[HN Gopher] Exploring Neural Graphics Primitives
       ___________________________________________________________________
        
       Exploring Neural Graphics Primitives
        
       Author : TheNumbat
       Score  : 82 points
       Date   : 2023-03-30 15:23 UTC (7 hours ago)
        
 (HTM) web link (thenumb.at)
 (TXT) w3m dump (thenumb.at)
        
       | bob1029 wrote:
       | I wonder if some basic pre-processing of the image would help.
       | 
       | Turning RGB into YCbCr is a pretty essential step for achieving
       | efficiency gain in techniques like JPEG. Quantization works a lot
       | better when we can do the luminance and chrominance separately.
       | 
       | If you want the AI to "see" more like a human, converting into
       | this kind of color space is an important first step IMO.
        
         | johntb86 wrote:
         | RGB<->YCbCr is a linear operation, so a deep enough network
         | shouldn't have trouble learning to do the transform if it's
         | useful. Gamma correction might be important, though.
        
           | dTal wrote:
           | Sure, but surely you'd rather not? I imagine you want the
           | network encoding the details of the specific image, not
           | wasting weights to learn about easy wins that apply to
           | _every_ image. You could burn a whole layer converting to
           | YCbCr, and the net in the article only had 3.
        
             | toxik wrote:
             | You don't know if that is actually the case, the whole
             | appeal, originally, is that you don't need this kind of
             | handcrafting of features. Now, clearly, people preprocess
             | their data all the time so that point is kind of moot.
             | 
             | However, if YCbCr was meaningfully better for the network,
             | I think it would be pretty well-known as a preprocessing
             | step. Also, the actual sensor data is RGB.
        
               | limbicsystem wrote:
               | But the human visual system is insensitive to high
               | resolution yb and red/green information so you can
               | essentially down sample those layers to almost nothing
               | before you even start. That's a fundamental trick which
               | is also used in another way by jpeg but presumably not by
               | this RGB algorithm.
        
               | bob1029 wrote:
               | If you extend the JPEG analogy all the way and consider
               | the 4:2:0 subsampling mode, then you could reduce your
               | input parameter counts by exactly 50% by consuming a pre-
               | converted image.
        
       | tehsauce wrote:
       | In the last section on hash tables, I was wondering how on earth
       | they make hash tables work well on the GPU.
       | 
       | The answer:
       | 
       | "What do we do about hash collisions? Nothing--the training
       | process will automatically find an encoding that is robust to
       | collisions."
       | 
       | Amazing. Makes one wonder what other classic data structures gain
       | new properties when ML is mixed in.
        
         | _a_a_a_ wrote:
         | There was some stuff quite recently about using AI (neural
         | networks I think) to help optimise lookups in SQL indexes[1].
         | I'm afraid I can't find it now.
         | 
         | [1] That caused some rebuttal papers to be published saying
         | "not so fast"
        
       | tehsauce wrote:
       | A great summary of this line of research! I think it should have
       | the (2022) label.
        
       ___________________________________________________________________
       (page generated 2023-03-30 23:01 UTC)