Bernstein algebras that are algebraic and the Kurosh problem (Q2109053)

The authors belong to Moscow algebraic school. They study the class of so called \textit{Bernstein algebras} that are algebraic, in the sense that each element generates a finite-dimensional subalgebra. Every Bernstein algebra has a maximal algebraic ideal, and the quotient algebra is a zero-multiplication algebra. Several equivalent conditions for a Bernstein algebra to be algebraic are given. In particular, known characterizations of Bernstein train algebras in terms of nilpotency are generalized to the case of \textit{locally train algebras}. The authors show that if a Banach Bernstein algebra is algebraic (respectively, locally train), then it is of bounded degree (respectively, train). Then authors investigate the \textit{Kurosh problem} for Bernstein algebras: \textit{whether a finitely generated Bernstein algebra which is algebraic of bounded degree is finite-dimensional}. This problem turns out to have a closed link with a question about associative algebras. In particular, when the bar-ideal is nil, the Kurosh problem asks \textit{whether a finitely generated Bernstein train algebra is finite-dimensional}. The authors prove that the answer is positive for some specific cases and for low degrees, and construct counter-examples in the general case. Now let me describe why this nice paper has a special interest for me. It is known that the Jacobian conjecture is equivalent to \textit{cubic homogeneous} case i.e. for mappings \[x_i\to x_i+f_i(\vec{x})\] were \(f_i\) are cubic homogeneous forms. This statement is equivalent of weak nilpotence of Engel algebras with symmetric triple multiplication operation with Capelli identities. From other hand Jacobian conjecture for quadratic mapping is known and it implies weak nilpotence of Engel algebras with symmetric binary multiplication operation with Capelli identities. see [\textit{I. Cheltsov} (ed.) et al., Automorphisms in birational and affine geometry. Papers based on the presentations at the conference, Levico Terme, Italy, October 29 -- November 3, 2012. Cham: Springer (2014; Zbl 1291.14005)] for details. Hence it seems interesting to provide direct proof of weak nilpotence for Engel algebras with by nary operation. The authors show that the generalized Jacobian conjecture for quadratic mappings holds for Bernstein algebras in a ring theoretical way for characteristics different from 2.