Showing 1–50 of 138 results for author: Keller, T

More on Landau's theorem and Conjugacy Classes

Authors: Burcu Çınarcı, Thomas Michael Keller, Attila Maróti, Iulian I. Simion

Abstract: Let $p$ be a prime. We construct a function $f$ on the natural numbers such that $f(x) \to \infty$ as $x \to \infty$ and $k_{p}(G)+k_{p'}(G)\geq f(|G|)$ for all finite groups $G$. Here $k_{p}(G)$ denotes the number of conjugacy classes of nontrivial $p$-elements in $G$ and $k_{p'}(G)$ denotes the number of conjugacy classes of elements of $G$ whose orders are coprime to $p$. This is a variation of… ▽ More Let $p$ be a prime. We construct a function $f$ on the natural numbers such that $f(x) \to \infty$ as $x \to \infty$ and $k_{p}(G)+k_{p'}(G)\geq f(|G|)$ for all finite groups $G$. Here $k_{p}(G)$ denotes the number of conjugacy classes of nontrivial $p$-elements in $G$ and $k_{p'}(G)$ denotes the number of conjugacy classes of elements of $G$ whose orders are coprime to $p$. This is a variation of an old theorem of Landau and is used to prove the following: There exists a number $c$ such that whenever $p$ is a prime and $G$ is a finite group of order divisible by $p$ with $|G|>c$, there exists a factorization $p-1 = ab$ with $a$ and $b$ positive integers such that $k_{p}(G) \geq a$ and $k_{p'}(G) \geq b$ with equalities in both cases if and only if $G=C_p \rtimes C_b$ with $C_G(C_p) = C_p$. △ Less

Submitted 17 June, 2024; originally announced June 2024.

Comments: 13 pages

On the anticyclotomic Iwasawa theory of newforms at Eisenstein primes of semistable reduction

Authors: Timo Keller, Mulun Yin

Abstract: Let $f$ be a newform of weight $k$ and level $N$ with trivial nebentypus. Let $\mathfrak{p}\nmid 2N$ be a maximal prime ideal of the coefficient ring of $f$ such that the self-dual twist of the mod-$\mathfrak{p}$ Galois representation of $f$ is reducible with constituents $φ,ψ$. Denote a decomposition group over the rational prime $p$ below $\mathfrak{p}$ by $G_p$. We remove the condition… ▽ More Let $f$ be a newform of weight $k$ and level $N$ with trivial nebentypus. Let $\mathfrak{p}\nmid 2N$ be a maximal prime ideal of the coefficient ring of $f$ such that the self-dual twist of the mod-$\mathfrak{p}$ Galois representation of $f$ is reducible with constituents $φ,ψ$. Denote a decomposition group over the rational prime $p$ below $\mathfrak{p}$ by $G_p$. We remove the condition $φ|_{G_p} \neq \mathbf{1}, ω$ from [CGLS22], and generalize their results to newforms of arbitrary weights. As a consequence, we prove some Iwasawa main conjectures and get the $p$-part of the strong BSD conjecture for elliptic curves of analytic rank $0$ or $1$ over $\mathbf{Q}$ in this setting. In particular, non-trivial $p$-torsion is allowed in the Mordell--Weil group. Using Hida families, we prove a Iwasawa main conjecture for newforms of weight $2$ of multiplicative reduction at Eisenstein primes. In the above situations, we also get $p$-converse theorems to the theorems of Gross--Zagier--Kolyvagin. The $p$-converse theorems have applications to Goldfeld's conjecture in certain quadratic twist families of elliptic curves having a $3$-isogeny. △ Less

Submitted 20 February, 2024; originally announced February 2024.

Comments: Comments welcome

MSC Class: 11G40 (Primary) 11G05; 11G10; 14G10 (Secondary)

Complete verification of strong BSD for many modular abelian surfaces over $\mathbf{Q}$

Authors: Timo Keller, Michael Stoll

Abstract: We develop the theory and algorithms necessary to be able to verify the strong Birch--Swinnerton-Dyer Conjecture for absolutely simple modular abelian varieties over $\mathbf{Q}$. We apply our methods to all 28 Atkin--Lehner quotients of $X_0(N)$ of genus $2$, all 97 genus $2$ curves from the LMFDB whose Jacobian is of this type and six further curves originally found by Wang. We are able to verif… ▽ More We develop the theory and algorithms necessary to be able to verify the strong Birch--Swinnerton-Dyer Conjecture for absolutely simple modular abelian varieties over $\mathbf{Q}$. We apply our methods to all 28 Atkin--Lehner quotients of $X_0(N)$ of genus $2$, all 97 genus $2$ curves from the LMFDB whose Jacobian is of this type and six further curves originally found by Wang. We are able to verify the strong BSD Conjecture unconditionally and exactly in all these cases; this is the first time that strong BSD has been confirmed for absolutely simple abelian varieties of dimension at least $2$. We also give an example where we verify that the order of the Tate--Shafarevich group is $7^2$ and agrees with the order predicted by the BSD Conjecture. △ Less

Submitted 12 December, 2023; originally announced December 2023.

