SlideShare una empresa de Scribd logo
1 de 18
Descargar para leer sin conexión
Warm Start Tuning with Prior Beliefs
Thursday, June 4, 2020
Michael McCourt — Head of Research, mccourt@sigopt.com
SigOpt. Confidential.
Abstract
SigOpt provides an extensive set of advanced features,
which help you, the expert, save time while
increasing model performance via experimentation.
Today, we will continue this talk series by discussing
how to incorporate your prior beliefs into a SigOpt
experiment to inform the optimization about your expert
knowledge.
SigOpt. Confidential.
Agenda
1. Overview of SigOpt
2. How to convey prior beliefs about parameter
performance to SigOpt
3. Discussion on benefits and risks of prior beliefs
SigOpt. Confidential.
Accelerate and amplify the
impact of modelers everywhere
SigOpt. Confidential.
Can be deployed on-premises
Training
Data
AI, ML, DL,
Simulation Model
Model Evaluation
or Backtest
Testing
Data
New
Configurations
Objective
Metric
Better
Results
EXPERIMENT INSIGHTS
Track, organize, analyze and
reproduce any model
ENTERPRISE PLATFORM
Built to fit any stack and scale
with your needs
OPTIMIZATION ENGINE
Explore and exploit with a
variety of techniques
RESTAPI
Parameters or
Hyperparameters
Your data
and models
stay private
SigOpt delivers iterative, automated optimization
SigOpt fits
any stack
and helps
you build
the best
models
SigOpt: API-enabled parameter tuning
5
SigOpt is a Black Box Tool
Goal: SigOpt has been designed to prioritize users privacy -- no modeling information is required.
Opportunity: Customers can own their modeling process and still benefit from efficient model tuning.
Complication: What if customers want to provide modeling information?
Your models are your own
Users have Relevant Knowledge
Goal: SigOpt has been designed to prioritize users privacy -- no modeling information is required.
Opportunity: Customers can own their modeling process and still benefit from efficient model tuning.
Complication: What if customers want to provide modeling information?
Example: After many years developing XGBoost models,
a user may have developed intuition regarding the learning
rates which perform the best.
• Could this intuition accelerate SigOpt’s
optimization engine?
• How can it be conveyed to SigOpt?
Previous experience could inform the optimization
Prior Beliefs Structure
Our new Prior Beliefs feature allows customers to communicate structured information about parameters.
• This information takes the form of a probability density function.
• Parameters with higher prior belief value will receive more interest from the optimizer.
• Prior beliefs become less important as more observations are reported.
What does the user believe about parameters?
Prior Beliefs Structure
Our new Prior Beliefs feature allows customers to communicate structured information about parameters.
• This information takes the form of a probability density function.
• Parameters with higher prior belief value will receive more interest from the optimizer.
• Prior beliefs become less important as more observations are reported.
Example: Using high-performing
learning rates from previous
months, we can build prior beliefs.
What does the user believe about parameters?
Domain
Boundary
Possible
Prior
Belief
Designing Prior Beliefs for SigOpt
At present, SigOpt allows two types of prior beliefs.
• Normal
• mean
• scale
• Beta
• shape_a
• shape_b
How are prior beliefs structured?
Designing Prior Beliefs for SigOpt
At present, SigOpt allows two types of prior beliefs.
• Normal
• mean
• scale
• Beta
• shape_a
• shape_b
How are prior beliefs structured?
https://app.sigopt.com/docs/overview/parameter_priors
Benefits of Prior Beliefs
How will prior beliefs affect an experiment?
This prior beliefs feature gives users
new power.
Previously: Features existed to
define or change the problem
under consideration.
Prior beliefs: The problem remains
the same, but the optimization
engine changes.
Choices made when defining prior
beliefs can affect the optimization
process in both good and bad ways.
Benefits of Prior Beliefs
This prior beliefs feature gives users
new power.
Previously: Features existed to
define or change the problem
under consideration.
Prior beliefs: The problem remains
the same, but the optimization
engine changes.
Choices made when defining prior
beliefs can affect the optimization
process in both good and bad ways.
How will prior beliefs affect an experiment?
Benefits and Dangers of Prior Beliefs
This prior beliefs feature gives users
new power.
Previously: Features existed to
define or change the problem
under consideration.
Prior beliefs: The problem remains
the same, but the optimization
engine changes.
Choices made when defining prior
beliefs can affect the optimization
process in both good and bad ways.
How will prior beliefs affect an experiment?
Benefits and Dangers of Prior Beliefs
This prior beliefs feature gives users
new power.
Previously: Features existed to
define or change the problem
under consideration.
Prior beliefs: The problem remains
the same, but the optimization
engine changes.
Choices made when defining prior
beliefs can affect the optimization
process in both good and bad ways.
How will prior beliefs affect an experiment?
Can provide early acceleration
Eventually, the
observed data will
overpower any
prior beliefs
More Content
https://app.sigopt.com/docs/overview/parameter_priors https://sigopt.com/blog/applying-prior-beliefs-in-sigopt/
Learn more about Prior Beliefs
SigOpt. Confidential.
Questions?
Ask now, or feel free to email:
contact@sigopt.com
Slides will be available on slideshare … check your
email in the next few days.
SigOpt. Confidential.
Check out our
YouTube channel: Learn more about SigOpt
Read our research and product blog at
blog.sigopt.com.
See more videos at
https://sigopt.com/resources/videos
Try our solution:
Sign up at
sigopt.com/try-it
today.
Click Here
Upcoming webinars:
● Detecting COVID-19 Cases
with Deep Learning
Tuesday, June 9 at 10am PT / 1pm ET

