forked from canadalearningcode/llc-intro-to-ai-master
-
Notifications
You must be signed in to change notification settings - Fork 0
/
slides.html
1405 lines (1036 loc) · 85 KB
/
slides.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, maximum-scale=1">
<title>Ladies Learning Code</title>
<link rel="stylesheet" href="framework/css/slideshow.css" data-noprefix>
<link rel="stylesheet" href="framework/css/fonts.css" data-noprefix>
<link rel="stylesheet" href="framework/css/highlightjs/github.css" data-noprefix>
<link rel="stylesheet" href="framework/css/styles.css" data-noprefix>
<link rel="shortcut icon" href="framework/img/favicon.ico">
<!-- Takes care of CSS3 prefixes -->
<script src="framework/scripts/prefixfree.min.js"></script>
<!-- opens all links in a new window -->
<base target="_blank">
</head>
<!-- Timer/progress bar: Define the presentation duration using "data-duration" in minutes. -->
<body class="en" data-duration="360">
<!-- <section class="slide" data-markdown>
<script type="text/template">
#Ladies Learning Code Slide Template
Created by [Parinaz Sobhani](http://georgianpartners.com) for Ladies Learning Code.
View the slide presentation for instructions regarding content, customization options and style guides.
Email questions & comments to <[email protected]>.
</script>
</section> -->
<main>
<section class="slide welcome highlight">
<h1><img class="logo-stacked" src="framework/img/llc-logo-stacked-white.png" alt="Ladies Learning Code logo">Welcome!</h1>
<div class="instructions">
<!-- ADD WIFI INFO HERE -->
<!-- <h2>Get Connected</h2>
<p><strong>Wifi:</strong>Network Name</p>
<p><strong>Password:</strong>Password</p>
<hr> -->
<h2>Download<br> & Install</h2>
<ol class="downloads">
<li>Learner files (zip file): <a href="https://github.com/ladieslearningcode/llc-intro-to-ai-master/archive/master.zip">http://bit.ly/llc-intro-to-ai</a>
<ul>
<li>unzip the learner file (<em>extract all</em> if you’re on a PC)</li>
<li>open <em>slides.html</em> in the browser to view the slides</li>
</ul>
</li>
<li>Dataiku: <a href="https://www.dataiku.com/">https://www.dataiku.com/</a> **A very special thanks to Amazon (AWS) for generously hosting all instances of dataiku today!** </li>
<li>Chrome Browser: <a href="https://www.google.ca/chrome/browser/desktop/">https://www.google.ca/chrome</a></li>
</li>
</ol>
</div>
<footer>
<a class="left" rel="license" href="http://creativecommons.org/licenses/by-nc/4.0/"><img alt="Creative Commons License" src="framework/img/cc-by-nc.png" /></a>
<p class="left">Content created by <a href="https://georgianpartners.com/team_member/parinaz-sobhani/">Parinaz Sobhani</a> for <a prefix="cc: http://creativecommons.org/ns#" href="http://ladieslearningcode.com" property="cc:attributionName" rel="cc:attributionURL">Ladies Learning Code</a></p>
<p class="right">Use the left <span class="arrow">←</span> and right <span class="arrow">→</span> arrow keys to navigate</p>
</footer>
</section>
<section class="slide intro">
<img class="logo" src="framework/img/llc-logo-white.png" alt="Ladies Learning Code logo">
<h1 class="heading-bg">
<span>Solving Problems with Data: <br> Intro to AI and Machine Learning</span>
</h1>
<!-- FILL IN INSTRUCTOR DETAILS -->
<!-- <img class="instructor" src="framework/img/workshop/Parinaz.jpg" alt="Instructor Name">
<h2><span class="cursive">with</span> Instructor Name</h2>
<ul>
<li><a href="mailto:">[email protected]</a></li>
<li><a href="http://yourwebsite.com">http://yourwebsite.com</a></li>
<li><a href="http://twitter.com/your-twitter-handle-here">@instructor-handle</a></li>
</ul> -->
<div class="sponsor">
<p>In partnership with <br> <img src="framework/img/logo-accenture.svg" alt="Accenture"></p>
</div>
<footer>
<a class="left" rel="license" href="http://creativecommons.org/licenses/by-nc/4.0/"><img alt="Creative Commons License" src="framework/img/cc-by-nc.png" /></a>
<p class="left">Content created by <a href="https://georgianpartners.com/team_member/parinaz-sobhani/">Parinaz Sobhani</a> for <a prefix="cc: http://creativecommons.org/ns#" href="http://ladieslearningcode.com" property="cc:attributionName" rel="cc:attributionURL">Ladies Learning Code</a></p>
<p class="right">Use the left <span class="arrow">←</span> and right <span class="arrow">→</span> arrow keys to navigate</p>
</footer>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# A message from our sponsor, Accenture:
A message from our National Learn to Code Day sponsor Accenture and Canada's Artificial Intelligence lead Jodie Wallis
<iframe width="560" height="315" src="https://www.youtube.com/embed/EccUbIYW6IQ" frameborder="0" allowfullscreen></iframe>
</script>
</section>
<section class="slide two-col-list" data-markdown>
<script type="text/template">
#Agenda
<div class="table-of-contents"></div>
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Today's Project
Today we'll be exploring data to solve problems!