Comments: 94 pages

MSC Class: 11G40 (Primary) 11G10; 14G05 (Secondary)

Image segmentation with traveling waves in an exactly solvable recurrent neural network

Authors: Luisa H. B. Liboni, Roberto C. Budzinski, Alexandra N. Busch, Sindy Löwe, Thomas A. Keller, Max Welling, Lyle E. Muller

Abstract: We study image segmentation using spatiotemporal dynamics in a recurrent neural network where the state of each unit is given by a complex number. We show that this network generates sophisticated spatiotemporal dynamics that can effectively divide an image into groups according to a scene's structural characteristics. Using an exact solution of the recurrent network's dynamics, we present a preci… ▽ More We study image segmentation using spatiotemporal dynamics in a recurrent neural network where the state of each unit is given by a complex number. We show that this network generates sophisticated spatiotemporal dynamics that can effectively divide an image into groups according to a scene's structural characteristics. Using an exact solution of the recurrent network's dynamics, we present a precise description of the mechanism underlying object segmentation in this network, providing a clear mathematical interpretation of how the network performs this task. We then demonstrate a simple algorithm for object segmentation that generalizes across inputs ranging from simple geometric objects in grayscale images to natural images. Object segmentation across all images is accomplished with one recurrent neural network that has a single, fixed set of weights. This demonstrates the expressive potential of recurrent neural networks when constructed using a mathematical approach that brings together their structure, dynamics, and computation. △ Less

Submitted 28 November, 2023; originally announced November 2023.

Bounding the number of p'-degrees from below

Authors: Thomas Keller, Yong Yang

Abstract: Let $G$ be a finite group of order divisible by a prime $p$ and let $P\in\Syl_p(G)$. We prove a recent conjecture by Hung stating that $|\Irr_{p'}(G)|\geq \frac{\exp(P/P')-1}{p-1}+2\sqrt{p-1}-1.$ Let $a\geq 2$ be an integer and suppose that $p^a$ does not exceed the exponent of the center of $P$. We then also show that the number of conjugacy classes of elements of $G$ for which $p^a$ is the exact… ▽ More Let $G$ be a finite group of order divisible by a prime $p$ and let $P\in\Syl_p(G)$. We prove a recent conjecture by Hung stating that $|\Irr_{p'}(G)|\geq \frac{\exp(P/P')-1}{p-1}+2\sqrt{p-1}-1.$ Let $a\geq 2$ be an integer and suppose that $p^a$ does not exceed the exponent of the center of $P$. We then also show that the number of conjugacy classes of elements of $G$ for which $p^a$ is the exact $p$-part of their order is at least $p^{a-1}$. △ Less

Submitted 19 November, 2023; originally announced November 2023.

MSC Class: 20C15; 20C10

Towards a classification of isolated $j$-invariants

Authors: Abbey Bourdon, Sachi Hashimoto, Timo Keller, Zev Klagsbrun, David Lowry-Duda, Travis Morrison, Filip Najman, Himanshu Shukla

Abstract: We develop an algorithm to test whether a non-CM elliptic curve $E/\mathbb{Q}$ gives rise to an isolated point of any degree on any modular curve of the form $X_1(N)$. This builds on prior work of Zywina which gives a method for computing the image of the adelic Galois representation associated to $E$. Running this algorithm on all elliptic curves presently in the $L$-functions and Modular Forms D… ▽ More We develop an algorithm to test whether a non-CM elliptic curve $E/\mathbb{Q}$ gives rise to an isolated point of any degree on any modular curve of the form $X_1(N)$. This builds on prior work of Zywina which gives a method for computing the image of the adelic Galois representation associated to $E$. Running this algorithm on all elliptic curves presently in the $L$-functions and Modular Forms Database and the Stein-Watkins Database gives strong evidence for the conjecture that $E$ gives rise to an isolated point on $X_1(N)$ if and only if $j(E)=-140625/8, -9317,$ $351/4$, or $-162677523113838677$. △ Less

Submitted 26 February, 2024; v1 submitted 13 November, 2023; originally announced November 2023.

