arXiv subjects

Anthony Joseph

Publications and source records attributed to Anthony Joseph.

At least 19 recordsLinked to original sources

arXiv2026

The Reverse Tableaux: a Gateway to the Surjectivity of the Component Map

Let $G$ be a simple algebraic group over $\mathbb C$, $B$ a fixed Borel subgroup, $P$ a parabolic subgroup, $P'$ its derived group acting on the Lie algebra $\mathfrak m$ of its nilradical. The nilfibre $\mathscr N$ is the zero locus of the augmentation $\mathcal I_+$ of the semiinvariant algebra $\mathcal I=\mathbb C[\mathfrak m]^{P'}$. Via Richardson's theorem, $\mathcal I$ is polynomial. Then the generators of $\mathcal I$ may be taken to be the Benlolo-Sanderson invariants \cite{BS}. In Y.Fittouhi and A.Joseph, The Magic and Mystery of Component Tableaux, Indag 2026, a set $\{\mathscr T^\mathcal C\}$ of component tableaux was constructed each encoding explicit combinatorial data $\mathcal C$. Each tableau $\{\mathscr T^\mathcal C\}$ defines a component $\mathscr C$ of $\mathscr N$ and the map $\{\mathscr T^\mathcal C\}\mapsto \mathscr C$ is injective. Here this data is simply encoded in a multiset called the Red Set. In the present work a set $\{\mathscr R^{\psi(\mathcal C)}\}$ of Reverse Tableaux is constructed through an Enabling Proposition. They define the same components as the component tableaux and furthermore give a factorisation of each new invariant in a chosen sequence via successive linearisation of preceding invariants. Via Krull's theorem this factorisation provides the required surjectivity. There can be several reverse tableaux for a given RedSet, yet each determine the same variety as the component tableaux with the given RedSet and define the same components. The flexibility of having several equivalent reverse tableaux absent from the more rigid component tableaux, is essential for factorisation This procedure is different from the classical approach to surjectivity requiring a \textit{geometric} description of the nilcone which seems unattainable. Nothing of this complexity has never been tackled before and the methods used here are entirely new

math.RT↗

arXiv2024

The Magic and Mystery of Component Tableaux

Let $G$ be a simple algebraic group over the complex field $\mathbb C$, $P$ a parabolic subgroup containing $B$ its Borel subgroup, $P'$ its derived group and $\mathfrak m$ the Lie algebra of its nilradical. The nilfibre $\mathscr N$ for this action is the zero locus of the augmentation $\mathscr I_+$ of the semi-invariant algebra $\mathscr I=\mathbb C[\mathfrak m]^{P'}$. For $G=SL(n)$ practically nothing was known previously. The only result of comparable, but lesser complexity, is for $\mathscr V:=\mathscr O\cap \mathfrak n$, with $\mathscr O$ a nilptent $G$ orbit and $\mathfrak n$ the set of strictly upper triangular matrices. Then $\mathscr V$ is equidimensional with components known as orbital varieties, parameterised by standard tableaux whose shape is dictated by $\mathscr O$. Here the components of $\mathscr N$ are studied for $G=SL(n)$. They increase exponentially in $n$ with no a priori discernable pattern. For each choice of numerical data $\mathcal C$, a semi-standard tableau $\mathscr T^\mathcal C$, is constructed from $\mathscr T$. A \textit{delicate and tightly interlocking} analysis constructs a set of excluded root vectors from $\mathfrak m$ such that the complementary space $\mathfrak u^\mathcal C$ has the following properties. First it is a subalgebra of $\mathfrak m$. Secondly $\mathscr C:=\overline{B.\mathfrak u^\mathcal C}$ lies in $\mathscr N$ to which, thirdly, a Weierstrass section can be associated. Fourthly $\dim \mathscr C = dim \mathfrak m-\textbf{g}$, where \textbf{g} is the number of generators of the polynomial algebra $\mathscr I$. Fifthly the Weierstrass section, is shown to imply that $\mathscr C$ an irreducible component of $\mathscr N$, yet $\mathscr C$ is \textit{ only sometimes} an orbital variety closure. The resulting Component Map $\mathscr T^\mathcal C\mapsto\mathscr C$ is shown to be injective. Evidence for its surjectivity is given.

