Archive | September 2021

” Artificial intelligence : 101 – Netflix User Case”

There is a general misconception in several quarters that Artificial Intelligence (AI) is primarily a work medium for technologically qualified personnel and data science experts. The answer is, however, a firm “No”.

AI actually lends itself to any discerning individual irrespective of qualifications but interested in : (a) making the world a better place ; (b) making people’s standard of living better than what is being enjoyed now ; and (c) solving problems in an interdisciplinary way utilising simple technologies and tools available to qualify and quantify information gathered from various sources. With short term courses one will do well to get a handle on tools and techniques of AI.

History records a number of studies and instances where the work carried out and inferences derived could be a part of AI studies without having been clearly designated as such.

My Study in brief :

A three year study I was engaged with beginning 1980 explored the incidences cum impact of Occupational Stress Management in industry (specially manufacturing sector) and the availability of limited scientific facilities for in-house monitoring, collecting, analysing, and interpreting relevant data in an attempt to assess the situation and to provide preventive solutions.

I had the luxury of sophisticated instrumentation in the Premier Institute I worked comparable to any leading institution in the world. ( This is quite akin to Computer facilities and Cloud for today’s AI ).

Coupled with scientific manpower it was possible to collect enormous amount of data – in that period data on 10000 individuals was supposed to be – in real life, work centered, human centered conditions and parameters from a range of industries as steel plants, textile Mills, manufacturing plants, assembly units, and power plants. (This is quite akin to Big Data in AI ).

The data collected was subjected to detailed analysis, and programs were developed to the interests of industries in assessing the risk status of individuals to health. ( This is quite akin to the algorithms written to analyse and create a graph /nomogram for prediction ).

Result :

  1. 1. Simple monitoring of stress parameters with instruments for the purpose which can be done by in-house personnel ( very economical and highly reliable ).
  2. 2. Small team with only three members to take relevant measurements saving time, resources, personnel, and effort.
  3. 3. Simple graphs, nomogram, and equations to help assess health risks and predict future to an extent.
  4. It did not occur to me to term that as an AI Study at that time since the term was not popular making its way in scientific investigations during that period.
  5. In retrospect when we go back to the details of the Study, the first thought that comes to my mind is that it could well be the prelude to AI Studies in that field from a developing nation. Much less sophisticated, much less tech induced, much less algorithmic, and much less revolutionary. But, it sure, was “data driven” and engaging. The study would easily pass for an AI part in today’s context.
  6. User Case : ” Netflix” : This considers the period prior to 2019 with special emphasis on 2018.
  7. Netflix is an international company that relies a lot on Content and Graphics catering to different mediums while adding value to its streaming service.
  8. Good value and engaging content are the chief driving factors of the company’s progressive actions. It centers on its principal focus on the process of ” delivering entertainment reimagined with tech”. Customers enjoy and get fully influenced to watch their contents of the video.
  9. Their movie titles are “stylistic” ;
  10. Their posters are “special” ; and
  11. Their images are “titillating”.
  12. These were the result of AI in their process of ideation and depiction. It was possible because of AI.
  13. To secure images for exciting and meaningful display to the viewers, they focused on three elements – focal point ; text placement ; and image clustering – and made effective variations with suitable algorithms.
  14. Focal Point : The chief area of the image like face, body, legs that the user is expected to view primarily is the focus here. (A focal point algorithm is used here).
  15. Text Placement : Deciding which is the correct place in a scene to put a text like, say, subtitles in a frame is the focus here. ( A text detection algorithm is used here).
  16. Image Clustering : Using different images with varying degrees of image cropping, resizing, colour correction, and title treatment utilising different characteristic qualities to make images more effective for different users globally is the focus here. ( Algorithms to determine a similarity in feel among groups of images is used here).
  17. Netflix has been successful in enhancing the enthusiasm of the viewers with these changes.
  18. ” Grouping the images and maintaining the image collection of more than 2 million has been the challenging task for Netflix with lofty hopes and aims down the road ” – admitted Netflix.
  19. Netflix also stood to gain profitably through the deals forged like : (a) Grey’s anatomy creator Shonda Rhymes ( $ 100 million ) ; (b) Ryan Murphy ( $ 300 million ) ; Netflix’s original projects already planned ( about $ 700 million) ; and (d) the highly valued ‘high 8 figure deal’ struck in May 2018 with Barack and Michelle Obama ( expected to be more than $ 65 million) which together can last several years.
  20. Reports then mentioned that Obamas were likely to create programs covering : ” Scripted series ; unscripted series ; docu-series ; documentaries ; and features.”
  21. Netflix had earned its mark with more than 150 million subscribers covering more than 190 countries by 2018- since unveiling its streaming services in 2007.
  22. Discuss :
  23. 1. What is the status of its competitive edge over rivals now? Specially after COVID-19?
  24. 2. How long can this revenue stream be sustained? Will it be impeded by any unforseen glitches or levels of enthusiasm of users?
  25. 3. Can this medium through AI be used to clear up the conflict and confrontation between nation(s) by focusing on the similarities of qualities of human beings in general?
  26. ” Dieu avec nous “
  27. Tuesday, September 7, 2021 – 9.59 p.m. (IST)
  28. Tidbit : ” Netflix’s ability to pass along price increases to customers without walking away is the sign of a great company “.