SlideShare a Scribd company logo
1 of 14
AI
system validation
Devesh Rajadhyax
Founder and CEO, Cere Labs
Contents
• Introduction to AI
• Machine Learning
• Model Validation
• Challenges
© Cere Labs Pvt. Ltd. 2
Introduction to AI
© Cere Labs Pvt. Ltd.3
About Cere Labs
© Cere Labs Pvt. Ltd. 4
• Three year old privately held company from Mumbai,
India
• Creator of AI platform ‘Cerescope’ used for
unstructured data processing
• Applications in banking, healthcare/pharma,
manufacturing, agriculture, retail
• 20+ people – mostly AI engineers, expertise in
Machine Learning, Deep Learning, Cognitive
Computing
• Part of SAP Co-innovation Labs
• Research interests – GA, GAN, Speech processing,
What is AI?
• AI is the pursuit of imitating human capabilities
© Cere Labs Pvt. Ltd. 5
Human capabilities
Perception
Action
Computation and memory
• Speech
recognition
• Computer vision
• Natural language
processing
• Writing recognition
• Robotics
• Natural language
generation
• Text to speech
• Conversation
• Predictive Analytics
• Anomaly Detection
• Planning and
reasoning
• Decisions
• Pattern recognition
AI based computation
© Cere Labs Pvt. Ltd. 6
Technologies
• Machine Learning
• Deep Learning
• Natural Language Processing
• Logic programming
Buddha and the child
Features of AI computing
• Learning not algorithmic
• Uncertain and not well-defined
problems
• Inaccurate
Examples
• Prediction
• Sales forecast
• Demand forecast
• Anomaly
• Fraud detection
• Disease detection
• Reasoning
• Chess programs
• Expert systems
Well known AI systems
© Cere Labs Pvt. Ltd. 7
Recommendation systems Demand forecasting
Failure prediction Customer behaviour
Machine Learning
© Cere Labs Pvt. Ltd.8
Introduction
© Cere Labs Pvt. Ltd. 9
Classification
• Separating given data in
classes
Regression
• Coming up with a number
f(x)x y
ML
algorithm
(x1,y1),
(x2,y2), …
f(x)
How ML works
Algorithms
• Supervised
• Linear regression
• Support Vector
Machines
• Random Forest
• Unsupervised
• Clustering
• PCA
ML project cycle
© Cere Labs Pvt. Ltd. 10
Credit Card Fraud
Detection
• Features
• Member data
• Transaction data
• Trend data
• Outcome
• Fraud/Normal
• Data set
• 1 million records
• Data use
• 80% training
• 10% testing
• 10% validation
Validation
© Cere Labs Pvt. Ltd.11
Testing model performance
• Accuracy measurement
• Expressed as percentage
• In case of regression
• Root Mean Error (RME)
© Cere Labs Pvt. Ltd. 12
Actual Predicted Remarks
Normal Normal Correct
Fraud Fraud Correct
Normal Fraud Incorrect
Fraud Normal Incorrect
• Why the training error is of no use?
• What is the difference between validation
and test data set?
• Choosing the test data - why a tester will
have to be data scientist?
Challenges
• Accuracy – not quite that simple
• Data bias
• Over-fitting
• Concept drift and model re-training
© Cere Labs Pvt. Ltd. 13
Actual Predicted Remarks
Normal Normal Model worked well
Fraud Fraud Model worked well
Normal Fraud Mistake but less sensitive
Fraud Normal Serious error
Thanks
Devesh Rajadhyax
CEO, Cere Labs Pvt. Ltd.
devesh.rajadhyax@cerelabs.com

More Related Content

Similar to AI Systems Validation - ATA Pune 18th Meetup

Hire india ppt it
Hire india ppt  itHire india ppt  it
Hire india ppt it
Hire India
 

Similar to AI Systems Validation - ATA Pune 18th Meetup (20)

Chanchal Chatterjee PARTNERS 2017 Oct24
Chanchal Chatterjee PARTNERS 2017 Oct24Chanchal Chatterjee PARTNERS 2017 Oct24
Chanchal Chatterjee PARTNERS 2017 Oct24
 
Building a Data Driven Culture and AI Revolution With Gregory Little | Curren...
Building a Data Driven Culture and AI Revolution With Gregory Little | Curren...Building a Data Driven Culture and AI Revolution With Gregory Little | Curren...
Building a Data Driven Culture and AI Revolution With Gregory Little | Curren...
 