1. First we'll learn about databases by collecting data and entering it into our own database.
2. We'll be given a challenge and solve it using machine learning to predict the future.
3. We'll find our own challenge to solve.
<img width="40%" src="framework/img/workshop/hilary-shocked.gif" alt="Hilary Clinton shocked">
</script>
</section>
<section class="slide title" data-markdown>
<script type="text/template">
# What is AI?
</script>
</section>
<section class="slide" data-markdown data-toc>
<script type="text/template">
# What is AI?
Artificial intelligence (AI) is an area of computer science that involves the creation of machines that work and react like humans.
<img width="50%" src="framework/img/workshop/ai-william.gif" alt="AI Robot">
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# What is AI? cont'd
Some of the human activities that could be programmed using artificial intelligence include:
- **Computer vision:** ability of computers to identify objects, scenes, and activities in images
- **Natural Language Processing:** ability of computers to understand meaning from text or generating text that is readable
- **Speech Processing:** automatically transcribing human speech or generating speech from the corresponding text
<hr>
###Resources
[Demistifying AI](http://www.theatlantic.com/sponsored/deloitte-shifts/demystifying-artificial-intelligence/257/)
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Where is AI used today?
Despite it seeming like a technology far off into the future, Artificial Intelligence and Machine Learning are part of many of the tools we use every day. Here are just a few:
<img src="framework/img/workshop/googlemaps-logo.png" alt="google-maps" width="150" style="float:left; margin-right: 30px">
**Google Maps** uses AI to update addresses and street names based on street view pictures.
<br style="clear:both">
<br style="clear:both;margin-top: 50px;">
<img src="framework/img/workshop/FB-f-Logo__blue_100.png" alt="fb" width="150" style="float:left; margin-right: 30px">
**Facebook** uses AI to analyze text posted by users and better suggest ads they may need or be interested in. It also uses AI to analyse photos and identify faces.
<br style="clear:both;margin-top: 50px;">
<br style="clear:both;margin-top: 50px;">
<img src="framework/img/workshop/amazon-logo_transparent.png" alt="amazon" width="150" style="float:left; margin-right: 30px">
**Amazon** uses AI in its recommendation engine.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# 5 Artificial Intelligence Stats That Will Blow You Away
1. AI bots will power 85% of customer service interactions by 2020
1. Self-driving cars will save an estimated 300,000 lives per decade by reducing fatal traffic accidents
1. AI will replace 16% of jobs over next decade
1. Digital assistants will "know you" by 2018
1. 6 billion devices will request AI support
<hr>
### Resources
- [From Motley Fool](https://www.fool.com/investing/2016/12/10/9-artificial-intelligence-stats-that-will-blow-you.aspx)
- [More From Motley Fool](https://www.fool.com/investing/2016/06/19/10-stats-about-artificial-intelligence-that-will-b.aspx)
</script>
</section>
<section class="slide title ai-history" data-markdown>
<script type="text/template">
# History of AI
</script>
</section>
<section class="slide" data-markdown data-doc>
<script type="text/template">
# History of AI: A New Age?
<div class="flex">
<div>
<h2>Industrial Age</h2>
<h3>19th Century</h3>
<p style="background: #b109aa; color: white"><strong style="color: white;"> Machines take away the dirty work.</strong></p>
<p>Industrial equipment from looms to the cotton gin. </p>
<p>Machines relieve humans of onerous manual labour. </p>
</div>
<div>
<h2>Information Age</h2>
<h3>20th Century</h3>
<p style="background: #b109aa; color: white"><strong style="color: white;">Machines take away the dull and routine work.</strong></p>
<p>Automated interfaces from airline kiosks to call centers.</p>
<p>Technology relieves humans of routine transactions and clerical chores. </p>
</div>
<div>
<h2>2nd Machine Age</h2>
<h3>21st Century</h3>
<p style="background: #b109aa; color: white"><strong style="color: white;">Machines take away decisions.</strong></p>
<p>Intelligent systems from airfare pricing to health diagnostics</p>
<p>Algorithms make better choices than humans reliably and fast.</p>
</div>
</div>
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Why Now?
Artificial Intelligence is rapidly moving from the laboratory towards business and consumer application.
- <strong>Big Data</strong>: Live, interactive, automatically generated, and often self-correcting data that fuels real-time decisions and real-time responses.
- <strong>Cheaper Computing</strong>: Cloud computing, massively parallel processing and new CPUs are powering AI techniques that simply weren’t practical before.
- <strong>Better Algorithms</strong>: AI techniques have existed for decades but there is a surge in innovation and performance with the rapid growth in computational infrastructure, data and sensors.
