Successfully reported this slideshow.
Your SlideShare is downloading. ×

Machine Learning on the web - moving from Terminator to Star Trek

Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad
Ad

Check these out next

1 of 98 Ad

More Related Content

Advertisement

Similar to Machine Learning on the web - moving from Terminator to Star Trek (20)

More from Christian Heilmann (20)

Advertisement

Recently uploaded (20)

Machine Learning on the web - moving from Terminator to Star Trek

  1. 1. Machine Learning on the web - moving from Terminator to Star Trek Chris Heilmann @codepo8, Future Decoded 2016 - 2nd November 2016
  2. 2. CHRIS HEILMANN @CODEPO8 CHRISTIAN.HEILMANN@MICROSOFT.COMWELCOME TO THE FUTURE (DECODED)
  3. 3. CHRIS HEILMANN @CODEPO8 CHRISTIAN.HEILMANN@MICROSOFT.COM
  4. 4. Machine Learning, Artificial Intelligence, Connected Devices, Security Breaches, Privacy and Surveillance…
  5. 5. What does the future hold? 🔮🍵🍂
  6. 6. 😱 Constant machine surveillance 😱 Replacement of humans with cheaper, unfeeling machines 😱 Robot takeover and end of humankind unless we find a saviour to kill them by making them get into a logic loop.
  7. 7. 😍 Ubiquitous computing as a resource to call in on human decisions 😍 Machines taking on human tasks that are dangerous or get us sick in the long run 😍 Machines improving humans to work around shortcomings 😍 A change of values: knowledge instead of monetary goals
  8. 8. Machine Learning and AI are incredibly hot topics and people are worried… 🔥💵📈💯📢
  9. 9. Expensive, creepy parrots… https://www.youtube.com/watch?v=W0_DPi0PmF0
  10. 10. Military research…
  11. 11. It is time for us as developers to show that it is not something in the future, but something humans can benefit from now. 🔓
  12. 12. The future of computing needs first and foremost people who are not afraid…
  13. 13. My first computer wasn’t able to do much…
  14. 14. This taught me to be not afraid of machines. 😨
  15. 15. Instead, it kept me curious and interested about them. And it made me always to look under the hood to see what I can use them for.🤔
  16. 16. Computers, programming and algorithms to me are tools to improve my work and do things I have to do in an easier fashion. 😨
  17. 17. WELCOME TO THE FOURTH INDUSTRIAL REVOLUTION! https://en.wikipedia.org/wiki/July_Revolution
  18. 18. Steam Engine https://de.wikipedia.org/wiki/Hugo_Lentz
  19. 19. Steam Engine > Electricity https://de.wikipedia.org/wiki/Telefon
  20. 20. Steam Engine > Electricity > Computer https://de.wikipedia.org/wiki/Computer
  21. 21. Each industrial revolution meant that people lost their jobs… 😨
  22. 22. At the same time the quality of life improved… 😄
  23. 23. Not for all, though, and this is where the problems start… 😠
  24. 24. What we call success these days turns out to be more and more of a scam. Making money by all means doesn’t help humanity, the market or the environment. And we only have one planet to live on - for now. 📉
  25. 25. And this is why this fourth industrial revolution is ours to define. We’re in a lucky position where our jobs are secure and we control what’s happening - for now. 📈
  26. 26. http://www.bbc.com/news/technology-34066941 http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf The robot takeover
  27. 27. http://www.bbc.com/news/technology-34066941 http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf The robot takeover
  28. 28. http://www.bbc.com/news/technology-34066941 http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf The robot takeover
  29. 29. http://www.bbc.com/news/technology-34066941 http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf The robot takeover
  30. 30. Robots and intelligent systems will take jobs away from us… It is not an if - it is a when. And it is soon. 🤖
  31. 31. http://www.bbc.com/news/technology-34066941 http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf How safe is my job? Abstract thinking Working with humans Images Words Numbers Routine Variety You’re out Get nervous You’re OK
  32. 32. Steam Engine > Electricity > Computer > Digital, physical and biological symbiosis… https://www.microsoft.com/microsoft-hololens/en-us
  33. 33. This feels a lot of like science fiction…
  34. 34. The Future, 1991 Star Trek TNG S04E19 - The nth degree - http://gb.imdb.com/title/tt0708813/
  35. 35. Lifts…
  36. 36. Tablet computers…
  37. 37. Server synchronisation
  38. 38. Language recognition
  39. 39. Finding people
  40. 40. Nowadays, we have supercomputers in our pockets…
  41. 41. http://www.smartinsights.com/mobile-marketing/mobile-marketing-analytics/mobile-marketing-statistics/ And mobile computing is constantly on the rise…
  42. 42. This means that the interaction between human and machine changed drastically…
  43. 43. Keyboard and Mouse Camera Speaker Microphone Device Movement Body functions
  44. 44. Text, images, sound and video… Speech recognition, speech synthesis, vibration, notifications…
  45. 45. Click and type Touch, move, get closer, asking questions and showing emotions…
