Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Machine Learning to Predict Responses

141 views

Published on

This is the slide deck from my talk at PHP[TEK] 2017 about Machine Learning. The case study is my invention S. A. M. (Suspicious Activity Monitor). S. A. M. is a predictive policing machine learning program that predicts crime given a set of circumstances. This talk reveals the secrets behind S. A. M. and takes conference attendees through a step-by-step process to create a machine learning program like S. A. M. To learn more about S. A. M., visit http://www.iamsam.tech/.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Machine Learning to Predict Responses

  1. 1. MACHINE LEARNING Amazon Machine Learning to Predict Responses “As a young girl, I believed that computers would soon control the world, so I wanted to be the one to control the computers.” - Kesha Williams Kesha Williams @KeshaUCI
  2. 2. Conference Information
  3. 3. MINORITY REPORT MOVIE "precrime" unit https://youtu.be/jdl6eAIx2K4
  4. 4. Suspicious Activity? Is this suspicious activity? Is a crime going to be committed? Can you look at a person and determine if they are going to commit a crime? SAM can! Twitter User
  5. 5. Suspicious Activity? Is this suspicious activity? Is a crime going to be committed? Can you look at a person and determine if they are going to commit a crime? SAM can! Twitter User
  6. 6. Suspicious Activity? Twitter User Is this suspicious activity? Is a crime going to be committed? Can you look at a person and determine if they are going to commit a crime? SAM can!
  7. 7. Suspicious Activity? Is this suspicious activity? Is a crime going to be committed? Can you look at a person and determine if they are going to commit a crime? SAM can! Twitter User
  8. 8. Suspicious Activity? Is this suspicious activity? Is a crime going to be committed? Can you look at a person and determine if they are going to commit a crime? SAM can! Twitter User
  9. 9. Psychological Science Says 1/10 of 1 second
  10. 10. About Me Software Engineer 20+ Years
  11. 11. S. A. M. (Suspicious Activity Monitor) http://www.iamsam.tech
  12. 12. Sound Far Fetched? Disbelief?
  13. 13. Civil Images FBI 4.3+ Million “Civil” Images
  14. 14. Predictive Policing ML at arraignments cut repeat domestic violence New Domestic Violence Cases50%
  15. 15. United Kingdom (UK) Metropolitan Police • ML produces daily hotspot maps • Deploy additional forces to hotspots 26% Violent Crimes
  16. 16. Pittsburgh, Pennsylvania Police Department September 2016
  17. 17. Chicago Police Department (CPD) Strategic Subject List Or Heat List
  18. 18. This Talk • Secrets Revealed • Covers The How • New to ML • No Computer Vision, Twitter APIs
  19. 19. Workshop Materials http://kesha-williams.thinkific.com/courses/machine-learning-tech-conf
  20. 20. Workshop Schedule Topic Lesson 1: Introducing S. A. M. Lesson 2: Development Environment Setup Lesson 3: Machine Learning Overview ****Break**** Lesson 4: S. A. M. Back-End Interface ****Break**** Lesson 5: S. A. M. Front-End Interface Lesson 6: Bonus Module – IoT Button ****Break**** Wrap-Up/Q&A Raffle for Prizes
  21. 21. Future of ML and SAM 2.0 • Future of ML • 1954 • 2017 and Beyond • S. A. M. 2.0 • Twitter StreamsAPI & AWS Kinesis • Facial Recognition • Retrain Model to include more factors • Substance abuse, criminal history, etc. • Voice Interface • Amazon Alexa
  22. 22. Next Step #1 SAM 2.0
  23. 23. Next Step #2 http://kesha- williams.thinkific.com/ courses/sam-full- edition/
  24. 24. Questions?
  25. 25. Resource Links • SlideShare • https://www.slideshare.net/Hyperlinks /amazon-machine-learning-to-predict- responses • GitHub • https://github.com/keshaTechConfSam ples/sam • Course Link • http://kesha- williams.thinkific.com/courses/machin e-learning-tech-conf http://www.iamsam.tech/ @iamsam_tech
  26. 26. Contact Kesha Williams @KeshaUCI https://www.linkedin.com/in/java-rock-star-kesha/ kesha@s4technology.com 678-364-2767
  27. 27. Raffle
  28. 28. Thank You

×