</script>
</section>
<section class="slide" data-markdown data-toc>
<script type="text/template">
# History of AI: Notable Dates
<table>
<tr>
<td><strong style="background: #b109aa; color: white;">1763</strong> Thomas Bayes develops a framework for reasoning about the probability of events. Bayesian inference will become a leading approach in machine learning.</td>
<td><img src="framework/img/workshop/Thomas_Bayes.gif" alt="Thomas Bayes"></td>
</tr>
<tr>
<td><img src="framework/img/workshop/capek-robot.jpg" alt="Capek with robot"></td>
<td><strong style="background: #b109aa; color: white;">1921</strong> Czech writer Karel Čapek introduces the word "robot" in his play R.U.R. (Rossum's Universal Robots). The word "robot" comes from the word "robota" (work).</td>
</tr>
<tr>
<td><strong style="background: #b109aa; color: white;">1950</strong> Alan Turing publishes “Computing Machinery and Intelligence” in which he proposes “the imitation game” which will later become known as the “Turing Test.”</td>
<td><img src="framework/img/workshop/turing.jpg" alt="Turing"></td>
</tr>
<tr>
<td><img src="framework/img/workshop/wp-Marvin-Minsky-Claude-Shannon-Ray-Solomonoff-Plus-2-Dartmouth-1956-Conference.jpg" alt="Marvin and friends"></td>
<td><strong style="background: #b109aa; color: white;">August 31, 1955</strong> The term “artificial intelligence” is coined in a proposal for a “2 month, 10 man study of artificial intelligence” submitted by John McCarthy (Dartmouth College), Marvin Minsky (Harvard University), Nathaniel Rochester (IBM), and Claude Shannon (Bell Telephone Laboratories). The workshop, which took place a year later, in July and August 1956, is generally considered as the official birthdate of the new field.</td>
</tr>
<tr>
<td><strong style="background: #b109aa; color: white;">1961</strong> The first industrial robot, Unimate, starts working on an assembly line in a General Motors plant in New Jersey.</td>
<td><img src="framework/img/workshop/unimate-robot.jpg" alt="Unimate Robot"></td>
</tr>
<tr>
<td><iframe width="560" height="315" src="https://www.youtube.com/embed/qDrDUmuUBTo?rel=0&controls=0&showinfo=0" frameborder="0" allowfullscreen></iframe></td>
<td><strong style="background: #b109aa; color: white;">1968</strong> The film 2001: Space Odyssey is released, featuring Hal a sentient computer</td>
</tr>
<tr>
<td><strong style="background: #b109aa; color: white;">2009</strong> Google starts developing, in secret, a driverless car. In 2014, it became the first to pass, in Nevada, a U.S. state self-driving test.</td>
<td><iframe width="560" height="315" src="https://www.youtube.com/embed/uHbMt6WDhQ8?rel=0&controls=0&showinfo=0" frameborder="0" allowfullscreen></iframe></td>
</tr>
<tr>
<td><img src="framework/img/workshop/leesedolvsAl.jpg" alt="Lee Sedol vs AlphaGo"></td>
<td><strong style="background: #b109aa; color: white;">March 2016</strong> Google DeepMind's AlphaGo defeats Go champion Lee Sedol.</td>
</tr>
</table>
<hr>
### Resources
- [More history](https://www.forbes.com/sites/gilpress/2016/12/30/a-very-short-history-of-artificial-intelligence-ai/#22672d226fba)
</script>
</section>
<section class="slide women-in-ai" data-toc data-markdown>
<script type="text/template">
# History of AI: Notable Women
While the AI sector has been historically dominated by men, women are also making contributions to the field of Artificial Intelligence. Some of them work right here, in Canada.
In 2005, four women, organized the first "Women in Machine Learning Conference". The conference has now been running for over 10 years and WiML has a directory of women in Machine Learning, 1000 women strong.
## These are just a few of the amazing women working in Machine Learning and Artificial Intelligence in Canada and around the world:
Dr. Nancy Reid is a University Professor of Statistical Sciences at the University of Toronto. She is an Officer of the Order of Canada and the Director of the Canadian Statistical Sciences Institute. She works in statistical inference, with an emphasis on likelihood-based methods and higher order asymptotics.
Dr. Kathryn Hume - Product management and marketing lead for Integrate.ai, a startup helping large enterprises reinvent customer engagement by applying new AI technologies.
Dr. Bonolo Mathibela is a Machine Learning Scientist at IBM Research Africa, where she builds machine learning models that autonomously improve road traffic flow.
Dr. Kiri Wagstaff is a Principal Researcher in Machine Learning and Tactical Uplink Lead for the Opportunity Mars Rover (Mars Exploration Rovers) at the NASA Jet Propulsion Laboratory. She develops machine learning methods for spacecraft and space applications.
<hr>
###Resources
[More bios](http://wimlworkshop.org/directory-of-women-in-machine-learning/)
[Even more bios](https://www.forbes.com/sites/mariyayao/2017/05/18/meet-20-incredible-women-advancing-a-i-research/#7bb6e3e726f9),
[Quam Proxime, Kathryn Hume's enlightening blog on AI](https://quamproxime.com/)
[Equality of Opportunity in Machine Learning from the Google Research Blog](https://research.googleblog.com/2016/10/equality-of-opportunity-in-machine.html)
[Attacking Discrimination in ML from the Google Research Blog](https://research.google.com/bigpicture/attacking-discrimination-in-ml/)
</script>
</section>
<section class="slide title why-care" data-markdown>
<script type="text/template">
# Why should you care?
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Why should you care?