math.RT↗

arXiv2022

The Canonical Component of the nilfibre for Parabolic adjoint action in type $A$

This work is a continuation of [Fittouhi and Joseph, Parabolic adjoint action, Weierstrass Sections and components of the nilfibre in type $A$]. Let $P$ be a parabolic subgroup of an irreducible simple algebraic group $G$, $P'$ its derived group and $\mathfrak m$ be the nilradical to its Lie algebra. A theorem of Richardson implies that the subalgebra $\mathbb C[\mathfrak m]^{P'}$, spanned by the $P$ semi-invariants in $\mathbb C[\mathfrak m]$, is polynomial. A linear subvariety $e+V$ of $\mathfrak m$ is is called a Weierstrass section for the action of $P'$ on $\mathfrak m$, if the restriction map induces an isomorphism of $\mathbb C[\mathfrak m]^{P'}$ onto $\mathbb C[e+V]$. Thus a Weierstrass section can exist only if the latter is polynomial, but even when this holds its existence is far from assured. The existence of a Weierstrass section $e+V$ in $\mathfrak m$ was established by a general combinatorial construction. Notably $e \in \mathscr N$ and is a sum of root vectors with linearly independent roots. The Weierstraass section $e+V$ looks very different for different choices of parabolics but nevertheless has a uniform construction and exists in all cases. It is called the "canonical Weierstrass section". It was announced in [Fittouhi and Joseph, loc. cit.] that one may augment $e$ to an element $e_{VS}$ by adjoining root vectors. Then the linear span $E_{VS}$ of these root vectors lies in $\mathscr N^e$ and its closure is just $\mathscr N^e$. Yet this result shows that $\mathscr N^e$ need not admit a dense $P$ orbit. However this theorem was only verified in the special case needed to obtain the example showing that $\mathscr N^e$ may fail to admit a dense $P$ orbit. Here a general proof is given. Finally a map from compositions to the set of distinct non-negative integers is defined. Its image is shown to determine the canonical Weierstrass section.

math.RT↗

arXiv2022

The Composition Tableau and Reconstruction of the Canonical Weierstrass Section for Parabolic Adjoint Action in type $A$

A "Composition map" is constructed, leaning heavily on earlier work [Y. Fittouhi and A. Joseph, Parabolic adjoint action, Weierstrass sections and components of the nilfibre in type $A$, Indag Math. and Y. Fittouhi and A. Joseph, The canonical component of the nilfibre for parabolic adjoint action, Weierstrass sections in type $A$, preprint, Weizmann, 2021]. It defines a composition tableau which recovers the "canonical" Weierstrass section $e+V$ described in the first paper above. Moreover \textit{without reference to this earlier work}, it is then shown that $e+V$ is indeed a Weierstrass section. This results in a huge simplification. Moreover one may read off from the composition tableau the "VS quadruplets'' of the second of the above papers, thereby describing the "canonical component" of the nil-fibre in which $e$ lies but does not of itself determine.

math.RT↗

arXiv2022

SCL: A Secure Concurrency Layer For Paranoid Stateful Lambdas

We propose a federated Function-as-a-Service (FaaS) execution model that provides secure and stateful execution in both Cloud and Edge environments. The FaaS workers, called Paranoid Stateful Lambdas (PSLs), collaborate with one another to perform large parallel computations. We exploit cryptographically hardened and mobile bundles of data, called DataCapsules, to provide persistent state for our PSLs, whose execution is protected using hardware-secured TEEs. To make PSLs easy to program and performant, we build the familiar Key-Value Store interface on top of DataCapsules in a way that allows amortization of cryptographic operations. We demonstrate PSLs functioning in an edge environment running on a group of Intel NUCs with SGXv2. As described, our Secure Concurrency Layer (SCL), provides eventually-consistent semantics over written values using untrusted and unordered multicast. All SCL communication is encrypted, unforgeable, and private. For durability, updates are recorded in replicated DataCapsules, which are append-only cryptographically-hardened blockchain with confidentiality, integrity, and provenance guarantees. Values for inactive keys are stored in a log-structured merge-tree (LSM) in the same DataCapsule. SCL features a variety of communication optimizations, such as an efficient message passing framework that reduces the latency up to 44x from the Intel SGX SDK, and an actor-based cryptographic processing architecture that batches cryptographic operations and increases throughput by 81x.

