahout2. 2©MapR Technologies - Confidential
whoami – Ted Dunning
Chief Application Architect, MapR Technologies
Committer, member, Apache Software Foundation
– particularly Mahout, Zookeeper and Drill
(we’re hiring)
Contact me at
tdunning@maprtech.com
tdunning@apache.com
ted.dunning@gmail.com
@ted_dunning
3. 3©MapR Technologies - Confidential
Slides and such (available late tonight):
– http://www.mapr.com/company/events/nyhug-03-05-2013
Hash tags: #mapr #nyhug #mahout
4. 4©MapR Technologies - Confidential
New in Mahout
0.8 is coming soon (1-2 months)
gobs of fixes
QR decomposition is 10x faster
– makes ALS 2-3 times faster
May include Bayesian Bandits
Super fast k-means
– fast
– online (!?!)
5. 5©MapR Technologies - Confidential
New in Mahout
0.8 is coming soon (1-2 months)
gobs of fixes
QR decomposition is 10x faster
– makes ALS 2-3 times faster
May include Bayesian Bandits
Super fast k-means
– fast
– online (!?!)
– fast
Possible new edition of MiA coming
– Japanese and Korean editions released, Chinese coming
6. 6©MapR Technologies - Confidential
New in Mahout
0.8 is coming soon (1-2 months)
gobs of fixes
QR decomposition is 10x faster
– makes ALS 2-3 times faster
May include Bayesian Bandits
Super fast k-means
– fast
– online (!?!)
– fast
Possible new edition of MiA coming
– Japanese and Korean editions released, Chinese coming
9. 9©MapR Technologies - Confidential
Bogus Dog Food is the Best!
Now available in handy 1 ton
bags!
Buy 5!
What
picture?
What tag-
line?
What call to
action?
10. 10©MapR Technologies - Confidential
The Challenge
Design decisions affect probability of success
– Cheesy web-sites don’t even sell cheese
The best designers do better when allowed to fail
– Exploration juices creativity
But failing is expensive
– If only because we could have succeeded
– But also because offending or disappointing customers is bad
11. 11©MapR Technologies - Confidential
More Challenges
Too many designs
– 5 pictures
– 10 tag-lines
– 4 calls to action
– 3 back-ground colors
=> 5 x 10 x 4 x 3 = 600 designs
It gets worse quickly
– What about changes on the back-end?
– Search engine variants?
– Checkout process variants?
12. 12©MapR Technologies - Confidential
Example – AB testing in real-time
I have 15 versions of my landing page
Each visitor is assigned to a version
– Which version?
A conversion or sale or whatever can happen
– How long to wait?
Some versions of the landing page are horrible
– Don’t want to give them traffic
13. 13©MapR Technologies - Confidential
A Quick Diversion
You see a coin
– What is the probability of heads?
– Could it be larger or smaller than that?
I flip the coin and while it is in the air ask again
I catch the coin and ask again
I look at the coin (and you don’t) and ask again
Why does the answer change?
– And did it ever have a single value?
14. 14©MapR Technologies - Confidential
A Philosophical Conclusion
Probability as expressed by humans is subjective and depends on
information and experience
19. 19©MapR Technologies - Confidential
Another Quick Diversion
Let’s play a shell game
This is a special shell game
It costs you nothing to play
The pea has constant probability of being under each shell
(trust me)
How do you find the best shell?
How do you find it while maximizing the number of wins?
21. 21©MapR Technologies - Confidential
Interim Thoughts
Can you identify winners or losers without trying them out?
Can you ever completely eliminate a shell with a bad streak?
Should you keep trying apparent losers?
23. 23©MapR Technologies - Confidential
Conclusions
Can you identify winners or losers without trying them out?
No
Can you ever completely eliminate a shell with a bad streak?
No
Should you keep trying apparent losers?
Yes, but at a decreasing rate
25. 25©MapR Technologies - Confidential
Bayesian Bandit
Compute distributions based on data so far
Sample p1, p2 and p2 from these distributions
Pick shell i where i = argmaxi pi
Lemma 1: The probability of picking shell i will match the
probability it is the best shell
Lemma 2: This is as good as it gets
26. 26©MapR Technologies - Confidential
And it works!
