Searcharxiv⌕ Search

arXiv subjects

Ekaterina Shemyakova

Publications and source records attributed to Ekaterina Shemyakova.

At least 19 recordsLinked to original sources

Berezinian expansion and super exterior powers

In supergeometry, the classical identification of top-degree differential forms with volume forms suitable for integration breaks down. To address this issue, several generalized notions of differential forms have been introduced, including integral forms and pseudo-differential forms of Bernstein-Leites and $r|s$-forms of Voronov-Zorich. The Baranov-Schwarz transformation maps pseudo-differential forms to $r|s$-forms of type II, and, on a supermanifold of dimension $n|m$, integral $r$-forms are isomorphic to $r|m$-forms. However, an explicit construction of $r|s$-forms for arbitrary $s$ has remained elusive. First, we observe that this problem can be related with expansions of $\mathrm{Ber}(E+zA)$ for a linear operator $A$ on a superspace $V$. It was already known that the coefficients of the expansions near zero and near infinity give the supertraces of the representations $Λ^{r|s}(A)$ in the two extreme cases $s=0$ and $s=m$, respectively. In this paper we show that the intermediate expansions of $\mathrm{Ber}(E+zA)$, taken in the annular regions between consecutive poles, encode the supertraces of representations on explicitly constructed vector spaces that model $Λ^{r|s}(V)$ for $0<s<m$. We introduce a formal analogue of the Baranov-Schwarz transformation and show that it relates these spaces to Voronov-Zorich $r|s$-forms. As a separate remark, we establish that $1|1$-forms at a point of a supermanifold of dimension $n|m$ can be realized as closed differential forms on the super projective space $\mathbb{P}^{m-1|n}$.

math.DG↗

A quantum anchor for higher Koszul brackets

It is well known that the chain map between the de Rham and Poisson complexes on a Poisson manifold also maps the Koszul bracket of differential forms into the Schouten bracket of multivector fields. In the generalized case of a $P_\infty$-structure, where a Poisson bivector $P$ is replaced by an arbitrary even multivector obeying $[[P,P]]=0$, an analog of the chain map and an $L_\infty$-morphism from the higher Koszul brackets into the Schouten bracket are also known; however, they differ significantly in nature. In the present paper, we address the problem of quantizing this picture. In particular, we show that the $L_\infty$-morphism is quantized into a single linear operator, which is a formal Fourier integral operator. This paper employs Voronov's thick morphism technique and quantum Mackenzie-Xu transformations in the framework of $L_\infty$-algebroids.

math-ph↗

The weighted projective superspace with weights $+1, -1$ and an analog of the Fubini--Study form

As a by-product of our work on super Plücker embedding, we came to the notion of a weighted projective superspace $P_{+1,-1}(V\oplus W)$ with weights $+1,-1$. The construction is not in itself super and makes sense in ordinary (purely even) framework. Unlike the familiar weighted projective spaces with positive weights, the (super)space $P_{+1,-1}(V\oplus W)$ is a smooth (super)manifold. We describe its structure and show that it possesses an analog of the Fubini--Study form.

math.DG↗

On super cluster algebras based on super Plücker and super Ptolemy relations

We study super cluster algebra structure arising in examples provided by super Plücker and super Ptolemy relations. We develop the super cluster structure of the super Grassmannians $\Gr_{2|0}(n|1)$ for arbitrary $n$, which was indicated earlier in our joint work with Th. Voronov. For the super Ptolemy relation for the decorated super Teichmüller space of Penner-Zeitlin, we show how by a change of variables it can be transformed into the classical Ptolemy relation with the new even variables decoupled from odd variables. We also analyze super Plücker relations for general super Grassmannians and obtain a new simple form of the relations for $\Gr_{r|1}(n|1)$. To this end, we establish properties of Berezinians of certain type matrices (which we call ``wrong'').