Comments: With an appendix by Maarten Derickx and Mark van Hoeij, to appear in Math. Comp

A new lower bound for the number of conjugacy classes

Authors: Burcu Çınarcı, Thomas Michael Keller

Abstract: In 2003, Héthelyi and Külshammer proposed that if $G$ is a finite group and $p$ is a prime dividing the group order, then $k(G)\geq 2\sqrt{p-1}$, and they proved this conjecture for solvable $G$ and showed that it is sharp for those primes $p$ for which $\sqrt{p-1}$ is an integer. This initiated a flurry of activity, leading to many generalizations and variations of the result; in particular, toda… ▽ More In 2003, Héthelyi and Külshammer proposed that if $G$ is a finite group and $p$ is a prime dividing the group order, then $k(G)\geq 2\sqrt{p-1}$, and they proved this conjecture for solvable $G$ and showed that it is sharp for those primes $p$ for which $\sqrt{p-1}$ is an integer. This initiated a flurry of activity, leading to many generalizations and variations of the result; in particular, today the conjecture is known to be true for all finite groups. In this note, we put forward a natural new and stronger conjecture, which is sharp for all primes $p$, and we prove it for solvable groups, and when $p$ is large, also for arbitrary groups. △ Less

Submitted 13 November, 2023; v1 submitted 13 October, 2023; originally announced October 2023.

Comments: A hypothesis for Theorem E is added that was inadvertently omitted in the previous version. Some other small changes have been made

Deep Generative Models of Music Expectation

Authors: Ninon Lizé Masclef, T. Anderson Keller

Abstract: A prominent theory of affective response to music revolves around the concepts of surprisal and expectation. In prior work, this idea has been operationalized in the form of probabilistic models of music which allow for precise computation of song (or note-by-note) probabilities, conditioned on a 'training set' of prior musical or cultural experiences. To date, however, these models have been limi… ▽ More A prominent theory of affective response to music revolves around the concepts of surprisal and expectation. In prior work, this idea has been operationalized in the form of probabilistic models of music which allow for precise computation of song (or note-by-note) probabilities, conditioned on a 'training set' of prior musical or cultural experiences. To date, however, these models have been limited to compute exact probabilities through hand-crafted features or restricted to linear models which are likely not sufficient to represent the complex conditional distributions present in music. In this work, we propose to use modern deep probabilistic generative models in the form of a Diffusion Model to compute an approximate likelihood of a musical input sequence. Unlike prior work, such a generative model parameterized by deep neural networks is able to learn complex non-linear features directly from a training set itself. In doing so, we expect to find that such models are able to more accurately represent the 'surprisal' of music for human listeners. From the literature, it is known that there is an inverted U-shaped relationship between surprisal and the amount human subjects 'like' a given song. In this work we show that pre-trained diffusion models indeed yield musical surprisal values which exhibit a negative quadratic relationship with measured subject 'liking' ratings, and that the quality of this relationship is competitive with state of the art methods such as IDyOM. We therefore present this model a preliminary step in develo** modern deep generative models of music expectation and subjective likability. △ Less

Submitted 5 October, 2023; originally announced October 2023.

Flow Factorized Representation Learning

Authors: Yue Song, T. Anderson Keller, Nicu Sebe, Max Welling

Abstract: A prominent goal of representation learning research is to achieve representations which are factorized in a useful manner with respect to the ground truth factors of variation. The fields of disentangled and equivariant representation learning have approached this ideal from a range of complimentary perspectives; however, to date, most approaches have proven to either be ill-specified or insuffic… ▽ More A prominent goal of representation learning research is to achieve representations which are factorized in a useful manner with respect to the ground truth factors of variation. The fields of disentangled and equivariant representation learning have approached this ideal from a range of complimentary perspectives; however, to date, most approaches have proven to either be ill-specified or insufficiently flexible to effectively separate all realistic factors of interest in a learned latent space. In this work, we propose an alternative viewpoint on such structured representation learning which we call Flow Factorized Representation Learning, and demonstrate it to learn both more efficient and more usefully structured representations than existing frameworks. Specifically, we introduce a generative model which specifies a distinct set of latent probability paths that define different input transformations. Each latent flow is generated by the gradient field of a learned potential following dynamic optimal transport. Our novel setup brings new understandings to both \textit{disentanglement} and \textit{equivariance}. We show that our model achieves higher likelihoods on standard representation learning benchmarks while simultaneously being closer to approximately equivariant models. Furthermore, we demonstrate that the transformations learned by our model are flexibly composable and can also extrapolate to new data, implying a degree of robustness and generalizability approaching the ultimate goal of usefully factorized representation learning. △ Less

Submitted 22 September, 2023; originally announced September 2023.