cs.CR↗

arXiv2021

Parabolic Adjoint Action, Weierstrass Sections and Components of the Nilfibre in Type $A$

This work is a continuation of [Y. Fittouhi and A. Joseph, Weierstrass Sections for Parabolic adjoint action in type $A$]. Let $G$ be an irreducible simple algebraic group and $B$ a Borel subgroup of $G$. Let $\mathfrak n$ be the Lie algebra of the nilradical of $B$. Consider an irreducible subgroup $P$ of $G$ containing $B$. Let $P'$ be the derived group of $P$. Let $\mathfrak m$ be the Lie algebra of the nilradical of $P$. A theorem of Richardson asserts that the algebra $\mathbb C[\mathfrak m]^{P'}$ of $P$ semi-invariants is multiplicity-free. A linear subvariety $e+V$ such that the restriction map induces an isomorphism of $\mathbb C[\mathfrak m]^{P'}$ onto $\mathbb C[e+V]$ is called a Weierstrass section for the action of $P'$ on $\mathfrak m$. Here in type $A$ such a section is constructed, but in better form than that given in Sect. 4, loc cit. Yet the main difference is a complete change of emphasis from the construction of a Weierstrass section, to its application. Let $\mathscr N$ be the nilfibre relative to this action. From the construction of a Weierstrass section $e+V$, it is shown that $e \in \mathscr N$. Then $P.e$ is contained in a unique irreducible component $\mathscr C$ of $\mathscr N$. The structure of $e+V$ is used to give a rather explicit description of $\mathscr C$ as a $B$ saturation set, that is of the form $\overline{B.\mathfrak u}$, where $\mathfrak u$ is a subalgebra of $\mathfrak n$ . This algebra is not necessarily complemented by a subalgebra in $\mathfrak n$ and so $\overline{B.\mathfrak u}$ is not necessarily an orbital variety closure (hence Lagrangian) but it can be. It is shown that $\mathscr C$ need not contain a dense $P$ orbit and this by a purely theoretical analysis. This occurs for an appropriate parabolic in $A_{10}$ and is possibly the simplest example. In this particular case $\mathscr C$ is not an orbital variety closure.

math.RT↗

arXiv2020

Weierstrass sections for parabolic adjoint action in type $A$

The notion of "Weierstrass Section", comes from Weierstrass canonical form for elliptic curves. In celebrated work [B. Kostant, Lie group representations on polynomial rings, Amer. J. Math. 85 (1963), 327-404] constructed such a section for the action of a semisimple Lie algebra on its dual using a principal s-triple. Actually it is enough to have an "adapted pair" and indeed the construction in [A. Joseph and D. Shafrir, Polynomiality of invariants, unimodularity and adapted pairs, Transform. Groups 15 (2010), no. 4, 851-882] works rather well for the coadjoint action of an algebraic, but not necessarily reductive Lie algebra. In the present work a Weierstrass section is constructed for the adjoint action of the derived algebra of a parabolic subalgebra on its nilradical in type $A$. The starting point is Richardson's theorem which implies the polynomiality of the invariant sub-algebra. Here adapted pairs seldom exist. A new construction is developed and this is mainly combinatorial based on joining boxes in the Young tableau associated to the "Richardson component". Indications are given for extending this construction in other types. The construction has relations to quivers [T. Brustle, L. Hille, Lutz, C.M. Ringel and G. Rohrle, The $\delta$-filtered modules without selfextensions for the Auslander algebra of $k[T]/Tn$. Algebr. Represent. Theory 2 (1999), no. 3, 295-312] and to hypersurface orbital varieties [A. Joseph and A. Melnikov, Quantization of hypersurface orbital varieties in $sl(n)$. The orbit method in geometry and physics (Marseille, 2000), 165-196, Progr. Math., 213].

