Showing 1–15 of 15 results for author: Axelsson, A

Using Large Language Models for Zero-Shot Natural Language Generation from Knowledge Graphs

Authors: Agnes Axelsson, Gabriel Skantze

Abstract: In any system that uses structured knowledge graph (KG) data as its underlying knowledge representation, KG-to-text generation is a useful tool for turning parts of the graph data into text that can be understood by humans. Recent work has shown that models that make use of pretraining on large amounts of text data can perform well on the KG-to-text task even with relatively small sets of training… ▽ More In any system that uses structured knowledge graph (KG) data as its underlying knowledge representation, KG-to-text generation is a useful tool for turning parts of the graph data into text that can be understood by humans. Recent work has shown that models that make use of pretraining on large amounts of text data can perform well on the KG-to-text task even with relatively small sets of training data on the specific graph-to-text task. In this paper, we build on this concept by using large language models to perform zero-shot generation based on nothing but the model's understanding of the triple structure from what it can read. We show that ChatGPT achieves near state-of-the-art performance on some measures of the WebNLG 2020 challenge, but falls behind on others. Additionally, we compare factual, counter-factual and fictional statements, and show that there is a significant connection between what the LLM already knows about the data it is parsing and the quality of the output text. △ Less

Submitted 22 August, 2023; v1 submitted 14 July, 2023; originally announced July 2023.

Comments: 9 pages, 3 pages appendices, 1 figure, 4 tables (incl. appendices)

MSC Class: 68T50 ACM Class: I.2.7; I.2.4

Uncovering the Handwritten Text in the Margins: End-to-end Handwritten Text Detection and Recognition

Authors: Liang Cheng, Jonas Frankemölle, Adam Axelsson, Ekta Vats

Abstract: The pressing need for digitization of historical documents has led to a strong interest in designing computerised image processing methods for automatic handwritten text recognition. However, not much attention has been paid on studying the handwritten text written in the margins, i.e. marginalia, that also forms an important source of information. Nevertheless, training an accurate and robust rec… ▽ More The pressing need for digitization of historical documents has led to a strong interest in designing computerised image processing methods for automatic handwritten text recognition. However, not much attention has been paid on studying the handwritten text written in the margins, i.e. marginalia, that also forms an important source of information. Nevertheless, training an accurate and robust recognition system for marginalia calls for data-efficient approaches due to the unavailability of sufficient amounts of annotated multi-writer texts. Therefore, this work presents an end-to-end framework for automatic detection and recognition of handwritten marginalia, and leverages data augmentation and transfer learning to overcome training data scarcity. The detection phase involves investigation of R-CNN and Faster R-CNN networks. The recognition phase includes an attention-based sequence-to-sequence model, with ResNet feature extraction, bidirectional LSTM-based sequence modeling, and attention-based prediction of marginalia. The effectiveness of the proposed framework has been empirically evaluated on the data from early book collections found in the Uppsala University Library in Sweden. Source code and pre-trained models are available at Github. △ Less

Submitted 29 January, 2024; v1 submitted 10 March, 2023; originally announced March 2023.

Intrinsic paramagnetism and aggregation of manganese dopants in SrTiO$_3$

Authors: Andrej Zorko, Matej Pregelj, Hubertus Luetkens, Anna-Karin Axelsson, Matjaz Valant

Abstract: Using local-probe magnetic-characterization techniques of muon spin relaxation ($μ$SR) and electron spin resonance (ESR) we have investigated the Mn-induced magnetism of the wide-bandgap perovskite SrTiO$_3$. Our results clearly demonstrate that this diluted magnetic oxide (DMO) remains paramagnetic down to low temperatures for both doping cases, i.e., when Mn substitutes for Sr or Ti. In addition… ▽ More Using local-probe magnetic-characterization techniques of muon spin relaxation ($μ$SR) and electron spin resonance (ESR) we have investigated the Mn-induced magnetism of the wide-bandgap perovskite SrTiO$_3$. Our results clearly demonstrate that this diluted magnetic oxide (DMO) remains paramagnetic down to low temperatures for both doping cases, i.e., when Mn substitutes for Sr or Ti. In addition, both experimental techniques have unveiled that the distribution of individual Mn$^{2+}$ and Mn$^{4+}$ ions is nonrandom, as these ions partially aggregate into nanosized clusters. △ Less

Submitted 22 March, 2014; v1 submitted 3 January, 2014; originally announced January 2014.