Comments: NeurIPS23

Traveling Waves Encode the Recent Past and Enhance Sequence Learning

Authors: T. Anderson Keller, Lyle Muller, Terrence Sejnowski, Max Welling

Abstract: Traveling waves of neural activity have been observed throughout the brain at a diversity of regions and scales; however, their precise computational role is still debated. One physically inspired hypothesis suggests that the cortical sheet may act like a wave-propagating system capable of invertibly storing a short-term memory of sequential stimuli through induced waves traveling across the corti… ▽ More Traveling waves of neural activity have been observed throughout the brain at a diversity of regions and scales; however, their precise computational role is still debated. One physically inspired hypothesis suggests that the cortical sheet may act like a wave-propagating system capable of invertibly storing a short-term memory of sequential stimuli through induced waves traveling across the cortical surface, and indeed many experimental results from neuroscience correlate wave activity with memory tasks. To date, however, the computational implications of this idea have remained hypothetical due to the lack of a simple recurrent neural network architecture capable of exhibiting such waves. In this work, we introduce a model to fill this gap, which we denote the Wave-RNN (wRNN), and demonstrate how such an architecture indeed efficiently encodes the recent past through a suite of synthetic memory tasks where wRNNs learn faster and reach significantly lower error than wave-free counterparts. We further explore the implications of this memory storage system on more complex sequence modeling tasks such as sequential image classification and find that wave-based models not only again outperform comparable wave-free RNNs while using significantly fewer parameters, but additionally perform comparably to more complex gated architectures such as LSTMs and GRUs. △ Less

Submitted 14 March, 2024; v1 submitted 3 September, 2023; originally announced September 2023.

Prime divisors and the number of conjugacy classes of finite groups

Authors: Thomas Michael Keller, Alexander Moretó

Abstract: We prove that there exists a universal constant $D$ such that if $p$ is a prime divisor of the index of the Fitting subgroup of a finite group $G$, then the number of conjugacy classes of G is at least $Dp/log_2 p$. We conjecture that we can take $D=1$ and prove that for solvable groups, we can take $D=1/3$. We prove that there exists a universal constant $D$ such that if $p$ is a prime divisor of the index of the Fitting subgroup of a finite group $G$, then the number of conjugacy classes of G is at least $Dp/log_2 p$. We conjecture that we can take $D=1$ and prove that for solvable groups, we can take $D=1/3$. △ Less

Submitted 17 July, 2023; v1 submitted 10 July, 2023; originally announced July 2023.

Comments: To appear in Math. Proc. Cambridge Philos. Soc

DUET: 2D Structured and Approximately Equivariant Representations

Authors: Xavier Suau, Federico Danieli, T. Anderson Keller, Arno Blaas, Chen Huang, Jason Ramapuram, Dan Busbridge, Luca Zappella

Abstract: Multiview Self-Supervised Learning (MSSL) is based on learning invariances with respect to a set of input transformations. However, invariance partially or totally removes transformation-related information from the representations, which might harm performance for specific downstream tasks that require such information. We propose 2D strUctured and EquivarianT representations (coined DUET), which… ▽ More Multiview Self-Supervised Learning (MSSL) is based on learning invariances with respect to a set of input transformations. However, invariance partially or totally removes transformation-related information from the representations, which might harm performance for specific downstream tasks that require such information. We propose 2D strUctured and EquivarianT representations (coined DUET), which are 2d representations organized in a matrix structure, and equivariant with respect to transformations acting on the input data. DUET representations maintain information about an input transformation, while remaining semantically expressive. Compared to SimCLR (Chen et al., 2020) (unstructured and invariant) and ESSL (Dangovski et al., 2022) (unstructured and equivariant), the structured and equivariant nature of DUET representations enables controlled generation with lower reconstruction error, while controllability is not possible with SimCLR or ESSL. DUET also achieves higher accuracy for several discriminative tasks, and improves transfer learning. △ Less

Submitted 17 November, 2023; v1 submitted 28 June, 2023; originally announced June 2023.

Comments: Accepted at ICML 2023

Latent Traversals in Generative Models as Potential Flows

Authors: Yue Song, T. Anderson Keller, Nicu Sebe, Max Welling

Abstract: Despite the significant recent progress in deep generative models, the underlying structure of their latent spaces is still poorly understood, thereby making the task of performing semantically meaningful latent traversals an open research challenge. Most prior work has aimed to solve this challenge by modeling latent structures linearly, and finding corresponding linear directions which result in… ▽ More Despite the significant recent progress in deep generative models, the underlying structure of their latent spaces is still poorly understood, thereby making the task of performing semantically meaningful latent traversals an open research challenge. Most prior work has aimed to solve this challenge by modeling latent structures linearly, and finding corresponding linear directions which result in `disentangled' generations. In this work, we instead propose to model latent structures with a learned dynamic potential landscape, thereby performing latent traversals as the flow of samples down the landscape's gradient. Inspired by physics, optimal transport, and neuroscience, these potential landscapes are learned as physically realistic partial differential equations, thereby allowing them to flexibly vary over both space and time. To achieve disentanglement, multiple potentials are learned simultaneously, and are constrained by a classifier to be distinct and semantically self-consistent. Experimentally, we demonstrate that our method achieves both more qualitatively and quantitatively disentangled trajectories than state-of-the-art baselines. Further, we demonstrate that our method can be integrated as a regularization term during training, thereby acting as an inductive bias towards the learning of structured representations, ultimately improving model likelihood on similarly structured data. △ Less

Submitted 1 July, 2023; v1 submitted 25 April, 2023; originally announced April 2023.