math.RT↗

arXiv2017

Convolutional Kitchen Sinks for Transcription Factor Binding Site Prediction

We present a simple and efficient method for prediction of transcription factor binding sites from DNA sequence. Our method computes a random approximation of a convolutional kernel feature map from DNA sequence and then learns a linear model from the approximated feature map. Our method outperforms state-of-the-art deep learning methods on five out of six test datasets from the ENCODE consortium, while training in less than one eighth the time.

q-bio.GN↗

arXiv2017

Trails, $S$-graphs and Identities in Demazure Modules

The Kashiwara crystal $B(\infty)$ parametrizes a basis for the Verma module of a Kac-Moody algebra. It has a deep combinatorial structure which one seeks to understand. For each sequence $J$ of reduced decompositions of elements of the Weyl group $W$, it has a realization as a subset $B_J(\infty)$ of a crystal $B_J$ which as a set is just $J$ copies of the natural numbers. The goal is to determine $B_J(\infty)$ and in particular to show that it is a polyhedral subset of $B_J$. In earlier work this led to the notion of an $S$-graph associated to a given simple root $\alpha$. Here the notion of a giant $S$-graph depending on a fixed simple root is introduced. It is essentially a union of $S$-graphs for each simple root with one distinguished vertex depending on $\alpha$. Its vertices, which forms a giant $S$-set, determine a set of dual Kashiwara functions. These are linear functions on $B_J$, whose common maximum determines the dual Kashiwara parameter with respect to $\alpha$. From these parameters one may compute $B_J(\infty)$ as an explicit polyhedral subset of $B_J$. For $W$ finite, Berenstein and Zelevinsky had studied this problem by introducing the notion of a trail in a fundamental module. The functions they define may also be viewed as a set of dual Kashiwara functions. The goal is to relate these two approaches and without restriction on $W$. It is shown under the hypothesis that no "false" trails exist, that the set of trails determines the "$Z$-convex envelope" of a giant $S$-set. The proof involves the study of identities in Demazure modules.

math.RT↗

arXiv2017

Convexity properties of the canonical S-graphs

Let n be a positive integer and c an n-tuple of natural numbers. A convex set in Euclidean n-space given by a family of linear relations in the elements of c and depending on their natural order is defined. The extremal elements of this convex set are shown to be the S-set obtained from the S-graph defined by c constructed in studying the Kashiwara crystal B associated to a Verma module for a Kac-Moody algebra. The result has applications to the precise description of B.

math.RT↗

arXiv2015

A new interpretation of Catalan numbers

Towards the study of the Kashiwara B(infinity) crystal, sets H^t of functions were introduced given by equivalence classes of unordered partitions satisfying certain boundary conditions. Here it is shown that H^t is a Catalan set of order t, that is to say the cardinality of H^t is the t-th Catalan number C(t). This is a new description of a Catalan set and moreover admits some remarkable features. Thus to H^t there is an associated labelled graph G_t which is shown to have a canonical decomposition into (t-1)! subgraphs each with 2^{t-1} vertices. These subgraphs, called S-graphs, have some tight properties which are needed for the study of B(infinity). They are described as labelled hypercubes whose edges connecting vertices with equal labels are missing. It is shown that the number of distinct hypercubes so obtained is again a Catalan number, namely C(t-1). They define functions which depend on a coefficient set of non-negative integers. When the latter are non-zero and pairwise distinct, the vertices of the S-graphs describe distinct functions. Moreover this property is retained if certain edges are deleted and certain vertices identified. In particular when these coefficients are all equal and non-zero, it is shown that every hypercube degenerates to a simplex, resulting in exactly t distinct functions, which for example are exactly those needed in the description of B(infinity) in type A.