11000 100 200 300 400 500 600 700 800 900 1000
0.12
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0.11
n
regret
ε- greedy, ε = 0.05
Bayesian Bandit with Gamma- Normal
28. 28©MapR Technologies - Confidential
The Code
Select an alternative
Select and learn
But we already know how to count!
n = dim(k)[1]
p0 = rep(0, length.out=n)
for (i in 1:n) {
p0[i] = rbeta(1, k[i,2]+1, k[i,1]+1)
}
return (which(p0 == max(p0)))
for (z in 1:steps) {
i = select(k)
j = test(i)
k[i,j] = k[i,j]+1
}
return (k)
29. 29©MapR Technologies - Confidential
The Basic Idea
We can encode a distribution by sampling
Sampling allows unification of exploration and exploitation
Can be extended to more general response models
30. 30©MapR Technologies - Confidential
The Original Problem
Bogus Dog Food is the Best!
Now available in handy 1 ton
bags!
Buy 5!
x1
x2
x3
31. 31©MapR Technologies - Confidential
Response Function
p(win) = w qi
i
å xi
æ
è
ç
ö
ø
÷
6- 6 - 4 - 2 0 2 4
1
0
0.5
x
y
32. 32©MapR Technologies - Confidential
Generalized Banditry
Suppose we have an infinite number of bandits
– suppose they are each labeled by two real numbers x and y in [0,1]
– also that expected payoff is a parameterized function of x and y
– now assume a distribution for θ that we can learn online
Selection works by sampling θ, then computing f
Learning works by propagating updates back to θ
– If f is linear, this is very easy
– For special other kinds of f it isn’t too hard
Don’t just have to have two labels, could have labels and context
E z[ ] = f (x, y |q)
33. 33©MapR Technologies - Confidential
Context Variables
Bogus Dog Food is the Best!
Now available in handy 1 ton
bags!
Buy 5!
x1
x2
x3
user.geo env.time env.day_of_week env.weekend
34. 34©MapR Technologies - Confidential
Caveats
Original Bayesian Bandit only requires real-time
Generalized Bandit may require access to long history for learning
– Pseudo online learning may be easier than true online
Bandit variables can include content, time of day, day of week
Context variables can include user id, user features
Bandit × context variables provide the real power
38. 38©MapR Technologies - Confidential
What is Quality?
Robust clustering not a goal
– we don’t care if the same clustering is replicated
Generalization is critical
Agreement to “gold standard” is a non-issue
45. 45©MapR Technologies - Confidential
For Example
Grouping these
two clusters
seriously hurts
squared distance
D4
2
(X) >
1
s 2
D5
2
(X)
47. 47©MapR Technologies - Confidential
Typical k-means Failure
Selecting two seeds
here cannot be
fixed with Lloyds
Result is that these two
clusters get glued
together
48. 48©MapR Technologies - Confidential
Ball k-means
Provably better for highly clusterable data
Tries to find initial centroids in each “core” of each real clusters
Avoids outliers in centroid computation
initialize centroids randomly with distance maximizing tendency
for each of a very few iterations:
for each data point:
assign point to nearest cluster
recompute centroids using only points much closer than closest cluster
49. 49©MapR Technologies - Confidential
Still Not a Win
Ball k-means is nearly guaranteed with k = 2
Probability of successful seeding drops exponentially with k
Alternative strategy has high probability of success, but takes
O(nkd + k3d) time
50. 50©MapR Technologies - Confidential
Still Not a Win
Ball k-means is nearly guaranteed with k = 2
Probability of successful seeding drops exponentially with k
Alternative strategy has high probability of success, but takes O(
nkd + k3d ) time
But for big data, k gets large
51. 51©MapR Technologies - Confidential
Surrogate Method
Start with sloppy clustering into lots of clusters
κ = k log n clusters
Use this sketch as a weighted surrogate for the data
Results are provably good for highly clusterable data
52. 52©MapR Technologies - Confidential
Algorithm Costs
Surrogate methods
– fast, sloppy single pass clustering with κ = k log n
– fast sloppy search for nearest cluster,
O(d log κ) = O(d (log k + log log n)) per point
– fast, in-memory, high-quality clustering of κ weighted centroids
O(κ k d + k3 d) = O(k2 d log n + k3 d) for small k, high quality
O(κ d log k) or O(d log κ log k) for larger k, looser quality
– result is k high-quality centroids
• Even the sloppy surrogate may suffice
53. 53©MapR Technologies - Confidential
Algorithm Costs
Surrogate methods
– fast, sloppy single pass clustering with κ = k log n
– fast sloppy search for nearest cluster,
O(d log κ) = O(d ( log k + log log n )) per point
– fast, in-memory, high-quality clustering of κ weighted centroids
O(κ k d + k3 d) = O(k2 d log n + k3 d) for small k, high quality
O(κ d log k) or O( d log k ( log k + log log n ) ) for larger k, looser quality
– result is k high-quality centroids
• For many purposes, even the sloppy surrogate may suffice
54. 54©MapR Technologies - Confidential
Algorithm Costs
How much faster for the sketch phase?
