OptiML is an optimization process for model selection and parametrization that automatically finds the best supervised model to help you solve classification and regression problems. OptiML is available from the BigML Dashboard, API, and WhizzML. This new resource creates and evaluates hundreds of supervised models (decision trees, ensembles, logistic regression, and deepnets) with multiple configurations to finally return a list of the best models for your data. OptiML helps to avoid the difficult and time-consuming work of hand-tuning multiple supervised algorithms until you find the optimal one that solves your specific problem.
2. BigML, Inc 2OptiML Release Webinar
OptiML Release
CHARLES PARKER, PH.D. - VP of Machine
Learning Algorithms
Please enter questions into chat box – We will
answer some via chat and others at the end of the
session
https://bigml.com/releases
ATAKAN CETINSOY - VP of Predictive Applications
Resources
Moderator
Speaker
Contact support@bigml.com
Twitter @bigmlcom
Questions
3. BigML, Inc 3OptiML Release Webinar
Parameter Optimization
• There are lots of algorithms and lots of parameters
• We don’t have time to try even close to everything
• If only we had a way to make a prediction . . .
Did I hear someone say
Machine Learning?
4. BigML, Inc 4OptiML Release Webinar
The Allure of ML
“Why don’t we just use
Machine Learning to predict
the quality of a set of
modeling parameters before
we train a model on them?”
— Every first year ML grad student ever
5. BigML, Inc 5OptiML Release Webinar
In This Webinar
• Technology Overview
• Metric Selection
• The Dangers of Naive Cross-validation
• Selecting the “Best” Model
• Caveat Emptor!
6. BigML, Inc 6OptiML Release Webinar
In This Webinar
• Technology Overview
• Metric Selection
• The Dangers of Naive Cross-validation
• Selecting the “Best” Model
• Caveat Emptor!
7. BigML, Inc 7OptiML Release Webinar
Bayesian Parameter Optimization
• The performance of a ML algorithm (with associated parameters) is
data dependent
• So: Learn from your previous attempts
• Train a model, then evaluate it
• After you’ve done a number of evaluations, learn a regression
model to predict the performance of future, as-yet-untrained
models
• Use this classifier to chose a promising set of “next models” to
evaluate
15. BigML, Inc 15OptiML Release Webinar
Some Other Tricks
• Use metalearning to select a good set of initial candidates
• Cross-validation is expensive, and there’s no reason to do it for models
with terrible performance; stop early in these cases
16. BigML, Inc 16OptiML Release Webinar
In This Webinar
• Technology Overview
• Metric Selection
• The Dangers of Naive Cross-validation
• Selecting the “Best” Model
• Caveat Emptor!
17. BigML, Inc 17OptiML Release Webinar
A Metric Selection Flowchart
YES
YES
YES
NO
NO
NO
YES
NO
Will you bother about
threshold setting?
Is yours a “ranking”
problem?
Is your dataset
imbalanced?
Do you care more
about the top-ranked
instances?
Max. Phi
KS-statistic
Area Under the ROC / PR curve
Kendall’s Tau
Spearman’s Rho
Accuracy
Phi coefficient
f-measure
18. BigML, Inc 18OptiML Release Webinar
Ranking Problems
Medical Diagnosis (no) vs. Stock Picking (yes)
20. BigML, Inc 20OptiML Release Webinar
In This Webinar
• Technology Overview
• Metric Selection
• The Dangers of Naive Cross-validation
• Selecting the “Best” Model
• Caveat Emptor!
21. BigML, Inc 21OptiML Release Webinar
Is Cross-Validation Right for You?
• Cross-validation is a good tool some
of the time
• Many other times, it is disastrously bad
• Overly optimistic
• False confidence in results
• This is why we offer the option for a
specific holdout set
22. BigML, Inc 22OptiML Release Webinar
Case #1: Market Direction
• Suppose you want to predict the direction of the stock market
• You have information for that market for each minute of each day
• But minutes next to each other are correlated in the input and objective field
• So if you have the answer for one minute, you can trivially predict the rest!
• Cross-validation will tell you your classifier is near-perfect!
All Negative
All Positive
Close of Day
23. BigML, Inc 23OptiML Release Webinar
Case #2: Photo Age Prediction
• Suppose you want to predict the age of a printed photograph (based on dye-
fade, paper watermarks, the presence and type of border, etc.)
• Your training set: A few thousand photos from a few dozen people
• But the age of one person’s photos are correlated in both the input and output
spaces! (same age, camera, storage conditions, etc.)
• So you can trivially do well predicting the age of some of one person’s photos if
you know the ages of the rest
• Cross-validation will tell you your classifier is near perfect!
24. BigML, Inc 24OptiML Release Webinar
Take Care!
• These situations are very common in all
cases where data comes in batches
(days, users, etc.)
• The solution is to hold out whole batches
of data (e.g., a specific test set) rather
than just random points from each one
(as in cross-validation)
• It’s possible that it isn’t a problem in your
dataset, but when in doubt, try both!
25. BigML, Inc 25OptiML Release Webinar
In This Webinar
• Technology Overview
• Metric Selection
• The Dangers of Naive Cross-validation
• Selecting the “Best” Model
• Caveat Emptor!
26. BigML, Inc 26OptiML Release Webinar
Which Model is Best?
• Performance isn’t the only issue!
• Retraining: Will the amount of data you have be different in the future?
• Fit stability: How confident must you be that the model’s behavior is invariant
to small data changes?
• Prediction speed: The difference can be orders of magnitude
27. BigML, Inc 27OptiML Release Webinar
Modeling Tradeoffs
Interpretability vs. Representability
Weak vs. Slow
Confidence vs. Performance
Biased vs. Data-hungry
Simple
(Logistic)
Complex
(Deepnets)
28. BigML, Inc 28OptiML Release Webinar
In This Webinar
• Technology Overview
• Metric Selection
• The Dangers of Naive Cross-validation
• Selecting the “Best” Model
• Caveat Emptor!
29. BigML, Inc 29OptiML Release Webinar
Mo’ Problems
• Model selection tends to take a lot of
data, and the more accurate you
want the search to be, the more data
you need.
• We had to define a search space that
would suit “most” datasets. It’s
possible that the right model for your
data isn’t in there!
30. BigML, Inc 30OptiML Release Webinar
https://bigml.com/releases/winter-2018
Learn More
https://bigml.com/whatsnew