  46. 46. All this generates an amazing amount of data that is not in a simple to parse format and at times very erratic…
  47. 47. This is why we needed to find a way to get computers to understand humans better… Sound, images and movement 00101011010010101…
  48. 48. In the fourth industrial revolution, information is oil.
  49. 49. Computers are great to churn through large sets of data and find patterns without getting tired or sloppy - very much what humans are not.
  50. 50. The more data we give the machines, the better the results will be…
  51. 51. Our job right now is to create mobile interfaces that are simple, human and fun to use…
  52. 52. Let’s acknowledge the elephant in the room first… https://www.flickr.com/photos/135045262@N04/30459872622/
  53. 53. Business and hype is all about the chat bots and personal assistants… 🗣🤖 Siri (Apple) Google Now (Google) Cortana (Microsoft) Alexa (Amazon) M (facebook)
  54. 54. https://techcrunch.com/2015/09/29/forget-apps-now-the-bots-take-over/
  55. 55. Explosive projections…
  56. 56. If you want to play with bots, we got your back… http://blog.decodedshow.com/2016/09/27/i-have-a-bot-for-that/ https://docs.botframework.com/en-us/node/builder/chat/UniversalBot/ @ritazh
  57. 57. If you want to play with bots, we got your back… http://blog.decodedshow.com/2016/09/27/i-have-a-bot-for-that/ https://docs.botframework.com/en-us/node/builder/chat/UniversalBot/ @ritazh
  58. 58. However, more traditional interfaces like fat client and web need more intelligent and simple interfaces and content, too. They aren’t going away…
  59. 59. And machine learning is everywhere…
  60. 60. SCRIPTING SOLUTIONS FOR ALTERNATIVE CONTENTIntelligent image resizing
  61. 61. https://github.com/jwagner/smartcrop.js/Smartcrop.js
  62. 62. High contrast tricks https://blog.imgix.com/2015/10/21/automatic- point-of-interest-cropping-with-imgix.html
  63. 63. Intelligent resizing http://cloudinary.com/blog/ automatically_art_directed_responsive_images
  64. 64. Making photos findable… https://photos.google.com
  65. 65. Making photos indexable and understandable… Image may contain: dog, outdoor and nature
  66. 66. How does that work? Image may contain: dog, outdoor and nature https://code.facebook.com/posts/561187904071636
  67. 67. Learning from good data… http://image-net.org/
  68. 68. Learning from *lots* of data… http://image-net.org/
  69. 69. Learning from *lots* of data… https://github.com/ openimages/dataset
  70. 70. Refining the data with language https://research.googleblog.com/2016/09/show-and-tell-image-captioning-open.html
  71. 71. Compare and enhance…
  72. 72. Detect syntax, mix and match
  73. 73. Add visual information
  74. 74. http://captionbot.ai APIs and demos to use yourself…
  75. 75. Detection Result: JSON: [ { "faceRectangle": { "width": 109, "height": 109, "left": 62, "top": 62 }, "attributes": { "age": 31, "gender": "male", "headPose": { "roll": "2.9", "yaw": "-1.3", "pitch": "0.0" } "faceLandmarks": { "pupilLeft": { "x": "93.6", "y": "88.2" }, "pupilRight": { "x": "138.4", "y": "91.7" }, ... Detect Faces…
  76. 76. Verification Result: JSON: [ { "isIdentical":false, "confidence":0.01 } ] Verify Faces
  77. 77. Cluster automatically…
  78. 78. https://www.microsoft.com/cognitive-services/en-us/video-api Understanding and fixing video automatically… 📹 Stabilise shaky videos 📹 Detect and track faces 📹 Detect motion 📹 Generate video thumbnails
  79. 79. Machines and code are there to help humans communicate better. The more useful our “intelligent” solutions are, the easier it will be for people to stop fearing innovation… 🎙
  80. 80. Fear of innovation in many cases shows itself by making sure the first use of an “intelligent” system leads to a failure. We love to show our supremacy over technology. 🍌
  81. 81. In the worst case, it is flat our trolling and sabotage - and the press loves this. 😒
  82. 82. 🤔
  83. 83. Quality is dependent on data and context… https://www.microsoft.com/cognitive-services/en-us/custom-recognition-intelligent-service-cris
  84. 84. Get inspired by the amazing stuff that is already happening - and create more human solutions. 🦄
  85. 85. https://www.youtube.com/watch?v=R2mC-NUAmMk Smart Glasses listening to questions, using OCR and image analysis…
  86. 86. https://channel9.msdn.com/Events/Build/2016/KEY01 “Live” translation of calls.
  87. 87. Writing and reading tools that change display of text catered to the needs of people with learning disabilities… https://blogs.office.com/2016/01/19/learning- tools-for-onenote-improves-learning-for-all/ https://www.youtube.com/watch?v=3Ztr44aKmQ8
  88. 88. Adding automated content in context… https://research.googleblog.com/2016/05/aw-so-cute-allo-helps-you-respond-to.html
  89. 89. It is up to us to make this revolution one that improves all mankind, and not only makes some quick money and shows us more ads. 8
  90. 90. Humans are great people :) No need to make them do things they are not meant to do.🤕
  91. 91. Lift assistant… https://www.flickr.com/photos/breynolds/353144443/
  92. 92. CHRIS HEILMANN @CODEPO8 CHRISTIANHEILMANN.COM THANKS!

×