SupplementaryMaterial

Neural Information Processing Systems 

Note that this setting is the sameastheonedescribed in[14]. In this section, we illustrate the empirical convergence performance of the proposed algorithms. Since Problem(18) contains nonconvex terms in both its objective function and constraints, the convergence ofAlgorithm 3iscarefully analyzed andestablished through Theorem 2inthemain manuscript. Proof To prove Theorem 2, we first establish the boundedness of the sequence Ll,Bl,Yl generatedbyAlgorithm3inLemma5,andthemonotonicityofLρ Ll,Bl,Yl inLemma6. Yl 1, and Rl 1 contains the corresponding eigenvectors.

Similar Docs  Excel Report  more

TitleSimilaritySource
None found