Comments: ICML 2023

Computing quadratic points on modular curves $X_0(N)$

Authors: Nikola Adžaga, Timo Keller, Philippe Michaud-Jacobs, Filip Najman, Ekin Ozman, Borna Vukorepa

Abstract: In this paper we improve on existing methods to compute quadratic points on modular curves and apply them to successfully find all the quadratic points on all modular curves $X_0(N)$ of genus up to $8$, and genus up to $10$ with $N$ prime, for which they were previously unknown. The values of $N$ we consider are contained in the set \[ \mathcal{L}=\{58, 68, 74, 76, 80, 85, 97, 98, 100, 103, 107, 1… ▽ More In this paper we improve on existing methods to compute quadratic points on modular curves and apply them to successfully find all the quadratic points on all modular curves $X_0(N)$ of genus up to $8$, and genus up to $10$ with $N$ prime, for which they were previously unknown. The values of $N$ we consider are contained in the set \[ \mathcal{L}=\{58, 68, 74, 76, 80, 85, 97, 98, 100, 103, 107, 109, 113, 121, 127 \}.\] We obtain that all the non-cuspidal quadratic points on $X_0(N)$ for $N\in \mathcal{L}$ are CM points, except for one pair of Galois conjugate points on $X_0(103)$ defined over $\mathbb{Q}(\sqrt{2885})$. We also compute the $j$-invariants of the elliptic curves parametrised by these points, and for the CM points determine their geometric endomorphism rings. △ Less

Submitted 25 August, 2023; v1 submitted 22 March, 2023; originally announced March 2023.

Comments: Minor corrections, to appear in Mathematics of Computation. (25 pages, 16 tables)

MSC Class: 11G05 (Primary); 14G05; 11G18 (Secondary)

Fermionization of a Few-Body Bose System Immersed into a Bose-Einstein Condensate

Authors: Tim Keller, Thomás Fogarty, Thomas Busch

Abstract: We study the recently introduced self-pinning transition [Phys. Rev. Lett. 128, 053401 (2022)] in a quasi-one-dimensional two-component quantum gas in the case where the component immersed into the Bose-Einstein condensate has a finite intraspecies interaction strength. As a result of the matter-wave backaction, the fermionization in the limit of infinite intraspecies repulsion occurs via a first-… ▽ More We study the recently introduced self-pinning transition [Phys. Rev. Lett. 128, 053401 (2022)] in a quasi-one-dimensional two-component quantum gas in the case where the component immersed into the Bose-Einstein condensate has a finite intraspecies interaction strength. As a result of the matter-wave backaction, the fermionization in the limit of infinite intraspecies repulsion occurs via a first-order phase transition to the self-pinned state, which is in contrast to the asymptotic behavior in static trap** potentials. The system also exhibits an additional superfluid state for the immersed component if the interspecies interaction is able to overcome the intraspecies repulsion. We approximate the superfluid state in an analytical model and derive an expression for the phase transition line that coincides with well-known phase separation criteria in binary Bose systems. The full phase diagram of the system is mapped out numerically for the case of two and three atoms in the immersed component. △ Less

Submitted 30 June, 2023; v1 submitted 2 February, 2023; originally announced February 2023.

Comments: 17 pages, 4 figures

Journal ref: SciPost Phys. 15, 095 (2023)

Specialization of Mordell-Weil ranks of abelian schemes over surfaces to curves

Authors: Timo Keller

Abstract: Using the Shioda-Tate theorem and an adaptation of Silverman's specialization theorem, we reduce the specialization of Mordell-Weil ranks for abelian varieties over fields finitely generated over infinite finitely generated fields $k$ to the the specialization theorem for Néron-Severi ranks recently proved by Ambrosi in positive characteristic. More precisely, we prove that after a blow-up of the… ▽ More Using the Shioda-Tate theorem and an adaptation of Silverman's specialization theorem, we reduce the specialization of Mordell-Weil ranks for abelian varieties over fields finitely generated over infinite finitely generated fields $k$ to the the specialization theorem for Néron-Severi ranks recently proved by Ambrosi in positive characteristic. More precisely, we prove that after a blow-up of the base surface $S$, for all vertical curves $S_x$ of a fibration $S \to U \subseteq \mathbf{P}^1_k$ with $x$ from the complement of a sparse subset of $|U|$, the Mordell-Weil rank of an abelian scheme over $S$ stays the same when restricted to $S_x$. △ Less

Submitted 30 January, 2023; originally announced January 2023.