Comments: a few minor changes were made

Journal ref: Phys. Rev. B 89, 094418 (2014)

Remarks on maximal regularity

Authors: Pascal Auscher, Andreas Axelsson

Abstract: We prove weighted estimates for the maximal regularity operator. Such estimates were motivated by boundary value problems. We take this opportunity to study a class of weak solutions to the abstract Cauchy problem. We also give a new proof of maximal regularity for closed and maximal accretive operators following from Kato's inequality for fractional powers and almost orthogonality arguments. We prove weighted estimates for the maximal regularity operator. Such estimates were motivated by boundary value problems. We take this opportunity to study a class of weak solutions to the abstract Cauchy problem. We also give a new proof of maximal regularity for closed and maximal accretive operators following from Kato's inequality for fractional powers and almost orthogonality arguments. △ Less

Submitted 22 December, 2009; originally announced December 2009.

Comments: submitted to a volume in honor of Amann's birthday

MSC Class: 47D06; 35K90; 47A60

Weighted maximal regularity estimates and solvability of non-smooth elliptic systems I

Authors: Pascal Auscher, Andreas Axelsson

Abstract: We develop new solvability methods for divergence form second order, real and complex, elliptic systems above Lipschitz graphs, with $L_2$ boundary data. The coefficients $A$ may depend on all variables, but are assumed to be close to coefficients $A_0$ that are independent of the coordinate transversal to the boundary, in the Carleson sense $\|A-A_0\|_C$ defined by Dahlberg. We obtain a number of… ▽ More We develop new solvability methods for divergence form second order, real and complex, elliptic systems above Lipschitz graphs, with $L_2$ boundary data. The coefficients $A$ may depend on all variables, but are assumed to be close to coefficients $A_0$ that are independent of the coordinate transversal to the boundary, in the Carleson sense $\|A-A_0\|_C$ defined by Dahlberg. We obtain a number of {\em a priori} estimates and boundary behaviour results under finiteness of $\|A-A_0\|_C$. Our methods yield full characterization of weak solutions, whose gradients have $L_2$ estimates of a non-tangential maximal function or of the square function, via an integral representation acting on the conormal gradient, with a singular operator-valued kernel. Also, the non-tangential maximal function of a weak solution is controlled in $L_2$ by the square function of its gradient. This estimate is new for systems in such generality, and even for real non-symmetric equations in dimension 3 or higher. The existence of a proof {\em a priori} to well-posedness, is also a new fact. As corollaries, we obtain well-posedness of the Dirichlet, Neumann and Dirichlet regularity problems under smallness of $\|A-A_0\|_C$ and well-posedness for $A_0$, improving earlier results for real symmetric equations. Our methods build on an algebraic reduction to a first order system first made for coefficients $A_0$ by the two authors and A. McIntosh in order to use functional calculus related to the Kato conjecture solution, and the main analytic tool for coefficients $A$ is an operational calculus to prove weighted maximal regularity estimates. △ Less

Submitted 15 September, 2010; v1 submitted 23 November, 2009; originally announced November 2009.

