Demystifying AL and ML

Presentation at Hour-of-Code 2021

Transcript

1. http://eazify.net/decks/hoc
3. a = int(input()) b = int(input()) c = a+b print(c) photos = db.GetPhotos(user) PhotoGallery.Display(photos) PhotoGallery.OnClick(p=> DisplayPhoto(p))
4. def koch_curve(n,x): if n==0: turtle.forward(x) else: koch_curve(n-1,x/3) turtle.left(60) koch_curve(n-1,x/3) turtle.right(120) koch_curve(n-1,x/3) turtle.left(60) koch_curve(n-1,x/3)
6. Let machines learn to take decisions by themselves by looking at data…
7. No # of pages Year of publication Book type Price 1 350 2017 Novel 450 2 100 2006 Detective 300 3 10 2015 Children 500 4 550 1950 Science 1200 5 20 10 Children 350 6 225 2000 Textbook ??? Features (Xi ) Label (Yi ) Training set
8. # of pages 150 x x x x x x x x x x 210 p. Price Model: Y = f(X) = aX+b Optimization: Find parameters a and b, such that L(a,b) → min Error: L(a,b) – sum of distances from point to the line
11. http://eazify.net/titanic_dataset pclass survived name gender age sibsp parch 1 1 Allen, Miss. Elisabeth Walton female 29 0 0 1 1 Allison, Master. Hudson Trevor male 0,9167 1 2 1 0 Allison, Miss. Helen Loraine female 2 1 2 1 0 Allison, Mr. Hudson Joshua Creighton male 30 1 2 1 0 Allison, Mrs. Hudson J C (Bessie Waldo Daniels) female 25 1 2 .. .. …. … … … …
12. Using Pivot Tables
14. Linear Model Tree Model
15. http://github.com/microsoft/ml-for-beginners
18. IF (at bus stop AND bus arrives) THEN action (get into the bus) IF (on the bus AND paid AND empty seat) THEN action (sit down). IF (on bus AND unpaid) THEN action (pay charges). IF (bus arrives at destination) THEN action (get down from the bus). Photo by Taylor Vatem and Alina Grubnyak on Unsplash
19. 𝑍 = 𝑓1 𝑊1 𝑋 + 𝑏1 𝑌 = 𝑓2 𝑊2 𝑍 + 𝑏2 Learning algorithm:
24. Sips of tea: 154 % Time at work: 64%
27. Computer Vision Face API Text Analytics LUIS Custom Vision Lobe.ai Automatic ML Azure ML Designer
28. Math Python Cloud Where to go: Microsoft Learn -- http://aka.ms/uchi
29. http://aka.ms/studstart http://aka.ms/az4stud http://aka.ms/azfree
30. Reactive Breain (algorithm) Knowledgebase Cloud Neural Net + statistics
31. arXiv:1610.05256 [cs.CL] arXiv:2009.13682v1 [cs.CV] DOI: 10.1109/ICCV.2015.123 arXiv:1803.05567 [cs.CL]
32. British scientists from John Lennon Institute in Liverpool last week have completed studies on the effect of a virus on musical tastes of British young people. They have found out that Brits are fans of choral music. The virus, which was given to them under the disguise of a vitamin supplement, is said to be harmless. "During the course of our studies, we accidentally discovered that the virus had another effect - it made people appreciate choral music," said Professor Ighorghgh. "It was very surprising. Our control group, which had not been given the virus, showed no musical preferences." The Liverpool scientists are now working with their colleagues in Estonia to find out how the virus does this. The Estonians are exploring the possibility that the virus changes the structure of the brain. Another possibility is that the virus modifies the gene that controls musical taste.
34. Some yummy goodness Text Analytics
36. Some yummy goodness
37. people_group 0.2578125 people_many 0.7109375 person 0.9945455 laptop 0.963514566 clothing 0.9370073 indoor 0.917977333 furniture 0.906451 woman 0.527442455 a group of people sitting at a table using a laptop
41. AVG LIKES BY MAIN CATEGORY
42. 2. Regression Run the Experiment and Train the Model
44. Affine Transformation Combines rotation, zooming and skew Is fully defined by transformation of 3 points cli = cf.FaceClient(endpoint, CognitiveServicesCredentials(key)) face_url = 'https://the.site/images/face.jpg' res = cli.face.detect_with_url(face_url, return_face_landmarks=True) {'pupil_left': {'x': 530.2, 'y': 387.6}, 'pupil_right': {'x': 725.6, 'y': 410.9}, 'nose_tip': {'x': 637.9, 'y': 514.3}, … def affine_transform(img,a): mc_x, mc_y = get_modmouth(a) tr = cv2.getAffineTransform( np.float32([(attrs['pupil_left']['x’], attrs['pupil_left']['y']), (attrs['pupil_right']['x’], attrs['pupil_right']['y']), (mc_x,mc_y)]), target_triangle return cv2.warpAffine(img,tr,(size,size))
45. http://github.com/shwars/DigiArt4Teens http://eazify.net/nbrun
46. Dmitry Soshnikov, «Growing Up» 2020, cognitive portrait
50. Q&A Artificial Intelligence Dmitry Soshnikov http://soshnikov.com dmitri@soshnikov.com
52. Knowledge Extraction from Text arXiv:2110.15453 [cs.CL] http://eazify.net/paper_analysis
53. FATE: Fairness, Accountability, Transparency, Ethical Fairness Accountability Transparency Ethical Is anyone treated differently? Is the training data representative/relevant? Are there systems in place to detect and address potential unfairness? Is there a mechanism to feedback/challenge results? Is there ultimately a human accountable? Is the AI behaviour verifiable (in the real world)? Is it clear an AI is being used? Is the AI clear, understandable and well-described? Are inadequacies/limitations clearly shared? Does it comply with privacy laws and contain appropriate use controls? What are the societal implications (think suffering, autonomy, equality)? Are there protections against unintended behaviour? Is there human monitoring? If lives are impacted, it’s important people understand how the decision was made