Comments: accepted for publication in International Journal of Number Theory

MSC Class: 11G10 (Primary) 11G05; 11G35 (Secondary)

The Prime Graphs of Groups With Arithmetically Small Composition Factors

Authors: Timothy J. Edwards, Thomas Michael Keller, Ryan M. Pesak, Karthik Sellakumaran Latha

Abstract: We continue the study of prime graphs of finite groups, also known as Gruenberg-Kegel graphs. The vertices of the prime graph of a finite group are the prime divisors of the group order, and two vertices $p$ and $q$ are connected by an edge if and only if there is an element of order $pq$ in the group. Prime graphs of solvable groups have been characterized in graph theoretical terms only, as have… ▽ More We continue the study of prime graphs of finite groups, also known as Gruenberg-Kegel graphs. The vertices of the prime graph of a finite group are the prime divisors of the group order, and two vertices $p$ and $q$ are connected by an edge if and only if there is an element of order $pq$ in the group. Prime graphs of solvable groups have been characterized in graph theoretical terms only, as have been the prime graphs of groups whose only nonsolvable composition factor is $A_5$. In this paper we classify the prime graphs of all groups whose composition factors have arithmetically small orders, that is, have no more than three prime divisors in their orders. We find that all such graphs have $3$-colorable complements, and we provide full characterizations of the prime graphs of such groups based on the exact type and multiplicity of the nonabelian composition factors of the group. △ Less

Submitted 6 December, 2022; originally announced December 2022.

Comments: 33 pages

MSC Class: 20D60; 05C25

A classification of prime graphs of pseudo-solvable groups

Authors: Ziyu Huang, Thomas Michael Keller, Shane Kissinger, Wen Plotnick, Maya Roma, Yong Yang

Abstract: The prime graph $Γ(G)$ of a finite group $G$ (also known as the Gruenberg-Kegel graph) has as its vertices the prime divisors of $|G|$, and $p\text-q$ is an edge in $Γ(G)$ if and only if $G$ has an element of order $pq$. Since their inception in the 1970s these graphs have been studied extensively; however, completely classifying the possible prime graphs for larger families of groups remains a di… ▽ More The prime graph $Γ(G)$ of a finite group $G$ (also known as the Gruenberg-Kegel graph) has as its vertices the prime divisors of $|G|$, and $p\text-q$ is an edge in $Γ(G)$ if and only if $G$ has an element of order $pq$. Since their inception in the 1970s these graphs have been studied extensively; however, completely classifying the possible prime graphs for larger families of groups remains a difficult problem. For solvable groups such a classification was found in 2015. In this paper we go beyond solvable groups for the first time and characterize prime graphs of a more general class of groups we call pseudo-solvable. These are groups whose composition factors are either cyclic or $A_5$. The classification is based on two conditions: the vertices $\{2,3,5\}$ form a triangle in $\overlineΓ(G)$ or $\{p,3,5\}$ form a triangle for some prime $p\neq 2$. △ Less

Submitted 28 November, 2022; originally announced November 2022.

MSC Class: 20D60; 05C25

How nanoporous silicon-polypyrrole hybrids flex their muscles in aqueous electrolytes: In operando high-resolution x-ray diffraction and electron tomography-based micromechanical computer simulations

Authors: Manuel Brinker, Marc Thelen, Manfred May, Dagmar Rings, Tobias Krekeler, Pirmin Lakner, Thomas F. Keller, Florian Bertram, Norbert Huber, Patrick Huber

Abstract: Macroscopic strain experiments revealed that Si crystals traversed by parallel, channel-like nanopores functionalized with the muscle polymer polypyrrole exhibit large and reversible electrochemo-mechanical actuation in aqueous electrolytes. On the microscopical level this system still bears open questions, as to how the electrochemical expansion and contraction of PPy acts on to np-Si pore walls… ▽ More Macroscopic strain experiments revealed that Si crystals traversed by parallel, channel-like nanopores functionalized with the muscle polymer polypyrrole exhibit large and reversible electrochemo-mechanical actuation in aqueous electrolytes. On the microscopical level this system still bears open questions, as to how the electrochemical expansion and contraction of PPy acts on to np-Si pore walls and how the collective motorics of the pore array emerges from the single-nanopore behavior. An analysis of in operando X-ray diffraction experiments with micromechanical finite element simulations, based on a 3D reconstruction of the nanoporous medium by TEM tomography, shows that the in-plane mechanical response is dominantly isotropic despite the anisotropic elasticity of the single crystalline host matrix. However, the structural anisotropy originating from the parallel alignment of the nanopores lead to significant differences between the in- and out-of-plane electromechanical response. This response is not describable by a simple 2D arrangement of parallel cylindrical channels. Rather, the simulations highlight that the dendritic shape of the Si pore walls, including pore connections between the main channels, cause complex, inhomogeneous stress-strain fields in the crystalline host. Time-dependent X-ray scattering on the dynamics of the actuator properties hint towards the importance of diffusion limitations, plastic deformation and creep in the nanoconfined polymer upon (counter-)ion adsorption and desorption, the very pore-scale processes causing the macroscopic electroactuation. From a more general perspective, our study demonstrates that the combination of TEM tomography-based micromechanical modeling with high-resolution X-ray scattering experiments provides a powerful approach for in operando analysis of nanoporous composites from the single-nanopore up to the porous-medium scale. △ Less

Submitted 28 November, 2022; originally announced November 2022.