math.CO↗

arXiv2015

A preparation theorem for the kashiwara $b(\infty)$ crystal

The Kashiwara $B(\infty)$ crystal pertains to a Verma module for a Kac- Moody Lie algebra. Ostensibly it provides only a parametrisation of the global/canonical basis for the latter. Yet it is much more having a rich combinatorial structure from which one may read of a parametrisation of the corresponding basis for any integrable highest weight module, describe the decomposition of the tensor products of highest weight modules, the Demazure submodules of integrable highest weight modules and Demazure flags for translates of Demazure modules. $B(\infty)$ has in general infinitely many presentations as subsets of countably many copies of the natural numbers each given by successive reduced decompositions of Weyl group elements. In each presentation there is an action of Kashiwara operators determined by Kashiwara functions. These functions are linear in the entries. Thus a natural question is to show that in each presentation the subset $B(\infty)$ is polyhedral. Here a new approach to this question is initiated based on constructing dual Kashiwara functions and in this it is enough to show that the latter are also linear in the entries. The present work resolves one of the two very difficult obstacles in a step-wise construction, namely that the resulting functions must satisfy a sum, or simply S, condition. It depends very subtly on inequalities between the coefficients occurring in functions obtained from the previous step. The only remaining obstacle, that sufficiently many functions are obtained, can at least be verified in many families of cases, though this is to be postponed to a subsequent paper. This theory has some intriguing numerology which involves the Catalan numbers in two different ways.

math.CO↗

arXiv2015

Adapted pairs and Weierstrass sections

Adapted pairs and Weierstrass sections are central to the invariant theory associated to the action of an algebraic Lie algebra a on a finite dimensional vector space X. In this a need not be a semisimple Lie algebra. Here their general properties are described particularly when a is the canonical truncation of a biparabolic subalgebra of a simple Lie algebra and X is the dual of a.

math.RT↗

arXiv2014

A Littelmann path model for crystals of Generalized Kac-Moody algebras revisited

A Littelmann path model is constructed for crystals pertaining to a not necessarily symmetrizable Borcherds-Cartan matrix. Here one must overcome several combinatorial problems coming from the imaginary simple roots. The main results are an isomorphism theorem and a character formula of Borcherds-Kac-Weyl type for the crystals. In the symmetrizable case, the isomorphism theorem implies that the crystals constructed by this path model coincide with those of Jeong, Kang, Kashiwara and Shin obtained by taking the limit at q=0 in the quantized enveloping algebra.

math.RT↗

arXiv2014

The integrality of an adapted pair

Let $\mathfrak a$ be an algebraic Lie algebra. An adapted pair for $\mathfrak a$ is pair $(h,η)$ consisting of an ad-semisimple element of $h \in \mathfrak a$ and a regular element of $η\in \mathfrak a^*$ satisfying $(ad \ h)η=-η$. An adapted pair $(h,η)$ is said to satisfy integrality if $ad \ h$ has integer eigenvalues on $\mathfrak a$. Integrality is shown to hold for any Frobenius Lie algebra which is a biparabolic subalgebra of a semisimple Lie algebra; but may fail in general. Call $\mathfrak a$ regular if there are no proper semi-invariant polynomial functions on $\mathfrak a^*$ and if the subalgebra of invariant functions is polynomial. In this case there are no known counter-examples to integrality. It is shown that if $\mathfrak a$ is the canonical truncation of a biparabolic subalgebra of a simple Lie algebra $\mathfrak g$ which is regular and admits an adapted pair $(h,η)$, then the eigenvalues of $ad \ h$ on $\mathfrak a$ lie in $\frac{1}{m}\mathbb Z$, where $m$ is a coefficient of a simple root in the highest root of $\mathfrak g$. Let $\mathfrak a$ be a regular Lie algebra admitting an adapted pair $(h,η)$. Let $\mathfrak a_\mathbb Z$ be the subalgebra spanned by the eigensubspaces of $ad \ h$ with integer eigenvalue. It is shown that the canonical truncation of $\mathfrak a_\mathbb Z$ is regular. Sufficient knowledge of the relation between the generators for the invariant polynomial functions on $\mathfrak a^*$ and on $\mathfrak a^*_\mathbb Z$ can then lead to establishing the integrality of $(h,η)$. This method is used to show the integrality of an adapted pair for a truncated parabolic subalgebra of a simple Lie algebra of type $C$.