– take k = 2000, d = 10, n = 100,000
– k d log n = 2000 x 10 x 26 = 500,000
– d (log k + log log n) = 10(11 + 5) = 170
– 3,000 times faster is a bona fide big deal
55. 55©MapR Technologies - Confidential
Algorithm Costs
How much faster for the sketch phase?
– take k = 2000, d = 10, n = 100,000
– k d log n = 2000 x 10 x 26 = 500,000
– d (log k + log log n) = 10(11 + 5) = 170
– 3,000 times faster is a bona fide big deal
56. 56©MapR Technologies - Confidential
How It Works
For each point
– Find approximately nearest centroid (distance = d)
– If (d > threshold) new centroid
– Else if (u > d/threshold) new cluster
– Else add to nearest centroid
If centroids > κ ≈ C log N
– Recursively cluster centroids with higher threshold
58. 58©MapR Technologies - Confidential
But Wait, …
Finding nearest centroid is inner loop
This could take O( d κ ) per point and κ can be big
Happily, approximate nearest centroid works fine
60. 60©MapR Technologies - Confidential
LSH Bit-match Versus Cosine
0 8 16 24 32 40 48 56 64
1
- 1
- 0.8
- 0.6
- 0.4
- 0.2
0
0.2
0.4
0.6
0.8
X Axis
YAxis
62. 62©MapR Technologies - Confidential
Parallel Speedup?
1 2 3 4 5 20
10
100
20
30
40
50
200
Threads
Timeperpoint(μs)
2
3
4
5
6
8
10
12
14
16
Threaded version
Non- threaded
Perfect Scaling
✓
63. 63©MapR Technologies - Confidential
Quality
Ball k-means implementation appears significantly better than
simple k-means
Streaming k-means + ball k-means appears to be about as good as
ball k-means alone
All evaluations on 20 newsgroups with held-out data
Figure of merit is mean and median squared distance to nearest
cluster
64. 64©MapR Technologies - Confidential
Contact Me!
We’re hiring at MapR in US and Europe
MapR software available for research use
Get the code as part of Mahout trunk (or 0.8 very soon)
Contact me at tdunning@maprtech.com or @ted_dunning
Share news with @apachemahout
Notes de l'éditeur The basic idea here is that I have colored slides to be presented by you in blue. You should substitute and reword those slides as you like. In a few places, I imagined that we would have fast back and forth as in the introduction or final slide where we can each say we are hiring in turn.The overall thrust of the presentation is for you to make these points:Amex does lots of modelingit is expensivehaving a way to quickly test models and new variables would be awesomeso we worked on a new project with MapRMy part will say the following:Knn basic pictorial motivation (could move to you if you like)describe knn quality metric of overlapshow how bad metric breaks knn (optional)quick description of LSH and projection searchpicture of why k-means search is coolmotivate k-means speed as tool for k-means searchdescribe single pass k-means algorithmdescribe basic data structuresshow parallel speedupOur summary should state that we have achievedsuper-fast k-means clusteringinitial version of super-fast knn search with good overlap