Predictive analytics
Predictive analytics Predictive analytics
Predictive analytics
 
Platform for Data Scientists
Platform for Data ScientistsPlatform for Data Scientists
Platform for Data Scientists
 
Think Big | Enterprise Artificial Intelligence
Think Big | Enterprise Artificial IntelligenceThink Big | Enterprise Artificial Intelligence
Think Big | Enterprise Artificial Intelligence
 
AI in the Enterprise at Scale
AI in the Enterprise at ScaleAI in the Enterprise at Scale
AI in the Enterprise at Scale
 
Large Scale Graph Processing & Machine Learning Algorithms for Payment Fraud ...
Large Scale Graph Processing & Machine Learning Algorithms for Payment Fraud ...Large Scale Graph Processing & Machine Learning Algorithms for Payment Fraud ...
Large Scale Graph Processing & Machine Learning Algorithms for Payment Fraud ...
 
Machine Learning Development Company in Mohali
Machine Learning Development Company in MohaliMachine Learning Development Company in Mohali
Machine Learning Development Company in Mohali
 
Webinar - Patient Readmission Risk
Webinar - Patient Readmission RiskWebinar - Patient Readmission Risk
Webinar - Patient Readmission Risk
 
AI & AWS DeepComposer
AI & AWS DeepComposerAI & AWS DeepComposer
AI & AWS DeepComposer
 
Drive Away Fraudsters With Driverless AI - Venkatesh Ramanathan, Senior Data ...
Drive Away Fraudsters With Driverless AI - Venkatesh Ramanathan, Senior Data ...Drive Away Fraudsters With Driverless AI - Venkatesh Ramanathan, Senior Data ...
Drive Away Fraudsters With Driverless AI - Venkatesh Ramanathan, Senior Data ...
 
Intelligent Digital Mesh Testing
Intelligent Digital Mesh TestingIntelligent Digital Mesh Testing
Intelligent Digital Mesh Testing
 
Helping B2B markerters to find more waldos
Helping B2B markerters to find more waldosHelping B2B markerters to find more waldos
Helping B2B markerters to find more waldos
 
Introduction to Machine Learning
Introduction to Machine LearningIntroduction to Machine Learning
Introduction to Machine Learning
 
Hire india ppt it
Hire india ppt  itHire india ppt  it
Hire india ppt it
 
Scaling Training Data for AI Applications
Scaling Training Data for AI ApplicationsScaling Training Data for AI Applications
Scaling Training Data for AI Applications
 
Building enterprise advance analytics platform
Building enterprise advance analytics platformBuilding enterprise advance analytics platform
Building enterprise advance analytics platform
 
Chatbots: Automated Conversational Model using Machine Learning
Chatbots: Automated Conversational Model using Machine LearningChatbots: Automated Conversational Model using Machine Learning
Chatbots: Automated Conversational Model using Machine Learning
 
Predictive Analytics: Advanced techniques in data mining
Predictive Analytics: Advanced techniques in data miningPredictive Analytics: Advanced techniques in data mining
Predictive Analytics: Advanced techniques in data mining
 
Functionalities in AI Applications and Use Cases (OECD)
Functionalities in AI Applications and Use Cases (OECD)Functionalities in AI Applications and Use Cases (OECD)
Functionalities in AI Applications and Use Cases (OECD)
 

More from Agile Testing Alliance

More from Agile Testing Alliance (20)

#Interactive Session by Anindita Rath and Mahathee Dandibhotla, "From Good to...
#Interactive Session by Anindita Rath and Mahathee Dandibhotla, "From Good to...#Interactive Session by Anindita Rath and Mahathee Dandibhotla, "From Good to...
#Interactive Session by Anindita Rath and Mahathee Dandibhotla, "From Good to...
 
#Interactive Session by Ajay Balamurugadas, "Where Are The Real Testers In T...
#Interactive Session by  Ajay Balamurugadas, "Where Are The Real Testers In T...#Interactive Session by  Ajay Balamurugadas, "Where Are The Real Testers In T...
#Interactive Session by Ajay Balamurugadas, "Where Are The Real Testers In T...
 
#Interactive Session by Jishnu Nambiar and Mayur Ovhal, "Monitoring Web Per...
#Interactive Session by  Jishnu Nambiar and  Mayur Ovhal, "Monitoring Web Per...#Interactive Session by  Jishnu Nambiar and  Mayur Ovhal, "Monitoring Web Per...
#Interactive Session by Jishnu Nambiar and Mayur Ovhal, "Monitoring Web Per...
 
#Interactive Session by Pradipta Biswas and Sucheta Saurabh Chitale, "Navigat...
#Interactive Session by Pradipta Biswas and Sucheta Saurabh Chitale, "Navigat...#Interactive Session by Pradipta Biswas and Sucheta Saurabh Chitale, "Navigat...
#Interactive Session by Pradipta Biswas and Sucheta Saurabh Chitale, "Navigat...
 
#Interactive Session by Apoorva Ram, "The Art of Storytelling for Testers" at...
#Interactive Session by Apoorva Ram, "The Art of Storytelling for Testers" at...#Interactive Session by Apoorva Ram, "The Art of Storytelling for Testers" at...
#Interactive Session by Apoorva Ram, "The Art of Storytelling for Testers" at...
 