math-ph↗

On super Plücker embedding and cluster algebras

We define a super analog of the classical Plücker embedding of the Grassmannian into a projective space. One of the difficulties of the problem is rooted in the fact that super exterior powers $Λ^{r|s}(V)$ are not a simple generalization from the completely even case (this works only for $r|0$ when it is possible to use $Λ^r(V)$). To construct the embedding we need to non-trivially combine a super vector space $V$ and its parity-reversion $ΠV$. Our "super Plücker map" takes the Grassmann supermanifold $G_{r|s}(V)$ to a "weighted projective space" $P\left(Λ^{r|s}(V)\oplus Λ^{s|r}(ΠV)\right)$ with weights $+1,-1$. A simpler map $G_{r|0}(V)\to P(Λ^r(V))$ works for the case $s=0$. We construct a super analog of Plücker coordinates, prove that our map is an embedding, and obtain "super Plücker relations". We analyze another type of relations (due to Khudaverdian) and show their equivalence with the super Plücker relations for $r|s=2|0$. We discuss application to much sought-after super cluster algebras and construct a super cluster structure for $G_2(\mathbb{R}^{4|1})$ and $G_2(\mathbb{R}^{5|1})$.

math.DG↗

Growing an architecture for a neural network

We propose a new kind of automatic architecture search algorithm. The algorithm alternates pruning connections and adding neurons, and it is not restricted to layered architectures only. Here architecture is an arbitrary oriented graph with some weights (along with some biases and an activation function), so there may be no layered structure in such a network. The algorithm minimizes the complexity of staying within a given error. We demonstrate our algorithm on the brightness prediction problem of the next point through the previous points on an image. Our second test problem is the approximation of the bivariate function defining the brightness of a black and white image. Our optimized networks significantly outperform the standard solution for neural network architectures in both cases.

cs.LG↗

On a Batalin--Vilkovisky operator generating higher Koszul brackets on differential forms

We introduce a formal $\hbar$-differential operator $Δ$ that generates higher Koszul brackets on the algebra of (pseudo)differential forms on a $P_{\infty}$-manifold. Such an operator was first mentioned by Khudaverdian and Voronov in \texttt{arXiv:1808.10049}. (This operator is an analogue of the Koszul--Brylinski boundary operator $\partial_P$ which defines Poisson homology for an ordinary Poisson structure.) Here we introduce $Δ=Δ_P$ by a different method and establish its properties. We show that this BV type operator generating higher Koszul brackets can be included in a one-parameter family of BV type formal $\hbar$-differential operators, which can be understood as a quantization of the cotangent $L_{\infty}$-bialgebroid. We obtain symmetric description on both $ΠTM$ and $ΠT^*M$. For the purpose of the above, we develop in detail a theory of formal $\hbar$-differential operators and also of operators acting on densities on dual vector bundles. In particular, we have a statement about operators that can be seen as a quantization of the Mackenzie--Xu canonical diffeomorphism. Another interesting feature is that we are able to introduce a grading, not a filtration, on our algebras of operators. When operators act on objects on vector bundles, we obtain a bi-grading.

math.DG↗

Laplace invariants of differential operators

We identify conditions giving large natural classes of partial differential operators for which it is possible to construct a complete set of Laplace invariants. In order to do that we investigate general properties of differential invariants of partial differential operators under gauge transformations and introduce a sufficient condition for a set of invariants to be complete. We also give a some mild conditions that guarantee the existence of such a set. The proof is constructive. The method gives many examples of invariants previously known in the literature as well as many new examples including multidimensional.

math-ph↗

Classification of Darboux transformations for operators of the form $\partial_x\partial_y +a \partial_x + b\partial_y +c$

Darboux transformations are non-group type symmetries of linear differential operators. One can define Darboux transformations algebraically by the intertwining relation $ML=L_1M$ or the intertwining relation $ML=L_1N$ in the cases when the first one is too restrictive. Here we show that Darboux transformations for operators of the form $\partial_x\partial_y +a \partial_x + b\partial_y +c$ (often referred to as 2D Schrödinger operators) are always compositions of atomic Darboux transformations of two different types. This is in contrast with the case of 1D Schrödinger operators and other 1D operators, where there is only one atomic kind of Darboux transformations.

math-ph↗

Differential operators on the algebra of densities and factorization of the generalized Sturm-Liouville operator