## 1. Without women and underrepresented groups, AI can have terrible consequences:
- When Siri was first introduced into the iPhone it was not able to understand women's voices because it had not been tuned to recognize higher pitch voices which women typically have. The developers of the speech recognition engine Siri uses programmed their own biases into the algorithms - as older men they were suffering from high frequency hearing loss. This is why it's so important to employ diverse teams of software developers.
- An AI system designed by Northpointe in the US to predict the likelihood that an alleged offender will commit another crime in the future was shown to demonstrate racial bias in its predictions.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Why should you care?
## 2. We need to solve the problems of each of us, not just the loudest amongst us.
We all need to care about being represented in the datasets and algorithms being used in AI today so that the machines that are made to help humanity, also help us.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Why should you care?
## 3. Women stand to lose more in an automated world:
According to the World Economic Forum, "twice as many women than men are likely to lose their jobs as automation replaces human labor." This is due to the fact that most jobs that can be automated are held by women. (e.g. cashiers)
<img class="medium-image" src="framework/img/workshop/cashier.jpg" alt="cashier">
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Why should you care?
## 4. Because our lives could be better!
Think about all the amazing apps that already make our lives better thanks to data and AI.
1. Chatbots:
<iframe src="https://player.vimeo.com/video/162458358" width="640" height="360" frameborder="0" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>
2. Traffic Data: Waze
<iframe width="560" height="315" src="https://www.youtube.com/embed/PPpZNzXqId0?rel=0&controls=0&showinfo=0" frameborder="0" allowfullscreen></iframe>
</script></section>
<section class="slide" data-markdown>
<script type="text/template">
# Why should you care?
## 5. Because with AI you can change the world!
AI is a fascinating field of discovery that can solve problems. So, why not get involved?
<hr>
### Resources
- [Why Women Are Twice As Likely As Men To Lose Their Jobs to Robots](https://www.weforum.org/agenda/2017/07/why-women-are-twice-as-likely-as-men-to-lose-their-job-to-robots)
- [Why we need diversity before ai takes over](http://www.lightreading.com/artificial-intelligence-machine-learning/why-we-need-diversity-before-ai-takes-over/a/d-id/729871)
- [Women vs the machine](http://foreignpolicy.com/2017/01/16/women-vs-the-machine/)
- [https://www.partnershiponai.org/#s-goals](https://www.partnershiponai.org/#s-goals)
- [An important group of women is changing things](https://www.fastcompany.com/3062932/mind-and-machine/ai-is-a-male-dominated-field-but-an-important-group-of-women-is-changing-th)
</script>
</section>
<section class="slide title data-everything" data-markdown>
<script type="text/template">
# Data
##Data is everything.
</script>
</section>
<section class="slide" data-markdown data-toc>
<script type="text/template">
# What is Data
Data is everything. More specifically, data is information.
- When we perceive the world around us, we are collecting and processing data.
- Data is also being collected all the time about our own actions and interactions by others
- Can you think of examples of when and where data about you is collected on a regular basis?
- stores ?
- government ?
- websites ?
- satellites ?
</script>
</section>
<section class="slide" data-toc data-markdown>
<script type="text/template">
# What are databases?
Databases are where data is saved and organized. Here are some examples of databases you might have seen before:
1. <img src="framework/img/workshop/rolodex-animated.gif" width="50%" style="float: right;" alt="rolodex"> The Old School Rolodex
(These were used to keep addresses organized in the old days when cell phones didn’t exist.)
<img width="50%" style="margin-right: 10px;clear: right; float: left;" src="framework/img/workshop/sample-db.png" alt="data table"><p style="margin-left: 20px;clear: right;">2. The Spreadsheet (You may have seen one on Excel, Google Sheets or Numbers)</p>
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Structured vs Unstructured Data
When we think of databases we're thinking of structured data. That is, data that can be neatly categorized and searched by a computer algorithm.
Unstructured data is a bit more messy and cannot neatly be categorized. Think for example of a **photo library**. We can now catalog our photos using advanced data markers like geo-tags, time stamps and even face recognition, but it's still very hard for a computer to really know what a photo is about.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
#Exercise 1: Our First Database (20 mins)
1. Get into a team of 4 with the students around you.
1. Select a team Data Input Specialist who will fill in everyone’s information in the database.
1. Go to the spreadsheet at: http://bit.ly/llc-ai-dataset
1. Go around your circle and give the Data Input Specialist the following information to enter into the spreadsheet. Be sure to only use the options listed below:
- Age Range: 18-29, 30-40, 40-50, 50+
- The number of hours watching TV per week
- The number of hours in social media per week
- The latitude of the city you're in (Clue: if the latitude is N of the equator, just add the number. if the latitude is S of the equator, add a "-" in front of the number. Do not add letters or non-numerical symbols to this column)
- The longitude of the city you're in (Clue: if the longitude is E of the prime meridian, just add the number. if the longitude is W of the prime meridian, add a "-" in front of the number. Do not add letters or non-numerical symbols to this column)
- Favourite Colour
- Favourite Protein: Fish, Beef, Chicken, Vegetarian/Vegan
- Favourite Film Genre: Comedy, Horror, Romantic Comedy, Drama, Action, Documentary
- Favourite Music Genre: Rock n Roll, Rap, Hip Hop, Punk, Classical, Indie, Pop, Jazz, World, Latin
- Favourite Pet: Cat, Dog, Fish, Other
It will start looking like this:
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
#Database Files
Databases can be saved in various formats.
