13 problems
Let be prefix Kolmogorov complexity, let be the true parameter, and let denote the complexity difference used in the MDL setup. Strengthe…
In the convex composite problem with closed proper convex functions , suppose the Douglas–Rachford operator has a fixed point and is -smooth. Let denot…
Consider the convex composite problem … where and are closed proper convex functions. Assume that the Douglas–Rachford operator has a fixed point , that is …
Let and be operators satisfying the paper's standing Assumption, let be -cocoercive, and let for a closed proper convex function . Let…
Asymptotic-rate conjecture. The exact sublinear rates for regimes and correspond to
Finite-iteration tightness conjecture. The DCA rates corresponding to regimes , , and are tight for any number of iterations .
Regular-graph extension conjecture. The bound given in Theorem remains valid in the simple symmetric case, for example for regular graphs without transitivity.
Kalman's convergence-rate conjecture. For any stable and with ,
The convergence-rate conjecture. Under these assumptions, the distance between the convergent subsequence of finite-depth minimisers and its limit is . This c…
Let be the unit circle with Haar probability measure , and let be the multiplication operator … Cons…
Let . For each , let , where is a primitive root o…
Let be the number of training examples and suppose the feature matrix has entries in . Let denote the exponential loss and let…
Let be any weight vector and let . For exponential loss and AdaBoost's iterates, quadratic convergence conjecture. For every…