We consider factorization problem for differential operators on the commutative algebra of densities (defined either algebraically or in terms of an auxiliary extended manifold) introduced in 2004 by Khudaverdian and Voronov in connection with Batalin-Vilkovisky geometry. We consider the case of the line, where unlike the familiar setting (where operators act on functions) there are obstructions for factorization. We analyze these obstructions. In particular, we study the "generalized Sturm-Liouville" operators acting on the algebra of densities on the line. This in a certain sense is in between the 1D and 2D cases. We establish a criterion of factorizabily for the generalized Sturm-Liouville operator in terms of solution of the classical Sturm-Liouville equation. We also establish the possibility of an incomplete factorization.

math-ph↗

Differential operators on the superline, Berezinians, and Darboux transformations

We consider differential operators on a supermanifold of dimension $1|1$. We define non-degenerate operators as those with an invertible top coefficient in the expansion in the "superderivative" $D$ (which is the square root of the shift generator, the partial derivative in an even variable, with the help of an odd indeterminate). They are remarkably similar to ordinary differential operators. We show that every non-degenerate operator can be written in terms of `super Wronskians' (which are certain Berezinians). We apply this to Darboux transformations (DTs), proving that every DT of an arbitrary non-degenerate operator is the composition of elementary first order transformations. Hence every DT corresponds to an invariant subspace of the source operator and, upon a choice of basis in this subspace, is expressed by a super-Wronskian formula. We consider also dressing transformations, i.e., the effect of a DT on the coefficients of the non-degenerate operator. We calculate these transformations in examples and make some general statements.

math.DG↗

Classification of Multidimensional Darboux Transformations: First Order and Continued Type

We analyze Darboux transformations in very general settings for multidimensional linear partial differential operators. We consider all known types of Darboux transformations, and present a new type. We obtain a full classification of all operators that admit Wronskian type Darboux transformations of first order and a complete description of all possible first-order Darboux transformations. We introduce a large class of invertible Darboux transformations of higher order, which we call Darboux transformations of continued Type I. This generalizes the class of Darboux transformations of Type I, which was previously introduced. There is also a modification of this type of Darboux transformations, continued Wronskian type, which generalize Wronskian type Darboux transformations.

math.DG↗

Darboux transformations for differential operators on the superline

We give a full description of Darboux transformations of any order for arbitrary (nondegenerate) differential operators on the superline. We show that every Darboux transformation of such operators factorizes into elementary Darboux transformations of order one. Similar statement holds for operators on the ordinary line.

math-ph↗

Orbits of Darboux groupoid for hyperbolic operators of order three

Darboux transformations are viewed as morphisms in a Darboux category. Darboux transformations of type I which we defined previously, make an important subgroupoid consists of Darboux transformations of type I. We describe the orbits of this subgroupoid for hyperbolic operators of order three. We consider the algebras of differential invariants for our operators. In particular, we show that the Darboux transformations of this class can be lifted to transformations of differential invariants (which we calculate explicitly).

math-ph↗

Invertible Darboux Transformations

For operators of many different kinds it has been proved that (generalized) Darboux transformations can be built using so called Wronskian formulae. Such Darboux transformations are not invertible in the sense that the corresponding mappings of the operator kernels are not invertible. The only known invertible ones were Laplace transformations (and their compositions), which are special cases of Darboux transformations for hyperbolic bivariate operators of order 2. In the present paper we find a criteria for a bivariate linear partial differential operator of an arbitrary order d to have an invertible Darboux transformation. We show that Wronkian formulae may fail in some cases, and find sufficient conditions for such formulae to work.

math-ph↗

Invariants for Darboux transformations of Arbitrary Order for $D_x D_y +aD_x + bD_y +c$

We develop the method of regularized moving frames of Fels and Olver to obtain explicit general formulas for the basis invariants that generate all the joint differential invariants, under gauge transformations, for the operators [ø{L}=D_xD_y +a(x,y) D_x + b(x,y) D_y +c(x,y)] and an operator of arbitrary order. The problem appeared in connection with invariant construction of Darboux transformations for $ø{L}$.

math-ph↗

Proof of the Completeness of Darboux Wronskian Formulae for Order Two

Darboux Wronskian formulas allow to construct Darboux transformations, but Laplace transformations, which are Darboux transformations of order one cannot be represented this way. It has been a long standing problem on what are other exceptions. In our previous work we proved that among transformations of total order one there are no other exceptions. Here we prove that for transformations of total order two there are no exceptions at all. We also obtain a simple explicit invariant description of all possible Darboux Transformations of total order two.

math.AP↗