Comments: This is an extended version of the paper, containing some new material and a road map to proofs on suggestion from the referees

MSC Class: 35J55; 42B25; 47N20

On a quadratic estimate related to the Kato conjecture and boundary value problems

Authors: Pascal Auscher, Andreas Axelsson, Alan McIntosh

Abstract: We provide a direct proof of a quadratic estimate that plays a central role in the determination of domains of square roots of elliptic operators and, as shown more recently, in some boundary value problems with $L^2$ boundary data. We develop the application to the Kato conjecture and to a Neumann problem. This quadratic estimate enjoys some equivalent forms in various settings. This gives new… ▽ More We provide a direct proof of a quadratic estimate that plays a central role in the determination of domains of square roots of elliptic operators and, as shown more recently, in some boundary value problems with $L^2$ boundary data. We develop the application to the Kato conjecture and to a Neumann problem. This quadratic estimate enjoys some equivalent forms in various settings. This gives new results in the functional calculus of Dirac type operators on forms. △ Less

Submitted 17 May, 2009; v1 submitted 17 October, 2008; originally announced October 2008.

Comments: Text of the lectures given at the El Escorial 2008 conference. Revised after the suggestions of the referee. Some historical material added. A short proof of the main result added under a further assumption. To appear in the Proceedings

MSC Class: 35J25; 35J55; 47N20; 47F05; 42B25

Solvability of elliptic systems with square integrable boundary data

Authors: Pascal Auscher, Andreas Axelsson, Alan McIntosh

Abstract: We consider second order elliptic divergence form systems with complex measurable coefficients $A$ that are independent of the transversal coordinate, and prove that the set of $A$ for which the boundary value problem with $L_2$ Dirichlet or Neumann data is well posed, is an open set. Furthermore we prove that these boundary value problems are well posed when $A$ is either Hermitean, block or co… ▽ More We consider second order elliptic divergence form systems with complex measurable coefficients $A$ that are independent of the transversal coordinate, and prove that the set of $A$ for which the boundary value problem with $L_2$ Dirichlet or Neumann data is well posed, is an open set. Furthermore we prove that these boundary value problems are well posed when $A$ is either Hermitean, block or constant. Our methods apply to more general systems of PDEs and as an example we prove perturbation results for boundary value problems for differential forms. △ Less

Submitted 29 September, 2008; originally announced September 2008.

Comments: This paper replaces its predecessor "A new approach to solvability of some elliptic pde's with square integrable boundary data" by the same authors

MSC Class: 35J25; 35J55; 47N20

Hilbert transforms and the Cauchy integral in euclidean space

Authors: Andreas Axelsson, Kit Ian Kou, Tao Qian

Abstract: We generalize the notion of harmonic conjugate functions and Hilbert transforms to higher dimensional euclidean spaces, in the setting of differential forms and the Hodge-Dirac system. These conjugate functions are in general far from being unique, but under suitable boundary conditions we prove existence and uniqueness of conjugates. The proof also yields invertibility results for a new class o… ▽ More We generalize the notion of harmonic conjugate functions and Hilbert transforms to higher dimensional euclidean spaces, in the setting of differential forms and the Hodge-Dirac system. These conjugate functions are in general far from being unique, but under suitable boundary conditions we prove existence and uniqueness of conjugates. The proof also yields invertibility results for a new class of generalized double layer potential operators on Lipschitz surfaces and boundedness of related Hilbert transforms. △ Less

Submitted 30 April, 2009; v1 submitted 24 September, 2008; originally announced September 2008.

Comments: Some minor corrections made

MSC Class: 45E05; 31B10

A new approach to solvability of some elliptic pde's with square integrable boundary data

Authors: Pascal Auscher, Andreas Axelsson, Alan McIntosh

Abstract: This paper has been withdrawn, and is replaced with paper "Solvability of elliptic systems with square integrable boundary data" by the same authors. This paper has been withdrawn, and is replaced with paper "Solvability of elliptic systems with square integrable boundary data" by the same authors. △ Less

Submitted 29 September, 2008; v1 submitted 5 February, 2008; originally announced February 2008.