Today we will be using datasets saved as **CSV files** (filename.csv) . CSV stands for *Comma Separated Values* because rows of data are saved separated by commas.
CSV files are really useful because they enable you to easily transfer data from one program to another.
We’ve been using Google Sheets so far but soon we’ll want to look at our data using data analysis software and CSV is compatible with it and most similar software.
Let's compare:
1. Data in our table:
2. Data in a csv file:
```CSV
Age Range,Weekly Hours of TV,Weekly Hours on Social Media,Favourite Color,Favourite Protein,Favourite Film Genre,Favourite Music Genre,Favourite Pet
30-40,5,2,Red,Fish,Romance,Indie,Cat
```
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Class Exercise: Create a CSV File From Google Sheets
Let's create our first CSV file from the dataset we built in Google Sheets to understand this type of file a bit more.
1. Open up the Google Sheet at [http://bit.ly/llc-ai-dataset](http://bit.ly/llc-ai-dataset). Be sure not to tamper with any of the data in the sheet.
1. Go to File > Download As > Comma Separated Values
1. Your file will begin downloading. Be sure to move your file into the **exercise-files** folder for this class.
Note: If needed, there is a backup CSV file with fake student data in the exercise-files folder already.
</script>
</section>
<section class="slide title" data-markdown>
<script type="text/template">
# DATAIKU
##Data Science Studio
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
#Dataiku
Today we will be using Dataiku to visualize and analyze our datasets. Dataiku is a platform that enables us to analyse our data and (later), run machine learning algorithms on our data, using a friendly Graphical User Interface(GUI). If we want, it also allows us to look behind the interface at the code itself to make changes.
(A special shout-out to the Dataiku team for supporting us in the installation of Dataiku for every student today!)
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Online Option
Go to the Get Started page on [dataiku.com](http://dataiku.com) and request the 14 day Free Trial. This will run in your Chrome or Firefox browser without any issues, as long as you’re connected to the internet.
Create an account and wait for the e-mail confirmation to come in. This could take 5-10 mins so feel free to take a break while you wait!
If you do not have internet, go to the next slide.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Offline Option for Mac: Installing Dataiku on an Apple Computer (15 mins)
1. Wait for the USB from the instructor with offline versions of dataiku and all the software you will need.
1. Open the DMG file and drop the application into the Applications folder.
1. Open up the application by finding it with Spotlight or clicking on its icon in the Applications folder.
1. You may need to authorize Dataiku to run on your OS. If so, Open System Preferences, go to "Security & Privacy" and authorize the application
DSS opens automatically in your browser (only Chrome and Firefox are supported). To reopen DSS, click on the DSS logo in the menubar or browse to [http://localhost:11200](localhost:11200)
1. Dataiku requires that you register for an account when you load up the app the first time. It's free. It may take 5-10 minutes to receive your account confirmation so this is a good time for a break.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Offline Option: Installing Dataiku on a Windows Computer (15 mins)
## For Windows computers, we’ll be using the Virtual Box option since Dataiku only runs locally on Linux and Apple computers. Get the USB from the instructor/mentor with the software you need.
1. Move the files from the Windows Installation folder to your computer.
1. Run the .exe file inside the folder to install Virtual Box.
1. Import the Data Science Studio Virtual Machine from the folder into Virtualbox (either by double-clicking or using File > Import Appliance)
1. The Appliance import wizard opens. You should not need to change any setting. Simply click on Import.
1. Once the virtual machine is imported, click on Start. The virtual machine boots. (It's possible that this will fail and you will get an error that says a variation of `VT-x is not available...`. This is because you need to activate Virtualization from the BIOS. Please ask a mentor for help with this.)
1. Once the Virtual Machine has finished booting, it displays a welcome banner with connection details.
1. You do not need to login into the virtual machine. Open your regular Chrome or Firefox browser (ie: not in the virtual machine). In that browser, open the URL that is displayed in the welcome banner, as "Data Science Studio interface". The DSS interface appears. This URL to open is often http://127.0.0.1:10000 (but not always, please check the welcome banner).
1. Dataiku requires that you register for an account when you load up the app the first time. It's free,but it may take you 10 minutes to get the confirmation e-mail if you haven't done this before. So go grab a coffee.
</script>
</section>
<section class="slide title" data-markdown style="background: url(framework/img/workshop/coffeebreak.gif") center 100% no-repeat;>
<script type="text/template">
# Installation Break
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# What goes in a database?