Comments: Supplementary see ancillary file. 20 pages, 11 figures

Journal ref: Phys. Rev. Materials 6, 116002 (2022)

Homomorphic Self-Supervised Learning

Authors: T. Anderson Keller, Xavier Suau, Luca Zappella

Abstract: In this work, we observe that many existing self-supervised learning algorithms can be both unified and generalized when seen through the lens of equivariant representations. Specifically, we introduce a general framework we call Homomorphic Self-Supervised Learning, and theoretically show how it may subsume the use of input-augmentations provided an augmentation-homomorphic feature extractor. We… ▽ More In this work, we observe that many existing self-supervised learning algorithms can be both unified and generalized when seen through the lens of equivariant representations. Specifically, we introduce a general framework we call Homomorphic Self-Supervised Learning, and theoretically show how it may subsume the use of input-augmentations provided an augmentation-homomorphic feature extractor. We validate this theory experimentally for simple augmentations, demonstrate how the framework fails when representational structure is removed, and further empirically explore how the parameters of this framework relate to those of traditional augmentation-based self-supervised learning. We conclude with a discussion of the potential benefits afforded by this new perspective on self-supervised learning. △ Less

Submitted 15 November, 2022; originally announced November 2022.

New bounds for numbers of primes in element orders of finite groups

Authors: Chiara Bellotti, Thomas Michael Keller, Timothy S. Trudgian

Abstract: Let $ρ(n)$ denote the maximal number of different primes that may occur in the order of a finite solvable group $G$, all elements of which have orders divisible by at most $n$ distinct primes. We show that $ρ(n)\leq 5n$ for all $n\geq 1$. As an application, we improve on a recent bound by Hung and Yang for arbitrary finite groups. Let $ρ(n)$ denote the maximal number of different primes that may occur in the order of a finite solvable group $G$, all elements of which have orders divisible by at most $n$ distinct primes. We show that $ρ(n)\leq 5n$ for all $n\geq 1$. As an application, we improve on a recent bound by Hung and Yang for arbitrary finite groups. △ Less

Submitted 10 November, 2022; originally announced November 2022.

Comments: 6 pages

MSC Class: 20D60; 20E34; 20C15; 11N05

On the Generation, Structure, and Symmetries of Minimal Prime Graphs

Authors: Ziyu Huang, Thomas Michael Keller, Shane Kissinger, Wen Plotnick, Maya Roma

Abstract: In this paper we continue the study of prime graphs of finite solvable groups. The prime graph, or Gruenberg-Kegel graph, of a finite group G has vertices consisting of the prime divisors of the order of G and an edge from primes p to q if and only if G contains an element of order pq. Since the discovery of a simple, purely graph theoretical characterization of the prime graphs of solvable groups… ▽ More In this paper we continue the study of prime graphs of finite solvable groups. The prime graph, or Gruenberg-Kegel graph, of a finite group G has vertices consisting of the prime divisors of the order of G and an edge from primes p to q if and only if G contains an element of order pq. Since the discovery of a simple, purely graph theoretical characterization of the prime graphs of solvable groups in 2015 these graphs have been studied in more detail from a graph theoretic angle. In this paper we explore several new aspects of these graphs. We characterize regular reseminant graphs and study the automorphisms of reseminant graphs for arbitrary base graphs. We then study minimal prime graphs on larger vertex sets by a novel regular graph construction for base graphs and by proving results on prime graph properties under graph products. Lastly, we present the first new way, different from vertex duplication, to obtain a new minimal prime graph from a given minimal prime graph. △ Less

Submitted 24 October, 2022; originally announced October 2022.