MSC Class: 35J25; 35J55; 47N20

Non unique solutions to boundary value problems for non symmetric divergence form equations

Authors: Andreas Axelsson

Abstract: We calculate explicitly solutions to the Dirichlet and Neumann boundary value problems in the upper half plane, for a family of divergence form equations with non symmetric coefficients with a jump discontinuity. It is shown that the boundary equation method and the Lax--Milgram method for constructing solutions may give two different solutions when the coefficients are sufficiently non symmetri… ▽ More We calculate explicitly solutions to the Dirichlet and Neumann boundary value problems in the upper half plane, for a family of divergence form equations with non symmetric coefficients with a jump discontinuity. It is shown that the boundary equation method and the Lax--Milgram method for constructing solutions may give two different solutions when the coefficients are sufficiently non symmetric. △ Less

Submitted 31 October, 2007; v1 submitted 14 September, 2007; originally announced September 2007.

Comments: Revised version. To appear in Transactions of the American Mathematical Society

MSC Class: 35J25; 42A50

Analyticity of layer potentials and $L^{2}$ solvability of boundary value problems for divergence form elliptic equations with complex $L^{\infty}$ coefficients

Authors: M. Alfonseca, P. Auscher, A. Axelsson, S. Hofmann, S. Kim

Abstract: We consider divergence form elliptic operators of the form $L=-\dv A(x)\nabla$, defined in $R^{n+1} = \{(x,t)\in R^n \times R \}$, $n \geq 2$, where the $L^{\infty}$ coefficient matrix $A$ is $(n+1)\times(n+1)$, uniformly elliptic, complex and $t$-independent. We show that for such operators, boundedness and invertibility of the corresponding layer potential operators on… ▽ More We consider divergence form elliptic operators of the form $L=-\dv A(x)\nabla$, defined in $R^{n+1} = \{(x,t)\in R^n \times R \}$, $n \geq 2$, where the $L^{\infty}$ coefficient matrix $A$ is $(n+1)\times(n+1)$, uniformly elliptic, complex and $t$-independent. We show that for such operators, boundedness and invertibility of the corresponding layer potential operators on $L^2(\mathbb{R}^{n})=L^2(\partial\mathbb{R}_{+}^{n+1})$, is stable under complex, $L^{\infty}$ perturbations of the coefficient matrix. Using a variant of the $Tb$ Theorem, we also prove that the layer potentials are bounded and invertible on $L^2(\mathbb{R}^n)$ whenever $A(x)$ is real and symmetric (and thus, by our stability result, also when $A$ is complex, $\Vert A-A^0\Vert_{\infty}$ is small enough and $A^0$ is real, symmetric, $L^{\infty}$ and elliptic). In particular, we establish solvability of the Dirichlet and Neumann (and Regularity) problems, with $L^2$ (resp. $\dot{L}^2_1)$ data, for small complex perturbations of a real symmetric matrix. Previously, $L^2$ solvability results for complex (or even real but non-symmetric) coefficients were known to hold only for perturbations of constant matrices (and then only for the Dirichlet problem), or in the special case that the coefficients $A_{j,n+1}=0=A_{n+1,j}$, $1\leq j\leq n$, which corresponds to the Kato square root problem. △ Less

Submitted 6 May, 2007; originally announced May 2007.

MSC Class: 42B20; 42B25; 35J25

Journal ref: Adv. Math. 226 (2011), no. 5, 4533--4606

Functional calculus of Dirac operators and complex perturbations of Neumann and Dirichlet problems

Authors: Pascal Auscher, Andreas Axelsson, Steve Hofmann

Abstract: We prove that the Neumann, Dirichlet and regularity problems for divergence form elliptic equations in the half space are well posed in $L_2$ for small complex $L_\infty$ perturbations of a coefficient matrix which is either real symmetric, of block form or constant. All matrices are assumed to be independent of the transversal coordinate. We solve the Neumann, Dirichlet and regularity problems… ▽ More We prove that the Neumann, Dirichlet and regularity problems for divergence form elliptic equations in the half space are well posed in $L_2$ for small complex $L_\infty$ perturbations of a coefficient matrix which is either real symmetric, of block form or constant. All matrices are assumed to be independent of the transversal coordinate. We solve the Neumann, Dirichlet and regularity problems through a new boundary operator method which makes use of operators in the functional calculus of an underlaying first order Dirac type operator. We establish quadratic estimates for this Dirac operator, which implies that the associated Hardy projection operators are bounded and depend continuously on the coefficient matrix. We also prove that certain transmission problems for $k$-forms are well posed for small perturbations of block matrices. △ Less

Submitted 3 May, 2007; v1 submitted 2 May, 2007; originally announced May 2007.