#Interactive Session by Nikhil Jain, "Catch All Mail With Graph" at #ATAGTR2023.
#Interactive Session by Nikhil Jain, "Catch All Mail With Graph" at #ATAGTR2023.#Interactive Session by Nikhil Jain, "Catch All Mail With Graph" at #ATAGTR2023.
#Interactive Session by Nikhil Jain, "Catch All Mail With Graph" at #ATAGTR2023.
 
#Interactive Session by Ashok Kumar S, "Test Data the key to robust test cove...
#Interactive Session by Ashok Kumar S, "Test Data the key to robust test cove...#Interactive Session by Ashok Kumar S, "Test Data the key to robust test cove...
#Interactive Session by Ashok Kumar S, "Test Data the key to robust test cove...
 
#Interactive Session by Seema Kohli, "Test Leadership in the Era of Artificia...
#Interactive Session by Seema Kohli, "Test Leadership in the Era of Artificia...#Interactive Session by Seema Kohli, "Test Leadership in the Era of Artificia...
#Interactive Session by Seema Kohli, "Test Leadership in the Era of Artificia...
 
#Interactive Session by Ashwini Lalit, RRR of Test Automation Maintenance" at...
#Interactive Session by Ashwini Lalit, RRR of Test Automation Maintenance" at...#Interactive Session by Ashwini Lalit, RRR of Test Automation Maintenance" at...
#Interactive Session by Ashwini Lalit, RRR of Test Automation Maintenance" at...
 
#Interactive Session by Srithanga Aishvarya T, "Machine Learning Model to aut...
#Interactive Session by Srithanga Aishvarya T, "Machine Learning Model to aut...#Interactive Session by Srithanga Aishvarya T, "Machine Learning Model to aut...
#Interactive Session by Srithanga Aishvarya T, "Machine Learning Model to aut...
 
#Interactive Session by Kirti Ranjan Satapathy and Nandini K, "Elements of Qu...
#Interactive Session by Kirti Ranjan Satapathy and Nandini K, "Elements of Qu...#Interactive Session by Kirti Ranjan Satapathy and Nandini K, "Elements of Qu...
#Interactive Session by Kirti Ranjan Satapathy and Nandini K, "Elements of Qu...
 
#Interactive Session by Sudhir Upadhyay and Ashish Kumar, "Strengthening Test...
#Interactive Session by Sudhir Upadhyay and Ashish Kumar, "Strengthening Test...#Interactive Session by Sudhir Upadhyay and Ashish Kumar, "Strengthening Test...
#Interactive Session by Sudhir Upadhyay and Ashish Kumar, "Strengthening Test...
 
#Interactive Session by Sayan Deb Kundu, "Testing Gen AI Applications" at #AT...
#Interactive Session by Sayan Deb Kundu, "Testing Gen AI Applications" at #AT...#Interactive Session by Sayan Deb Kundu, "Testing Gen AI Applications" at #AT...
#Interactive Session by Sayan Deb Kundu, "Testing Gen AI Applications" at #AT...
 
#Interactive Session by Dinesh Boravke, "Zero Defects – Myth or Reality" at #...
#Interactive Session by Dinesh Boravke, "Zero Defects – Myth or Reality" at #...#Interactive Session by Dinesh Boravke, "Zero Defects – Myth or Reality" at #...
#Interactive Session by Dinesh Boravke, "Zero Defects – Myth or Reality" at #...
 
#Interactive Session by Saby Saurabh Bhardwaj, "Redefine Quality Assurance –...
#Interactive Session by  Saby Saurabh Bhardwaj, "Redefine Quality Assurance –...#Interactive Session by  Saby Saurabh Bhardwaj, "Redefine Quality Assurance –...
#Interactive Session by Saby Saurabh Bhardwaj, "Redefine Quality Assurance –...
 
#Keynote Session by Sanjay Kumar, "Innovation Inspired Testing!!" at #ATAGTR2...
#Keynote Session by Sanjay Kumar, "Innovation Inspired Testing!!" at #ATAGTR2...#Keynote Session by Sanjay Kumar, "Innovation Inspired Testing!!" at #ATAGTR2...
#Keynote Session by Sanjay Kumar, "Innovation Inspired Testing!!" at #ATAGTR2...
 
#Keynote Session by Schalk Cronje, "Don’t Containerize me" at #ATAGTR2023.
#Keynote Session by Schalk Cronje, "Don’t Containerize me" at #ATAGTR2023.#Keynote Session by Schalk Cronje, "Don’t Containerize me" at #ATAGTR2023.
#Keynote Session by Schalk Cronje, "Don’t Containerize me" at #ATAGTR2023.
 