What kinds of information can we add to a database? If information is EVERYTHING, how do we store that in a computer in a way that it will understand?
There are established **types** of data that we can define in a database. For example:
- Integer (int, whole numbers)
- Float (float, numbers containing decimals)
- String (str, words, sentences, literal text)
- Boolean (True, False)
- Bytes and byte arrays, e.g. a jpeg image file.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Data Types
## Integers (Int)
These are whole numbers (no decimals). E.g. The number of attendees that came to your party = 35.
## Float(float)
These are decimal numbers. E.g. Usain Bolt’s 100m world-record-breaking running speed: 9.58 seconds.
## String (str)
Strings are words, sentences or text. E.g. Your favourite colour: “Red”
## Booleans
True or False. E.g. Are you attending your friend’s wedding? Yes(True).
## Bytes and byte arrays
This datatype is used for images such as jpegs.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Class Exercise: Our First Dataiku Project
Now that we have an account in Dataiku, let's import the csv file we created in Google Sheets into our first Dataiku project.
1. If you have the cloud edition, just go back to the dataiku website where you left off. If you installed the offline versions, Go to dataiku in your browser. (http://localhost:11200 on a mac and most likely http://127.0.0.1:10000 if you’re on Virtual Machine. )
1. Select the plus sign on the left hand corner to create a new project. 
1. Give your project a name. E.g. “My first Data project”
1. Import your first dataset. (this is the CSV file we created in Google Sheets) 
1. There are lots of ways to import datasets. We want to click on “Upload Files” under “Files”.
1. Select the CSV file you created earlier in Google Sheets. You’ll be able to preview it right away.
1. Click Create to finish the import. 
Here's an animation of the process of uploading your dataset:
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Exercise 2: Explore the Data (15 minutes)
Now it’s time to explore the data!
1. Once you click create, you should be sent directly the table of data with the exploration menu at the top. If not, click on the “Data sets” icon from the dashboard.
1. The dataset should open and enable you to explore using the exploration menu above it.
1. On the menu above it click on "Explore" 
1. In the Explore tab you can see all of your data in table format. In the top row you’ll see that every single column is first recognized as a string. This is because all CSV file values are stored as strings. Below the String you’ll see the actual data type DataIku will recognize for each of your columns. Are they all correct? Check against the data type definitions we talked about eariler.
1. Create a chart! If you go into the Charts tab you can start playing around with some data visualization! The default chart is a Histogram Bar Graph. You can change the type of graph you want by clicking on the graph-type dropdown. 
1. You’ll notice that some chart fields can only be filled with numerical columns. For instance, the Y column of your bar graph, must be a numerical data type for it to work. Play around with the columns and charts and see if there's any correlation between them! Have fun!
1. Bonus: Experiment with some of the geographical charts and your longitude and latitude columns! See if you can map everyone's favourite colour by location!
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Let's use Data to solve problems!
Now that we have explored our simple database, let's work with some real data.
**Problem:** The HR Department at IBM has a problem. They want to reduce the number of employees that leave the company.
**Solution:** In a dream world, the HR department would like to be able to predict which employees are most likely to leave the company so that they can stop them with new incentives.
How do we implement this solution? Machine learning.
<img src="framework/img/workshop/willis.gif" alt="Willis">
</script>
</section>
<section class="slide title" data-toc data-markdown>
<script type="text/template">
# Machine Learning
</script>
</section>
<section class="slide" data-markdown>f
<script type="text/template">
#What is Machine Learning?
Machine learning involves feeding data into a special kind of computer program, specifying a particular outcome, and having a machine develop its own algorithm to achieve the outcome. (Will Knight, July 20, 2017)
For instance, presented with a database of information about credit card transactions, such as date, time, merchant, merchant location, price, and whether the transaction was legitimate or fraudulent, a machine learning system learns patterns that are predictive of fraud.
- Traditionally data goes into the computer, the algorithm does its job **and gives out the result**.
- Learners turn this around: in goes the data and the desired result and **out comes the model** that turns one into the other.
- Learners are algorithms that make models **capture decision and action rules**.
- They learn knowledge / process that **underlie skills**
<hr>
### Resources
Definitions of Machine Learning and learning approaches to follow inspired in part by [Kathryn Hume](http://twitter.com/humekathryn)'s presentation at [TECH2025](http://tech2025.com)'s inaugural Toronto event "Explain It Like I'm 5: AI, ML, DL, and NLP" in July, 2017.
[AI Fight Club Could Help Save Us from a Future of Super-Smart Cyberattacks](https://www.technologyreview.com/s/608288/ai-fight-club-could-help-save-us-from-a-future-of-super-smart-cyberattacks/)
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Machine Learning is a core of AI
<img src="framework/img/workshop/ai-ml-dl-01.svg" alt="AI, ML, DL" width="50%" style="display: block; margin: 0 auto;">
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# ML Terms
## Instance (example, case, record):
A single object of the world from which a model will be learned, or on which a model will be used. Instances are described by feature vectors.