Comments: Some changes made in the introduction of the paper

MSC Class: 35J25; 35J55; 47N20

The Kato square root problem for mixed boundary value problems

Authors: Andreas Axelsson, Stephen Keith, Alan McIntosh

Abstract: We solve the Kato square root problem for second order elliptic systems in divergence form under mixed boundary conditions on Lipschitz domains. This answers a question posed by J.-L. Lions in 1962. To do this we develop a general theory of quadratic estimates and functional calculi for complex perturbations of Dirac-type operators on Lipschitz domains. We solve the Kato square root problem for second order elliptic systems in divergence form under mixed boundary conditions on Lipschitz domains. This answers a question posed by J.-L. Lions in 1962. To do this we develop a general theory of quadratic estimates and functional calculi for complex perturbations of Dirac-type operators on Lipschitz domains. △ Less

Submitted 24 October, 2005; v1 submitted 16 December, 2004; originally announced December 2004.

Comments: To appear in the Journal of the London Mathematical Society. This arxiv version is distinct from the final version

MSC Class: 35J25; 47B44; 47F05

Quadratic estimates and functional calculi of perturbed Dirac operators

Authors: Andreas Axelsson, Stephen Keith, Alan McIntosh

Abstract: We prove quadratic estimates for complex perturbations of Dirac-type operators, and thereby show that such operators have a bounded functional calculus. As an application we show that spectral projections of the Hodge--Dirac operator on compact manifolds depend analytically on $L_\infty$ changes in the metric. We also recover a unified proof of many results in the Calderón program, including the… ▽ More We prove quadratic estimates for complex perturbations of Dirac-type operators, and thereby show that such operators have a bounded functional calculus. As an application we show that spectral projections of the Hodge--Dirac operator on compact manifolds depend analytically on $L_\infty$ changes in the metric. We also recover a unified proof of many results in the Calderón program, including the Kato square root problem and the boundedness of the Cauchy operator on Lipschitz curves and surfaces. △ Less

Submitted 4 July, 2005; v1 submitted 16 December, 2004; originally announced December 2004.

Comments: To appear in Inventiones Mathematicae. Minor final changes added 4/7 2005

Temperature dependence of microwave and THz dielectric response in Srn+1TinO3n+1 (n=1-4)

Authors: D. Noujni, S. Kamba, A. Pashkin, V. Bovtun, J. Petzelt, A-K. Axelsson, N. McN Alford, P. L. Wise, I. M. Reaney

Abstract: The microwave, near-millimetre and infrared (IR) dielectric response of Srn+1TinO3n+1 (n=1-4) Ruddlesden-Popper homologous series was studied in the temperature range 10 to 300 K. Remarkable softening of the polar optical mode was observed in Sr4Ti3O10 and Sr5Ti4O13 which explains the increase in microwave permittivity and dielectric loss upon cooling. However, both samples have a distinct conte… ▽ More The microwave, near-millimetre and infrared (IR) dielectric response of Srn+1TinO3n+1 (n=1-4) Ruddlesden-Popper homologous series was studied in the temperature range 10 to 300 K. Remarkable softening of the polar optical mode was observed in Sr4Ti3O10 and Sr5Ti4O13 which explains the increase in microwave permittivity and dielectric loss upon cooling. However, both samples have a distinct content of SrTiO3 dispersed between SrO layers. It is proposed therefore that the observed soft mode originates from the SrTiO3 microscopic inclusions. △ Less

Submitted 22 January, 2004; originally announced January 2004.

Comments: The 10th European Meeting on Ferroelectricity, August 2003, Cambridge, UK Ferroelectrics, in press

Journal ref: Integrated Ferroelectrics 62, 199-203 (2004)