math.RT↗

arXiv2013

Modules for Relative Yangians (Family Algebras) and Kazhdan-Lusztig Polynomials

Let g be a complex simple Lie algebra and let V be a finite dimensional U(g) module. A relative Yangian is defined with respect to this pair. According to recent work of Khoroshkin and Nazarov the finite dimensional simple modules of the Yangians for g = sl(n) or the twisted Yangians for g = sp(2n); so(n) are described by the simple modules of relative Yangians for some V using the appropriate simple Lie algebra g. Here a classification of the simple modules of a relative Yangian is obtained simply and briefly as an advanced exercise in Frobenius reciprocity inspired by a Bernstein- Gelfand equivalence of categories. An unexpected fact is that the dimension of these modules are determined by the Kazhdan-Lusztig polynomials and conversely the latter are described in terms of dimensions of certain extension groups associated to finite dimensional modules of relative Yangians.

math.RT↗

arXiv2013

Adapted pairs in type $A$ and regular nilpotent elements

Let $\mathfrak g$ be a simple Lie algebra over an algebraically closed field $\bf k$ of characteristic zero and $\bf G$ its adjoint group. Let $\mathfrak q$ be a biparabolic subalgebra of $\mathfrak g$. The algebra $Sy(\mathfrak q)$ of semi-invariants on $\mathfrak q^*$ is polynomial in most cases, in particular when $\mathfrak g$ is simple of type $A$ or $C$. On the other hand $\mathfrak q$ admits a canonical truncation $\mathfrak q_{\Lambda}$ such that $Sy(\mathfrak q)=Sy(\mathfrak q_{\Lambda})=Y(\mathfrak q_{\Lambda})$ where $Y(\mathfrak q_{\Lambda})$ denotes the algebra of invariant functions on $\mathfrak q_{\Lambda}^*$. An adapted pair for $\mathfrak q_{\Lambda}$ is a pair $(h,\,\eta)\in \mathfrak q_{\Lambda}\times\mathfrak q_{\Lambda}^*$ such that $\eta$ is regular and $(ad\,h)\eta=-\eta$. In a previous paper of A. Joseph (2008) adapted pairs for every truncated biparabolic subalgebra $\mathfrak q_{\Lambda}$ of a simple Lie algebra $\mathfrak g$ of type $A$ were constructed and then provide Weierstrass sections for $Y(\mathfrak q_{\Lambda})$ in $\mathfrak q_{\Lambda}^*$. These latter are linear subvarieties $\eta+V$ of $\mathfrak q_{\Lambda}^*$ such that the restriction map induces an algebra isomorphism of $Y(\mathfrak q_{\Lambda})$ onto the algebra of regular functions on $\eta+V$. Here we show that for each of the adapted pairs $(h,\,\eta)$ constructed in the paper mentioned above one can express $\eta$ as the image of a regular nilpotent element $y$ of $\mathfrak g^*$ under the restriction to $\mathfrak q$. Since $y$ must be a $\bf G$ translate of the standard regular nilpotent element defined in terms of the already chosen set $\pi$ of simple roots, one may attach to $y$ a unique element of the Weyl group. Ultimately one can then hope to be able to describe adapted pairs (in general) through the Weyl group.

math.RT↗

arXiv2011

Analogue Zhelobenko Invariants and the Kostant Clifford Algebra Conjecture

Let g be a complex simple Lie algebra and h a Cartan subalgebra. The Clifford algebra C(g) of g admits a Harish-Chandra map. Kostant conjectured (as communicated to Bazlov in about 1997) that the value of this map on a (suitably chosen) fundamental invariant of degree 2m+1 is just the zero weight vector of the simple 2m+1-dimensional module of the principal s-triple obtained from the Langlands dual of g. Bazlov settled this conjecture positively in type A. The Kostant conjecture was reformulated (Alekseev-Bazlov-Rohr) in terms of the Harish-Chandra map for the enveloping algebra U(g) composed with evaluation at the half sum of the positive roots. In an earlier work we settled an analogue of the Kostant conjecture obtained by replacing the Harish-Chandra map by a "generalized Harish-Chandra" map whose image is described via Zhelobenko invariants. Here we show that there are analogue Zhelobenko invariants which describe the image of the Harish-Chandra map. Following this a similar proof goes through.

math.RT↗