Más contenido relacionado

Más de SigOpt

Tuning Data Augmentation to Boost Model Performance
Tuning Data Augmentation to Boost Model PerformanceTuning Data Augmentation to Boost Model Performance
Tuning Data Augmentation to Boost Model PerformanceSigOpt
 
Advanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise WebinarAdvanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise WebinarSigOpt
 
Modeling at Scale: SigOpt at TWIMLcon 2019
Modeling at Scale: SigOpt at TWIMLcon 2019Modeling at Scale: SigOpt at TWIMLcon 2019
Modeling at Scale: SigOpt at TWIMLcon 2019SigOpt
 
Tuning 2.0: Advanced Optimization Techniques Webinar
Tuning 2.0: Advanced Optimization Techniques WebinarTuning 2.0: Advanced Optimization Techniques Webinar
Tuning 2.0: Advanced Optimization Techniques WebinarSigOpt
 
SigOpt at Ai4 Finance—Modeling at Scale
SigOpt at Ai4 Finance—Modeling at Scale SigOpt at Ai4 Finance—Modeling at Scale
SigOpt at Ai4 Finance—Modeling at Scale SigOpt
 
Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...
Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...
Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...SigOpt
 
Machine Learning Infrastructure
Machine Learning InfrastructureMachine Learning Infrastructure
Machine Learning InfrastructureSigOpt
 
SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...
SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...
SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...SigOpt
 
SigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms
SigOpt at O'Reilly - Best Practices for Scaling Modeling PlatformsSigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms
SigOpt at O'Reilly - Best Practices for Scaling Modeling PlatformsSigOpt
 
SigOpt at GTC - Tuning the Untunable
SigOpt at GTC - Tuning the UntunableSigOpt at GTC - Tuning the Untunable
SigOpt at GTC - Tuning the UntunableSigOpt
 
SigOpt at GTC - Reducing operational barriers to optimization
SigOpt at GTC - Reducing operational barriers to optimizationSigOpt at GTC - Reducing operational barriers to optimization
SigOpt at GTC - Reducing operational barriers to optimizationSigOpt
 
Lessons for an enterprise approach to modeling at scale
Lessons for an enterprise approach to modeling at scaleLessons for an enterprise approach to modeling at scale
Lessons for an enterprise approach to modeling at scaleSigOpt
 