## Attribute (field, variable, feature)
A quantity describing an instance. An attribute has a domain defined by the attribute type, which denotes the values that can be taken by an attribute. A feature is the specification of an attribute and its value. For example, color is an attribute. ``Color is blue'' is a feature of an example.
The following domain types are common:
### Categorical
A finite number of discrete values. The type nominal denotes that there is no ordering between the values, such as last names and colors. The type ordinal denotes that there is an ordering, such as in an attribute taking on the values low, medium, or high.
### Continuous (quantitative)
Commonly, subset of real numbers, where there is a measurable difference between the possible values. Integers are usually treated as continuous in practical problems.
## Predictions:
model output that predicts the outcome by discovering patterns in the data
## Model:
A structure and corresponding interpretation that summarizes or partially summarizes a set of data, for description or prediction. Most inductive algorithms generate models that can then be used as classifiers, as regressors, as patterns for human consumption, and/or as input to subsequent stages of the KDD process.
## Classification Model/ Classifier: A mathematical function that maps from unlabeled instances to (discrete) classes.
For example, mapping each employee at IBM to the classes of leaving the job in the next 6 months or not.
## Regression Model/ Regressor:
A mapping from unlabeled instances to a real value, such as “amount” or “weight”.
<hr>
### References
[http://robotics.stanford.edu/~ronnyk/glossary.html](http://robotics.stanford.edu/~ronnyk/glossary.html)
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# What are the types of learning approaches?
- **Supervised**: Learn rules that map inputs to target outputs.
- **Unsupervised**: Learn to cluster and label similar inputs.
- **Deep** learn through a hierarchy of simple to complex concepts.
- **Reinforced**: Learn by continually interacting with an environment.
<img width="50%;" src="framework/img/workshop/urkel.gif" alt="Urkel">
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Supervised
<img width="60%" src="framework/img/workshop/supervised-learning.svg" alt="supervised learning ellen selfie">
Supervised learning happens when we feed the computer data and the outcome we want from that data. Then we let the computer find a correlation between the data and the outcome so that it may predict outcomes with new data without us providing the outcome.
For example:
1. We could give a computer a set of data about many different patients.
2. We could then tell the computer which of those patients have cancer and which ones don't.
3. The computer will find a correlation between the patients' data and whether or not they have cancer (that would be the model).
4. Then it will use this model which is no more than a mathematical function, to figure out which patients have cancer from the new dataset we give it without outcomes.
The **goal** is to minimize the error between the model's predictions and the actual outcomes.
In a **perfect world** we would minimize errors on all possible inputs.
In **reality**, we usually don't have enough inputs with corresponding outcomes to teach the computer well enough.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Deep Learning
<img class="medium-image" src="framework/img/workshop/puppy.jpg" alt="puppy">
<img class="medium-image" src="framework/img/workshop/kitten.jpg" alt="kitten">
Imagine we gave a computer millions of images of puppies and kittens. What characteristics about each might we pick to enable a computer to learn to identify future puppies and kittens? With images it's quite hard. This is where **deep learning** comes in.
<img src="framework/img/workshop/deeplearning.png" alt="kitten">
- Deep neural networks are a subtype of supervised learning.
- Deep networks are multilayer networks on top of each other where each layer corresponds to a different level of abstraction.
- There are two main reasons behind the popularity and the high impact of deep neural networks in various fields such as computer vision and speech recognition:
- The emergence of modern parallel computing architectures providing low-cost and fast computation for a large number of parameters of the deep networks.
- The availability of vast amounts of images, video, speech and text on the Internet providing sufficient data for training these networks.
- Each layer in the deep architecture provides a nonlinear information processing.
- The output is a parameterized function of the inputs and the output of each layer is the input for the higher layer.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Unsupervised Learning
<img src="framework/img/workshop/clustering-01.svg" alt="AI, ML, DL" width="80%" style="display: block;">
In unsupervised learning we give the computer unlabeled data (Input representations without their corresponding outcomes).
**The hacker in the haystack:** Imagine if we wanted a computer to figure out which internet users were hackers. We likely wouldn't have enough data for supervised learning so we would instead use unsupervised learning and the computer would find clusters of users, some who could be potential hackers based on online behaviour that is different than normal.
**Clustering:** group data based on some similarity metric.
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Reinforcement Learning
<img src="framework/img/workshop/reinforced-learning-01.svg" alt="AI, ML, DL" width="50%">
- **Training Experience:** the agent interacts with the environment and receives numerical reward signals
- **What to learn:** Best action for each state of environment- A behavior that maximize the reward in the long run
- **Challenges in RL:** Designing the problem domain: state representation, action choice, cost/reward signal
- **Example:** Alpha Go, Self Driving Cars
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
# Data Scientists WorkFlow
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-03.png" alt="Data Collection">
# Data Collection
Collecting and preparing the INPUT and the OUTCOME are the most important steps in solving problems using machine learning
In supervised ML, the model learns from the labeled instances that we feed into the model. The trained model can later be used to predict the outcome of data that you do not have the answer for.
