Skynet is coming!


Published on

Yes, it really is! Computer scientists around the world are teaching machines how to learn, how to think. You already use machines that are pretty smart, e.g. your camera which recognizes faces, your phone that you can converse with. Soon, you will be using a car which drives by itself. Come to this fun (and no-math-involved) talk to understand how these cool technologies work, and where all this is taking us. I promise several ‘Wow, this is amazing’ moments!

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Skynet is coming!

  1. 1. Skynet is coming! Deepak Azad M.Sc. Student, Computer Science Junior Fellows Speaker Series St. John’s College UBC Image: 19, 2013
  2. 2. Wait… What is Skynet?Image:
  3. 3. Overview• Do we know how we think?• Machines that can think• How do they work?• Implications for mankind
  4. 4. Do we know how we see?–– Image:
  5. 5. Do we know how we think?• The doctor has good news and bad news• Bad news: You have tested positive for HIV, and that the test is 99% accurate (i.e., the probability of testing positive given that you have the disease is 0.99, as is the probability of testing negative given that you don’t have the disease).• Good news: Only 1 in 10,000 people are HIV- Image: 878683/4411521534_thinking20web20pic_ans positive wer_3_xlarge.jpeg• How worried should you be?
  6. 6. Do we know how we think?• I have some money in my pocket and I will give $100 to the person(s) who ask the best question(s), as judged by me.• How many of you think that someone here will get $100? Image:
  7. 7. Let’s do the math• Lets say 1,000,000 people took the test.• Number of people who have the disease = 100 (Remember 1 in 10,000)• Number of people who tested positive AND who have the disease = 0.99 * 100 = 99• Number of people who tested positive BUT did not have the disease = 0.01 * 999900 = 9999• Probability that you actually have the disease = 99 / (99+9999) = 0.0098 Only 0.98% content/uploads/2009/12/DontWorryBeHappy.jpg
  8. 8. When should you worry?• Lets say you underwent another test which is also 99% accurate and you tested positive. = 49.5% 12/04/worry.gif
  9. 9. If all your friends jumped off a bridge…
  10. 10. Lesson posterior = likelihood x prior or new knowledge = data x old knowledge• This is known as Bayes’ rule http://skepticism-images.s3-website-us-east- Bayes.jpg
  11. 11. Lessons• We only ever see a tiny bit• We are pretty good at learning from data• We can recognize patterns quite well
  12. 12. Autonomous Car•
  13. 13. Face Recognition
  14. 14. Pedestrian detection•
  15. 15. Twitter Sentiment Analysis• %20Maple%20leafs• ver%20Canucks• %20flames
  16. 16. Microsoft Kinect
  17. 17. Google Glass•
  18. 18. How does all this work ?Temperature Hours of Sun
  19. 19. FormallyEquation of line :
  20. 20. A little more complicated…Temperature Hours of Sun
  21. 21. Using prior knowledge and dealing with uncertanity
  22. 22. Watson playing JeapordyImages:
  23. 23. Implications: Will machines take over? Images: content/uploads/2012/11/halglados1.jpg
  24. 24. Implications: Will we need to work? Image:
  25. 25. Ethics and Morality• Doctors and Insurance companies use the same tools – one to cure and one to deny insurance• Should people be allowed to drive? – People’s freedom vs reliable machines