Modeling at scale in systematic trading
Modeling at scale in systematic tradingModeling at scale in systematic trading
Modeling at scale in systematic tradingSigOpt
 
SigOpt at MLconf - Reducing Operational Barriers to Model Training
SigOpt at MLconf - Reducing Operational Barriers to Model TrainingSigOpt at MLconf - Reducing Operational Barriers to Model Training
SigOpt at MLconf - Reducing Operational Barriers to Model TrainingSigOpt
 
Machine Learning Infrastructure
Machine Learning InfrastructureMachine Learning Infrastructure
Machine Learning InfrastructureSigOpt
 
Tuning the Untunable - Insights on Deep Learning Optimization
Tuning the Untunable - Insights on Deep Learning OptimizationTuning the Untunable - Insights on Deep Learning Optimization
Tuning the Untunable - Insights on Deep Learning OptimizationSigOpt
 
Machine Learning Fundamentals
Machine Learning FundamentalsMachine Learning Fundamentals
Machine Learning FundamentalsSigOpt
 
Tips and techniques for hyperparameter optimization
Tips and techniques for hyperparameter optimizationTips and techniques for hyperparameter optimization
Tips and techniques for hyperparameter optimizationSigOpt
 
MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...
MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...
MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...SigOpt
 
Using Optimal Learning to Tune Deep Learning Pipelines
Using Optimal Learning to Tune Deep Learning PipelinesUsing Optimal Learning to Tune Deep Learning Pipelines
Using Optimal Learning to Tune Deep Learning PipelinesSigOpt
 

Más de SigOpt (20)

Tuning Data Augmentation to Boost Model Performance
Tuning Data Augmentation to Boost Model PerformanceTuning Data Augmentation to Boost Model Performance
Tuning Data Augmentation to Boost Model Performance
 
Advanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise WebinarAdvanced Optimization for the Enterprise Webinar
Advanced Optimization for the Enterprise Webinar
 
Modeling at Scale: SigOpt at TWIMLcon 2019
Modeling at Scale: SigOpt at TWIMLcon 2019Modeling at Scale: SigOpt at TWIMLcon 2019
Modeling at Scale: SigOpt at TWIMLcon 2019
 
Tuning 2.0: Advanced Optimization Techniques Webinar
Tuning 2.0: Advanced Optimization Techniques WebinarTuning 2.0: Advanced Optimization Techniques Webinar
Tuning 2.0: Advanced Optimization Techniques Webinar
 
SigOpt at Ai4 Finance—Modeling at Scale
SigOpt at Ai4 Finance—Modeling at Scale SigOpt at Ai4 Finance—Modeling at Scale
SigOpt at Ai4 Finance—Modeling at Scale
 
Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...
Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...
Interactive Tradeoffs Between Competing Offline Metrics with Bayesian Optimiz...
 
Machine Learning Infrastructure
Machine Learning InfrastructureMachine Learning Infrastructure
Machine Learning Infrastructure
 
SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...
SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...
SigOpt at Uber Science Symposium - Exploring the spectrum of black-box optimi...
 
SigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms
SigOpt at O'Reilly - Best Practices for Scaling Modeling PlatformsSigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms
SigOpt at O'Reilly - Best Practices for Scaling Modeling Platforms
 
SigOpt at GTC - Tuning the Untunable
SigOpt at GTC - Tuning the UntunableSigOpt at GTC - Tuning the Untunable
SigOpt at GTC - Tuning the Untunable
 
SigOpt at GTC - Reducing operational barriers to optimization
SigOpt at GTC - Reducing operational barriers to optimizationSigOpt at GTC - Reducing operational barriers to optimization
SigOpt at GTC - Reducing operational barriers to optimization
 
Lessons for an enterprise approach to modeling at scale
Lessons for an enterprise approach to modeling at scaleLessons for an enterprise approach to modeling at scale
Lessons for an enterprise approach to modeling at scale
 
