-
Notifications
You must be signed in to change notification settings - Fork 0
/
InnovationDays2018PromotionalVideoScript.txt
126 lines (77 loc) · 5.76 KB
/
InnovationDays2018PromotionalVideoScript.txt
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
Innovation Days 2018 Team FlexML Promotional Video Script
intro:
https://www.imdb.com/list/ls053181649/videoplayer/vi3416832537?ref_=vp_pl_1 from 3:53 to 4:10
Narrator (Texas/southern accent?):
- Let me tell you a story about a little app we have here at IMDB. Mhm well, yes, that IMDB, with the movies n stuff (bell tower explodes).
(pause)
Scene (User Value):
A) Show the IMDB web site
Go to Aquaman page (https://www.imdb.com/title/tt1477834/?pf_rd_m=A2FGELUUNOQJNL&pf_rd_p=f5d98e4a-da46-4e37-9028-a6e10c8f442e&pf_rd_r=Y89AN2BNQ9FTAB6PJG99&pf_rd_s=hero&pf_rd_t=15061&pf_rd_i=homepage&ref_=hm_hp_cap_pri_2)
Scroll down
Go to Justice league page (https://www.imdb.com/title/tt0974015/?ref_=nv_sr_2)
B) Scroll down to User Reviews
Create a review:
This movie is awesome! (See score going from 6.6 to 6.8)
(1 sec pause)
edit to:
This movie is awesomish! (See score going from 6.8 to 6.7)
Narrator:
We call it THE IMDB USER REVIEW MOVIE SCORE EXTRACTOR, URMSE for short, and is an integral part of the entry for every movie. And we have entries for all the movies! Ever! (IMDB site scene (A))
Once a user has creates a new review the USER REVIEW MOVIE SCORE EXTRACTOR would analyse the review and through the power of Machine Learning would extract an honest score based on the review's sentiment (create review scene (B)).
Scene (How it is done):
Show team at work
Narrator:
This is all fancy-shmancy but let me tell you how it is done and who are the men behind this app.
A (Bulent, Product manager at IMDB): We started off with Kinvey, cause it is reliable and secure and has all these great data conectors.
B (narrator): Don't forget that it is GDPR and HIIPA compliant!
A (Bulent, Product manager at IMDB): Oh, yes, it is GDPR and HIIPA compliant too. Which means that if we were to create an app for the Healthcare or the Financing industry it would still be our first choice! (wink)
B (narrator): Please stay on topic!
A (Bulent, Product manager at IMDB): - (awkward pause)
Scene (Data Consumption)
A (Georgi P - Georgi A with a moustache, developer at IMDB): So we have launched imdb.com some 27 years ago and we have aggregated a large quantity of reviews. (show entries in Kinvey)
We have analysed those based on the sentiment of the review and can build a MACHINE LEARNING model so that we can pretty surely asses the sentiment of new reviews automatically without the need for a human analyser.
B (narrator): Wow, that sounds great but also hard
A (Alex, developer at IMDB): It's not hard at all.
You see, Kinvey has now integrated an Intelligence as a Service capabilities (show Jupyter)
that runs on Kubernetes (show Kubernetes console)
We load the existing data on which we build our MACHINE LEARNING model here (show Load Data repl)
And here we create, compile and fit the model. (show Build model repl)
Because Kinvey's Intelligence as a Service has a Jupyter based online editor it's really easy to find the best algorithm and tweak the parameters so that you get the model with the highest score while preventing overfitting.
B (narrator): I see. But you'll need to expose an interface so you can get the results from outside.
A (Georgi A with a moustache, developer at IMDB): Oh yes, we have that too.
This is where we save the compiled model. (show the Save model repl)
And we export it with a REST API to the outside world using Kinvey's Flex services. (show the Export repl/Flex service in the Console API)
And it is really easy to test the Flex service using the Console API
For a negative review we expect a negative result (show negative review)
And we get one
Let's change the review to something positive (show positive review)
And we get a positive result.
Easy as that!
B (narrator): Wow Flex services! Those scale automatically, correct?
A (Bulent, product manager at IMDB): Yes, that was one more reason to choose Kinvey. Scalability was crucial as we have more than 80 million registered users. (show https://en.wikipedia.org/wiki/IMDb, blur everything but 83 million users)
B (narrator): I'm speachless! So you say you have built your app using
Kinvey which is reliable, secure has all the data connectors, is GDPR and HIIPA compliant (show Bulent's As Is Kinvey schema)
and Kinvey now has this Intelligence as a Service module (show Bulent's Kinvey with Intelligence as service schema)
which makes it really easy to find the best MACHINE LEARNING model for your problem and tweak the parameters as it has a Jupyter-based online editor! (show list with advantages from Bulent's slides)
And you export the model with just a few clicks and the interface scales out of the box as it runs on a Flex service.
A (Georgi P - Georgi A with a moustache, developer at IMDB): (awkward pause) Yes (awkward pause)
Scene (Variant 1):
Black screen, large font:
Fun fact #1:
The IMDB USER REVIEW MOVIE SCORE EXTRACTOR does not use Kinvey.
(2 secs pause) Yet.
(2 secs pause) In fact, it does exist.
(2 secs pause) Yet.
Fun fact #2:
No one of the people appearing in this video works or has ever worked for IMDB. (Start typing Yet. Stop at Ye. Delete. Type Fullstop in words)
(2 secs pause) They all work for Progress Software.
Type 'Kinvey, now with Intelligence as a Service capabilities' (Alex, Georgi and Bulent sing Kinveeey. by Progress)
END
Scene (Variant 2):
Go to Aquaman page (https://www.imdb.com/title/tt1477834/?pf_rd_m=A2FGELUUNOQJNL&pf_rd_p=f5d98e4a-da46-4e37-9028-a6e10c8f442e&pf_rd_r=Y89AN2BNQ9FTAB6PJG99&pf_rd_s=hero&pf_rd_t=15061&pf_rd_i=homepage&ref_=hm_hp_cap_pri_2)
Go to 'Did you know?' section
- First entry reads:
IMDB Review Aggregator does not use Kinvey. See more(link) // Blur everything else.
- Click on See more link (land on https://www.imdb.com/title/tt1477834/trivia?ref_=tt_trv_trv)
- First entry reads:
IMDB Review Aggregator does not use Kinvey. Yet // Blur everything else.