#Interactive Session by Chidambaram Vetrivel and Venkatesh Belde, "Revolution...
#Interactive Session by Chidambaram Vetrivel and Venkatesh Belde, "Revolution...#Interactive Session by Chidambaram Vetrivel and Venkatesh Belde, "Revolution...
#Interactive Session by Chidambaram Vetrivel and Venkatesh Belde, "Revolution...
 
#Interactive Session by Aniket Diwakar Kadukar and Padimiti Vaidik Eswar Dat...
#Interactive Session by Aniket Diwakar Kadukar and  Padimiti Vaidik Eswar Dat...#Interactive Session by Aniket Diwakar Kadukar and  Padimiti Vaidik Eswar Dat...
#Interactive Session by Aniket Diwakar Kadukar and Padimiti Vaidik Eswar Dat...
 
#Interactive Session by Vivek Patle and Jahnavi Umarji, "Empowering Functiona...
#Interactive Session by Vivek Patle and Jahnavi Umarji, "Empowering Functiona...#Interactive Session by Vivek Patle and Jahnavi Umarji, "Empowering Functiona...
#Interactive Session by Vivek Patle and Jahnavi Umarji, "Empowering Functiona...
 

Recently uploaded

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
vu2urc
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Joaquim Jorge
 

Recently uploaded (20)

Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
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...
 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day Presentation
 
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
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonets
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
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
 
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
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
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
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
 
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
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 

AI Systems Validation - ATA Pune 18th Meetup

  • 2. Contents • Introduction to AI • Machine Learning • Model Validation • Challenges © Cere Labs Pvt. Ltd. 2
  • 3. Introduction to AI © Cere Labs Pvt. Ltd.3
  • 4. About Cere Labs © Cere Labs Pvt. Ltd. 4 • Three year old privately held company from Mumbai, India • Creator of AI platform ‘Cerescope’ used for unstructured data processing • Applications in banking, healthcare/pharma, manufacturing, agriculture, retail • 20+ people – mostly AI engineers, expertise in Machine Learning, Deep Learning, Cognitive Computing • Part of SAP Co-innovation Labs • Research interests – GA, GAN, Speech processing,
  • 5. What is AI? • AI is the pursuit of imitating human capabilities © Cere Labs Pvt. Ltd. 5 Human capabilities Perception Action Computation and memory • Speech recognition • Computer vision • Natural language processing • Writing recognition • Robotics • Natural language generation • Text to speech • Conversation • Predictive Analytics • Anomaly Detection • Planning and reasoning • Decisions • Pattern recognition
  • 6. AI based computation © Cere Labs Pvt. Ltd. 6 Technologies • Machine Learning • Deep Learning • Natural Language Processing • Logic programming Buddha and the child Features of AI computing • Learning not algorithmic • Uncertain and not well-defined problems • Inaccurate Examples • Prediction • Sales forecast • Demand forecast • Anomaly • Fraud detection • Disease detection • Reasoning • Chess programs • Expert systems
  • 7. Well known AI systems © Cere Labs Pvt. Ltd. 7 Recommendation systems Demand forecasting Failure prediction Customer behaviour
  • 8. Machine Learning © Cere Labs Pvt. Ltd.8
  • 9. Introduction © Cere Labs Pvt. Ltd. 9 Classification • Separating given data in classes Regression • Coming up with a number f(x)x y ML algorithm (x1,y1), (x2,y2), … f(x) How ML works Algorithms • Supervised • Linear regression • Support Vector Machines • Random Forest • Unsupervised • Clustering • PCA
  • 10. ML project cycle © Cere Labs Pvt. Ltd. 10 Credit Card Fraud Detection • Features • Member data • Transaction data • Trend data • Outcome • Fraud/Normal • Data set • 1 million records • Data use • 80% training • 10% testing • 10% validation
  • 11. Validation © Cere Labs Pvt. Ltd.11
  • 12. Testing model performance • Accuracy measurement • Expressed as percentage • In case of regression • Root Mean Error (RME) © Cere Labs Pvt. Ltd. 12 Actual Predicted Remarks Normal Normal Correct Fraud Fraud Correct Normal Fraud Incorrect Fraud Normal Incorrect • Why the training error is of no use? • What is the difference between validation and test data set? • Choosing the test data - why a tester will have to be data scientist?
  • 13. Challenges • Accuracy – not quite that simple • Data bias • Over-fitting • Concept drift and model re-training © Cere Labs Pvt. Ltd. 13 Actual Predicted Remarks Normal Normal Model worked well Fraud Fraud Model worked well Normal Fraud Mistake but less sensitive Fraud Normal Serious error
  • 14. Thanks Devesh Rajadhyax CEO, Cere Labs Pvt. Ltd. devesh.rajadhyax@cerelabs.com