Scribd
Upload
28 views2 pages

Machine Learning Canvas (v1.1)

Uploaded by

FernandoCardoso
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
28 views2 pages

Machine Learning Canvas (v1.1)

Uploaded by

FernandoCardoso
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

THE MACHINE LEARNING CANVAS Designed for: Designed by:

Date: Iteration: .

PREDICTION TASK DECISIONS VALUE DATA DATA SOURCES


PROPOSITION COLLECTION
Type of task? Entity on How are predictions Where can we get (raw)
which predictions are made? turned into proposed value for Who is the end-user? What are Strategy for initial train set & information on entities and
Possible outcomes? Wait time the end-user? Mention their objectives? How will they continuous update. Mention observed outcomes? Mention
before observation? parameters of the process / benefit from the ML system? collection rate, holdout on database tables, API methods,
application that does that. Mention workflow/interfaces. production entities, websites to scrape, etc.
cost/constraints to observe
outcomes.

IMPACT MAKING BUILDING FEATURES


SIMULATION PREDICTIONS MODELS
Input representations
Can models be deployed? Which When do we make real-time / How many prod models are available at prediction
test data to assess performance? batch pred.? Time available for needed? When would we time, extracted from raw data
Cost/gain values for (in)correct this + featurization + post- update? Time available for this sources.
decisions? Fairness constraint? processing? Compute target? (including featurization and
analysis)?

MONITORING
Metrics to quantify value
creation and measure the ML
system’s impact in production
(on end-users and business)?

Version 1.1. Created by Louis Dorard, Ph.D. Licensed under a Creative Commons Attribution-ShareAlike 4.0
International License. OWNML.CO
Please keep this mention and the link to ownml.co when sharing.
Ready for the next step?
Check out the ML Project Checklist!

Lead ML implementation with confidence with the CRISP—


OWNML methodology (Cross-Industry Standard Process to
create your own Machine Learning system) and its
checklist. End-to-end ML projects are broken down into 9
phases of 4-5 tasks each.

The checklist serves as a roadmap, listing in detail what


you need to do, and in which order, so you can minimize
risks and make the most efficient use of your (and
your team’s) time.

Learn more at ownml.co/checklist

You might also like