Modeling at scale in systematic trading
Modeling at scale in systematic tradingModeling at scale in systematic trading
Modeling at scale in systematic trading
 
SigOpt at MLconf - Reducing Operational Barriers to Model Training
SigOpt at MLconf - Reducing Operational Barriers to Model TrainingSigOpt at MLconf - Reducing Operational Barriers to Model Training
SigOpt at MLconf - Reducing Operational Barriers to Model Training
 
Machine Learning Infrastructure
Machine Learning InfrastructureMachine Learning Infrastructure
Machine Learning Infrastructure
 
Tuning the Untunable - Insights on Deep Learning Optimization
Tuning the Untunable - Insights on Deep Learning OptimizationTuning the Untunable - Insights on Deep Learning Optimization
Tuning the Untunable - Insights on Deep Learning Optimization
 
Machine Learning Fundamentals
Machine Learning FundamentalsMachine Learning Fundamentals
Machine Learning Fundamentals
 
Tips and techniques for hyperparameter optimization
Tips and techniques for hyperparameter optimizationTips and techniques for hyperparameter optimization
Tips and techniques for hyperparameter optimization
 
MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...
MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...
MLconf 2017 Seattle Lunch Talk - Using Optimal Learning to tune Deep Learning...
 
Using Optimal Learning to Tune Deep Learning Pipelines
Using Optimal Learning to Tune Deep Learning PipelinesUsing Optimal Learning to Tune Deep Learning Pipelines
Using Optimal Learning to Tune Deep Learning Pipelines
 

Último

Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityPrincipled Technologies
 
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slideHistor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slidevu2urc
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherRemote DBA Services
 
08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking MenDelhi Call girls
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProduct Anonymous
 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024The Digital Insurer
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationMichael W. Hawkins
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024Rafal Los
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsJoaquim Jorge
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptxEIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptxEarley Information Science
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
GenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdfGenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdflior mazor
 
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonAnna Loughnan Colquhoun
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century educationjfdjdjcjdnsjd
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Igalia
 
Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)wesley chun
 
IAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI SolutionsIAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI SolutionsEnterprise Knowledge
 
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdfUnderstanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdfUK Journal
 

Último (20)

Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
 
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slideHistor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slide
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day Presentation
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptxEIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
EIS-Webinar-Prompt-Knowledge-Eng-2024-04-08.pptx
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
GenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdfGenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdf
 
Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
 
Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)Powerful Google developer tools for immediate impact! (2023-24 C)
Powerful Google developer tools for immediate impact! (2023-24 C)
 
IAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI SolutionsIAC 2024 - IA Fast Track to Search Focused AI Solutions
IAC 2024 - IA Fast Track to Search Focused AI Solutions
 
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdfUnderstanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
 