## Each instance in your data must contain two elements:
- **Input** - The raw input that should be mapped to numerical values as variables/features
- **Outcome** – what you want your model to predict.
## Sources for labeled data:
- **Experienced data:** Historical data captured from processes with inputs and outcomes.
- **Human Annotation:** Providing input to human annotators and ask for the labels
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-03.png" alt="Data Collection">
# Feature Engineering
When we talk about features, we're talking about the columns of our database.
The performance of machine learning models significantly depends on the representation of the data
Each piece of information included in the representation of the input is called a feature.
**Example:** Representing a patient- doctor examines the patient, measures blood pressure, body temperature, takes MRI, CT scan, blood sugar test and provides this information to an AI system as feature inputs
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-03.png" alt="Data Collection">
# Exploring the Data
Columns are called **input variables** or **features** or **attributes**
- The outcome (which we are trying to predict) are called **output variables or targets**
- A row in the table is called **training example or instance**
- The whole table is called the **data set**.
</script>
</section>
<section class="slide" dat-toc data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-06.png" alt="Data Collection">
# Clean, Prepare and Manipulate Data
Pre-processing and cleaning data are important tasks that must be applied before using data to train an ML model
Why? To avoid "garbage in, garbage out"
## The problems with raw data
- Incomplete: having missing values for different attributes per records
- Noisy: non-correct values for some attributes, outliers
- Inconsistent
## Preprocessing Steps:
- **Data cleaning:** fill in missing values, smooth noisy data, identify or remove outliers, and resolve inconsistencies.
- **Data integration:** using multiple databases, data cubes, or files.
- **Data transformation:** normalization and aggregation.
- **Data reduction:** reducing the volume but producing the same or similar analytical results.
- **Data discretization:** part of data reduction, replacing numerical attributes with nominal ones.
<hr>
### Resources
http://www.cs.ccsu.edu/~markov/ccsu_courses/datamining-3.html, https://docs.microsoft.com/en-us/azure/machine-learning/machine-learning-data-science-prepare-data
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-03.png" alt="Data Collection">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-06.png" alt="Data Collection">
# Exercise 3: Getting and Cleaning the data (8 minutes)
Now, let's get back to our challenge of figuring out who might leave their job next at IBM so we can stop attrition in its tracks.
1. We’re going to create another project. Click on the plus sign on the left hand side of the DataIku dashboard. Let’s call it “Machine Learning Project”
1. We're then going to download the necessary dataset from [http://bit.ly/llc-ai-hr-dataset](http://bit.ly/llc-ai-hr-dataset) (There is a backup file of this already in the exercise-files folder)
1. Click on Import Dataset. From the project folder choose the IBM HR Analytics Employee Attrition & Performance file. It looks like a zip file but when you upload it it will become csv.
1. Click on the create button to finalize the import.
1. Ensure that all of the column names have been imported properly.
1. Ensure that all data-types have been identified correctly. **Tip:** There is a mistake on one of the columns! Here's a clue: 
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-05.png" alt="Data Collection">
# Train Machine Learning Models
## A supervised machine learning model is a function that takes an input and returns an output
F(X)=y, where F(X) is a “good predictor” for the value of y.
e.g. # of years at the job might be a good predictor of attrition (whether someone will stay or leave their job)
## Steps to solve a problem using ML:
1. Identify what the input-output pairs are.
2. Encode inputs and outputs.
3. Choose a class of ML models and experimentally find the best model
<hr>
### Resources
[A Non-Technical Introduction to Machine Learning](https://medium.com/safegraph/a-non-technical-introduction-to-machine-learning-b49fce202ae8)
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-04.png" alt="Data Collection">
# Test the Predictions of the Model
## How to evaluate the quality of a learned model?
- The data is split into two subsets:
1. A training and validation set used only to find the right function predicator
2. A test set is used to report the prediction error of the model
- These sets must be disjoint!
- Normally, we use 70% of the data as a Training and validation set and the rest as a test set
- In case of having smaller datasets, we use cross-validation
</script>
</section>
<section class="slide" data-markdown>
<script type="text/template">
<img class="flow-hexagon" src="framework/img/workshop/datascienceflow-05.png" alt="Data Collection">
# Class Exercise: Choosing our inputs and output
We need to choose which inputs we will use for our learning, and what the expected output is.
<ol>
<li>
<p>Back in our Machine Learning Project we left off at a cleaned table. Next we'll go to a new place: the Lab</p>
<p>Here we'll want to select "Prepare data and build models" in the Visual Analysis option. <img src="framework/img/workshop/ml-ex/Prepare-data-an-build-models.png" alt="buiod model"></p></li>
<li><p>Right click on the column that you want to predict and select "Create prediction model" This column will be the output. In this case we want to predict attrition.</p>
<img src="framework/img/workshop/ml-ex/create-prediction-model.png" alt="build models">
</li>
<li>Start with Decision Tree as it has the best interpretability</li>
<li>Press the green Train button. <img src="framework/img/workshop/ml-ex/lab-step-one.gif" alt="1-4"></li>
</ol>
</script>
</section>