# Carl Feynman

 620 Reputation 1912 views

## Registered User

 Name Carl Feynman Member for 2 years Seen yesterday Website Location Boston, Massachusetts, USA Age
 May21 comment Are there some potentially good methods of comparing the variances of two distribution function?This problem is underspecified. You haven't told us anything about the distributions except for the properties they don't have, i.e. that they are different and that they don't have analytic variances. At this point, we don't know that they have variances at all, or even means, or are even computable. What you need to do is to tell us everything you do know about the distributions. For example, do they have compact support? Is there some way to sample from the distributions? What space is the random variable over? May16 answered Constrained minimum maximal distance. May9 comment bipartite graph coloringFor the case where all the V1 vertices have two neighbors, this reduces to vertex-coloring an arbitrary graph, with all its NP-complete goodness. The structure you are describing as "a bipartite graph with..." is more commonly called a hypergraph. V2 is the vertices of the hypergraph and V1 is the hyperedges. May6 comment Expected edit distanceYou can tighten the bound slightly. The number of sequences with the given number of insertions, deletions and substitutions is less than you give because it doen't make sense to have an insertion next to a deletion. Apr25 awarded ● Editor Apr25 revised iteratively (approximately) solving a sum of exponentialsdeleted 2 characters in body Apr25 answered iteratively (approximately) solving a sum of exponentials Apr23 comment Constraint optimization problem for any dimensionality $n>1$.Whoops, I see that my comment above is incorrect; Mr Kallus didn't give the amplitude either. Apr23 comment Constraint optimization problem for any dimensionality $n>1$.The period is $n$. I failed to correctly specify the amplitude, but Yoav Kallus kindly supplied it for us. I was going to write up a more precise answer, but Mr Kallus beat me to it. Apr23 comment Constraint optimization problem for any dimensionality $n>1$.That solution is a sine wave. That suggests the general solution is of the form $a_i=sin(2 \pi i / n)$. Apr17 answered In Szemerédi’s Regularity Lemma, how many blocks are in the partition? Mar27 comment Machin-like formulas for logarithmsYou mention that you are looking for numbers "sandwiched between smooth numbers." What is the importance of smooth numbers for this problem? Mar21 awarded ● Yearling Mar21 answered Is there a database somewhere for sharing translations of mathematical works? (Or, is anyone interested in a translation of a letter Weil wrote to de Rham in 1946?) Feb16 comment What machine learning algorithm is appropriate for predicting one time-series from another?I imagine that both the teacher and the student are more likely to look at interesting things then they are to look at boring things. Then areas the teacher looks at more are more likely to be looked at by the student, but not necessarily at the same time. This is impossible to capture in a Markov model because it is time-independent. On the other hand, it is space-dependent, so density estimation on the complete teacher time sequence could capture this dependence. Feb10 answered Find a convex hull that contains given points? Feb4 comment Solving a linear recurrence relation with variable coefficients.One can produce any sequence whatever of (nonzero) vectors for $X[n]$, by choosing $f$ and $g$ correctly. And once $X$ becomes zero, it will remain zero thereafter. In the absence of any knowledge of $f$ and $g$, there is nothing more to be said. Perhaps if you explained what your actual problem is, someone might be able to help. I don't know why you're trying to keep $f$ and $g$ secret from us-- you must have some theory that constrains them. Feb3 comment What is the cardinality of the family of unlabelled bipartite graphs on n vertices?In the limit for large $n$, the average graph has no automorphisms, so the number of unlabeled graphs is simply the number of labeled graphs divided by the number of possible labelings, i.e. $n!$. Jan12 answered equation for bowling ball on a trampoline Dec27 comment Integration in the surreal numbersThis thesis says in the conclusion that 'A “natural” definition of integral that gives the right answer for the exponential function ... is the main open problem of this thesis.' So that reference doesn't answer the question. Dec23 comment Gradient Descent for Primal Kernel SVM with Soft-Margin(Hinge) LossYou're not going to get nice results trying to find the derivative of a function like $max$, which is not everywhere differentiable. I see two ways out. First, use a conditional in your expression, like equation (12) in your reference. Second, use a differentiable approximation to $max$, like softmax (read the last section of the Wikipedia article on softmax for the approximation I'm thinking about.) Dec10 answered Estimate number of distinct items Dec6 comment Sampling without replacement: probability for total successes from successes in sample?This is a job for Bayes' theorem. Dec5 comment softmax activation function with infinite support ?And once again, I ask, how are the $q_i$ determined? If you know they're produced by some stochastic process, there is some hope of being able to provide a value for their softmax. But if you don't know that process, there's no hope. Dec4 awarded ● Commentator Dec4 comment softmax activation function with infinite support ?I'd be happy to answer this question if you clarified what you were asking. What do you know about the $q_i$? How are they determined? If you don't have any information about the $q_i$, there is no way to provide an estimate. Dec2 comment distinguishing random orthogonal matrix from Gaussian random matrix A nonzero total variation distance doesn't necessarily mean that the distributions can be told apart from samples. The place where the distributions differ substantially might take up a vanishingly small fraction of the probability weight of the distributions.