Warm Start Tuning and Prior Beliefs Webinar

  • 1. Warm Start Tuning with Prior Beliefs Thursday, June 4, 2020 Michael McCourt — Head of Research, mccourt@sigopt.com
  • 2. SigOpt. Confidential. Abstract SigOpt provides an extensive set of advanced features, which help you, the expert, save time while increasing model performance via experimentation. Today, we will continue this talk series by discussing how to incorporate your prior beliefs into a SigOpt experiment to inform the optimization about your expert knowledge.
  • 3. SigOpt. Confidential. Agenda 1. Overview of SigOpt 2. How to convey prior beliefs about parameter performance to SigOpt 3. Discussion on benefits and risks of prior beliefs
  • 4. SigOpt. Confidential. Accelerate and amplify the impact of modelers everywhere
  • 5. SigOpt. Confidential. Can be deployed on-premises Training Data AI, ML, DL, Simulation Model Model Evaluation or Backtest Testing Data New Configurations Objective Metric Better Results EXPERIMENT INSIGHTS Track, organize, analyze and reproduce any model ENTERPRISE PLATFORM Built to fit any stack and scale with your needs OPTIMIZATION ENGINE Explore and exploit with a variety of techniques RESTAPI Parameters or Hyperparameters Your data and models stay private SigOpt delivers iterative, automated optimization SigOpt fits any stack and helps you build the best models SigOpt: API-enabled parameter tuning 5
  • 6. SigOpt is a Black Box Tool Goal: SigOpt has been designed to prioritize users privacy -- no modeling information is required. Opportunity: Customers can own their modeling process and still benefit from efficient model tuning. Complication: What if customers want to provide modeling information? Your models are your own
  • 7. Users have Relevant Knowledge Goal: SigOpt has been designed to prioritize users privacy -- no modeling information is required. Opportunity: Customers can own their modeling process and still benefit from efficient model tuning. Complication: What if customers want to provide modeling information? Example: After many years developing XGBoost models, a user may have developed intuition regarding the learning rates which perform the best. • Could this intuition accelerate SigOpt’s optimization engine? • How can it be conveyed to SigOpt? Previous experience could inform the optimization
  • 8. Prior Beliefs Structure Our new Prior Beliefs feature allows customers to communicate structured information about parameters. • This information takes the form of a probability density function. • Parameters with higher prior belief value will receive more interest from the optimizer. • Prior beliefs become less important as more observations are reported. What does the user believe about parameters?
  • 9. Prior Beliefs Structure Our new Prior Beliefs feature allows customers to communicate structured information about parameters. • This information takes the form of a probability density function. • Parameters with higher prior belief value will receive more interest from the optimizer. • Prior beliefs become less important as more observations are reported. Example: Using high-performing learning rates from previous months, we can build prior beliefs. What does the user believe about parameters? Domain Boundary Possible Prior Belief
  • 10. Designing Prior Beliefs for SigOpt At present, SigOpt allows two types of prior beliefs. • Normal • mean • scale • Beta • shape_a • shape_b How are prior beliefs structured?
  • 11. Designing Prior Beliefs for SigOpt At present, SigOpt allows two types of prior beliefs. • Normal • mean • scale • Beta • shape_a • shape_b How are prior beliefs structured? https://app.sigopt.com/docs/overview/parameter_priors
  • 12. Benefits of Prior Beliefs How will prior beliefs affect an experiment? This prior beliefs feature gives users new power. Previously: Features existed to define or change the problem under consideration. Prior beliefs: The problem remains the same, but the optimization engine changes. Choices made when defining prior beliefs can affect the optimization process in both good and bad ways.
  • 13. Benefits of Prior Beliefs This prior beliefs feature gives users new power. Previously: Features existed to define or change the problem under consideration. Prior beliefs: The problem remains the same, but the optimization engine changes. Choices made when defining prior beliefs can affect the optimization process in both good and bad ways. How will prior beliefs affect an experiment?
  • 14. Benefits and Dangers of Prior Beliefs This prior beliefs feature gives users new power. Previously: Features existed to define or change the problem under consideration. Prior beliefs: The problem remains the same, but the optimization engine changes. Choices made when defining prior beliefs can affect the optimization process in both good and bad ways. How will prior beliefs affect an experiment?
  • 15. Benefits and Dangers of Prior Beliefs This prior beliefs feature gives users new power. Previously: Features existed to define or change the problem under consideration. Prior beliefs: The problem remains the same, but the optimization engine changes. Choices made when defining prior beliefs can affect the optimization process in both good and bad ways. How will prior beliefs affect an experiment? Can provide early acceleration Eventually, the observed data will overpower any prior beliefs
  • 17. SigOpt. Confidential. Questions? Ask now, or feel free to email: contact@sigopt.com Slides will be available on slideshare … check your email in the next few days.
  • 18. SigOpt. Confidential. Check out our YouTube channel: Learn more about SigOpt Read our research and product blog at blog.sigopt.com. See more videos at https://sigopt.com/resources/videos Try our solution: Sign up at sigopt.com/try-it today. Click Here Upcoming webinars: ● Detecting COVID-19 Cases with Deep Learning Tuesday, June 9 at 10am PT / 1pm ET