An Artificial Intelligence Company Inspired by Brains
Press Release
GENISAMA Announces
Consumer Products toward Strong AI
​     GENISAMA LLC is a startup based in East Lansing, Michigan, USA.  It is a spin-off from AI research at Michigan State University.  Nov. 3, 2017, it announced several consumer products that will attract those who like to keep on top of current technology for mobile 3D video recording and mobile viewing. “They are also our first products toward strong AI”, company founder Dr. Juyang Weng said.
     3D Camera and Machine learner, 3DCam for short, is a set of custom-made gadgets that work together to expand the recording and viewing power of smartphones.  You can use the set's binocular cameras to record 3D video and take 3D photos.  The synchronization of the two binocular cameras makes the resulting 3D video very stable to watch, unlike some unsynchronized 3D cameras for mobile phones. The provided 3D goggle is used to view 3D contents on smartphones, tablets, laptop computers, and 3D game consoles, and is more comfortable than anaglyph and methods that mask half of the pixels for each eye. Its other function is machine learning, based on MSU patented technology Developmental Networks (DN).  This represents a newly developed avenue for consumer enjoyment:  You can train the product for your own personal wishes and likings.
     3DTube is a video-based Internet platform emphasizing 3D contents.  You can share with your friends 3D movies of your wedding, vacation, or your child's first steps.  The growing computing power of mobile phones will make 3D video and 3D photos increasingly common in our lives. For example, facial features are not changed or lost in 3D photos as they are in 2D photos.
     AOS is a new kind of operating system.  Its core is the DN, which can auto-program for general purposes. Companies and individuals do not need to understand DN in order to use the AOS to train their own AI systems. AOS is different from traditional AI that requires a human to hand-craft programs for any specific task plus task-specific machine learning.  AOS auto-programs: The machine learns in a way similar to how parents and teachers physically interact with children, from simple to complex.  “This relieves humans from handcrafting AI rules, as such rules are too many and too muddy for any human to program,” Weng explains.  For more detail, email [email protected]