Houdini Automated Visual Effects, Animation And Modeling

   19556   41   4
User Avatar
Member
148 posts
Joined: March 2016
Offline
The algorithms I have developed are getting better through machine learning experience. Lets take this discussion further.

Recently I have had discussions with several founders of machine learning who I have known since the late 1970s regarding the current trajectory of artificial intelligence. As an example, Google created software that won matches against a GO master in 2015 for the first time ever (https://en.wikipedia.org/wiki/AlphaGo). [en.wikipedia.org] GO is a complicated game associated with what might be called Category IV machine reasoning defined here ( http://fvg.com/ai-categories.html [fvg.com] ) and in the table below. Many thought leader researchers in Artificial Intelligence believe that what they call a singularity will take place in 20 years (well within our lifetimes) where machines will have full Category IV human intelligence.

In this context, here is a thought experiment question based upon this forum discussion thread on “What defines art”:

Let us say there is a machine, which is defined by algorithms. These algorithms are extremely complex and implement full Category IV machine reasoning (defined below in the table), but they are still algorithms, i.e. complex configurations of what reduce to “if-than-else” statements. The Category IV machine in this thought experiment is assigned a task to create a novel character for a shot based upon some parameter constraints. Without access to outside references from which to copy, it produces the Mona Lisa. There is no intent, only complex algorithms responding to a set of parametric constraints. Is this art? More specifically computers are being used to do just this: http://qz.com/495614/computers-can-now-paint-like-van-gogh-and-picasso/ [qz.com]

Here is a table I created several years ago for and at the request of Dr Arno Penzias (https://en.wikipedia.org/wiki/Arno_Allan_Penzias) [en.wikipedia.org] delineating algorithms with Category I through IV machine reasoning, for reference. Penzias, former president of Bell Labs Research, agreed with the delineation in the table:



Here are Category III algorithms currently producing these Houdini animated models automatically from picture references:

Albert Einstein: https://en.wikipedia.org/wiki/Albert_Einstein [en.wikipedia.org]



Jim Morrison: https://en.wikipedia.org/wiki/Jim_Morrison [en.wikipedia.org]



John Wayne: https://en.wikipedia.org/wiki/John_Wayne [en.wikipedia.org]




Edited by LARSX - Sept. 10, 2016 16:51:19
Lars Wood
Future Vision Guide
Advanced Research And Development
User Avatar
Staff
2493 posts
Joined: Sept. 2007
Offline
Sorry folks. As entertaining as this thread is, I'm locking it down.
Chris McSpurren
Senior Quality Assurance Specialist
SideFX
  • Quick Links