5 ms·
Intuitive crutches for higher dimensional thinking
- saeranv 6y agoTerence Tao's answer here fascinates me, even though I only partially understand his response. > For instance, one can view a high-dimensional vector space as a state space for a system with many degrees of freedom. A megapixel image, for instance, is a point in a million-dimensional vector space; by varying the image, one can explore the space, and various subsets of this space correspond to various classes of images. I don't understand this. How is he thinking of a 2D data structure as a n-dim vector space? The only way I can think of mapping n-dims to 2d is maybe taking two n-dimensional vectors and visualizing the 2D plane between them, or projecting n-dimensions onto 2D space... > For instance, the fact that most of the mass of a unit ball in high dimensions lurks near the boundary of the ball can be interpreted as a manifestation of the law of large numbers, using the interpretation of a high-dimensional vector space as the state space for a large number of trials of a random variable. I think this came up recently in another hnews thread. The way I interpret this statement is that you can interpret the probability of events occurring in a portion/segment of a n-dimensional space as permutations of those segmented events in each dimension: 1/k^n where: k = number of segments n = number of dimensions So for example, if you have a 1D line, and you want to find the probability of ending up in a third of the line, the formula is: 1/3^1 = 0.333 If we think extend this to a plane, and look at the probability of ending up in the third of 2 dimensions: 1/3^2 = 1/9 Would be curious if someone has another interpretation.