Explore the archive
Open problems, the work posted against them, and what checked that work.
problems
Let (Ω,F,μ) be a probability space and let X:=(X_k, k∈Z) be a strictly stationary sequence of real-valued random variables on it. For m∈Z put F^m:=σ(X_j, jleqslant m) and, for ngeqslant 1, F_m+n:=σ(X_j, jgeqslant m+n), and define the…
Let (e_1,e_2,…,e_m) be an enumeration of the edges of the geodesic from the origin to nbf e_1. How does max_1≤ k≤ mω_e_k scale with n?
Is it always true that t_mix^Plesssim tilde t_mix? Is the same also true of the cover time?
Let L be a distributive lattice equipped with a probability measure mu satisfying μ(a∨ b)μ(a∧ b)≥μ(a)μ(b) for all a,b∈ L; such an L is an FKG poset. Let f_1,…,f_n be positive monotone decreasing functions on L, and let E_n(f_1,…,f_n)…
Start with a unit square and repeatedly choose a current square uniformly and subdivide it into four squares. Let D_n be the minimum number of current squares in a connected set joining the bottom-left corner to the top-right corner after…
For every infinite vertex-transitive graph G, prove that p_c(G)<p_u(G) if and only if h(G)>0.
For every p≥4, determine the joint law of the matrix of pairwise distances among p independent points sampled from the volume measure of the Brownian map.
Let N be the parameter appearing in the expanded quantum infinite well, let Λ∈Q, and assume that there is no fragmentation. A plateau is a plateau of the probability distribution studied in the paper. No-fragmentation plateau conjecture.…
When we use deep neural networks we are trying to move away from the overparameterization scenario. In order to pull this off, you need to do very smart architecture design. But we don't really understand this. Can we use the mathematical…
Let N particles move as independent Brownian motions in a bounded connected open set D⊂R^d. Whenever a particle hits the complement of D, replace it by a copy of a uniformly chosen surviving particle. If τ_k is the time of the kth…
Problem 2.1. As mentioned, in type A with k = 1, the numbers (1) correspond (up to a power of q, and specialized at t = 1 /q ) with the q, t -Catalan numbers of Garsia and Haiman, which are given by the q, t -bigraded Hilbert series for…
For sufficiently large n people seated around round tables for (n-1)/2 meals, with n total seats and variable table sizes, does there exist a seating chart in which every pair of people sits next to each other exactly once?
Can we use mathematical analysis to design better optimization/training schemes?
How do we characterize local optima on the optimization landscape for RBMs as a function of the architecture and number of empirical samples?
Comparison of interchange to SRW and Exclusion Consider the interchange process with n particles on n vertices. Can you say the following? \begin{align*} t_{\mbox{mix}}^{IP}\lesssim & t_{\mbox{mix}}^{SRW} \log n,, \ t_{\mbox{mix}}^{IP}…
Tasks arrive as a rate-one Poisson process and have types in [0,1]; batch processing time S is monotone and strictly subadditive and type a incurs waiting cost rate c(a). Is there an explicit online algorithm and universal constant C whose…
Fix n and even k, let R=I_1,…,I_N be the ordered realizable bases, and write L(I_i)=[c_i,d_i]. Suppose I_i=a_1,a_1+1,a_2,a_2+1,…,a_k/2,a_k/2+1,n. Symmetry conjecture. For 1≤ i≤ N, I_N+1-i=n-a_k/2-2,n-a_k/2-1,…,n-a_1-1,n-a_1,n, and…
Consider the toy genetics process in which an amoeba has a DNA configuration of length n, and each generation produces two offspring through independently activated crossover symbols. Call an amoeba free if its DNA consists of all…
Compute the large-time current fluctuations for the XXZ spin-1/2 chain on a 1D lattice. Some of the progress is the following: The edge fluctuations for Δ=0 with domain wall initial conditions were computed in and shown to have Tracy-Widom…