Comments: 22 pages, 7 figures

Extending MIEZE spectroscopy towards thermal wavelengths

Authors: Johanna K. Jochum, Christian Franz, Thomas Keller, Christian Pfleiderer

Abstract: We propose a Modulation of intensity with zero effort (MIEZE) set-up for high-resolution neutron spectroscopy at momentum transfers up to 3Å$^{-1}$,energy transfers up to ~ 20 meV, and an energy resolution in the $μ$eV-range using both thermal and cold neutrons. MIEZE has two prominent advantages compared to classical neutron spin-echo. The first one is the possibility to investigate spin-depolari… ▽ More We propose a Modulation of intensity with zero effort (MIEZE) set-up for high-resolution neutron spectroscopy at momentum transfers up to 3Å$^{-1}$,energy transfers up to ~ 20 meV, and an energy resolution in the $μ$eV-range using both thermal and cold neutrons. MIEZE has two prominent advantages compared to classical neutron spin-echo. The first one is the possibility to investigate spin-depolarizing samples or samples in strong magnetic fields without loss of signal amplitude and intensity. This allows for the study of spin fluctuations in ferromagnets, and facilitates the study of samples with strong spin-incoherent scattering. The second advantage is that multi-analyzer setups can be implemented with comparatively small effort. The use of thermal neutrons increases the range of validity of the spin-echo approximation towards shorter spin-echo times. In turn, the thermal MIEZE option for greater ranges (TIGER) closes the gap between classical neutron spin-echo spectroscopy and conventional high-resolution neutron spectroscopy techniques such as triple-axis, time-of-flight, and back-scattering. To illustrate the feasibility of TIGER we present the details of an implementation at the beamline RESEDA at FRM II by means of an additional velocity selector, polarizer and analyzer. △ Less

Submitted 18 November, 2022; v1 submitted 19 July, 2022; originally announced July 2022.

Journal ref: J. Appl. Cryst. (2022). 55

Critical magnetic fluctuations in the layered ruthenates Ca$_2$RuO$_4$ and Ca$_3$Ru$_2$O$_7$

Authors: H. Trepka, T. Keller, M. Krautloher, J. Xu, K. Habicht, M. Böhm, B. Keimer, M. Hepting

Abstract: Materials realizing the XY model in two dimensions (2D) are sparse. Here we use neutron triple-axis spectroscopy to investigate the critical static and dynamical magnetic fluctuations in the square-lattice antiferromagnets Ca$_{2}$RuO$_{4}$ and Ca$_{3}$Ru$_2$O$_{7}$. We probe the temperature-dependence of the antiferromagnetic (AFM) Bragg-intensity, the $Q$-width, the amplitude, and the energy-wid… ▽ More Materials realizing the XY model in two dimensions (2D) are sparse. Here we use neutron triple-axis spectroscopy to investigate the critical static and dynamical magnetic fluctuations in the square-lattice antiferromagnets Ca$_{2}$RuO$_{4}$ and Ca$_{3}$Ru$_2$O$_{7}$. We probe the temperature-dependence of the antiferromagnetic (AFM) Bragg-intensity, the $Q$-width, the amplitude, and the energy-width of the magnetic diffuse scattering in vicinity to the Néel temperature $T_N$ to determine the critical behavior of the magnetic order parameter $M$, correlation length $ξ$, susceptibility $χ$, and the characteristic energy $Γ$ with the corresponding critical exponents $β$, $ν$, $γ$, and $z$, respectively. We find that the critical behaviors of the single-layer compound Ca$_{2}$RuO$_{4}$ follow universal scaling laws that are compatible with predictions of the 2D-XY model. The bilayer compound Ca$_{3}$Ru$_2$O$_{7}$ is only partly consistent with the 2D-XY theory and best described by the three-dimensional Ising (3D-I) model, which is likely a consequence of the intra-bilayer exchange interactions in combination with an orthorhombic single-ion anisotropy. Hence, our results suggest that layered ruthenates are promising solid-state platforms for research on the 2D-XY model and the effects of 3D interactions and additional spin-space anisotropies on the magnetic fluctuations. △ Less

Submitted 3 June, 2022; originally announced June 2022.

Comments: 19 pages, 17 figures

Journal ref: Phys. Rev. Research 4, 023181 (2022)

On the $p$-torsion of the Tate-Shafarevich group of abelian varieties over higher dimensional bases over finite fields

Authors: Timo Keller

Abstract: We prove a finiteness theorem for the first flat cohomology group of finite flat group schemes over integral normal proper varieties over finite fields. As a consequence, we can prove the invariance of the finiteness of the Tate-Shafarevich group of Abelian schemes over higher dimensional bases under isogenies and alterations over/of such bases for the $p$-part. Along the way, we generalize previo… ▽ More We prove a finiteness theorem for the first flat cohomology group of finite flat group schemes over integral normal proper varieties over finite fields. As a consequence, we can prove the invariance of the finiteness of the Tate-Shafarevich group of Abelian schemes over higher dimensional bases under isogenies and alterations over/of such bases for the $p$-part. Along the way, we generalize previous results on the Tate-Shafarevich group in this situation. △ Less

Submitted 11 March, 2022; originally announced March 2022.

Comments: 14 pages, accepted for publication in Journal de Théorie des Nombres de Bordeaux

MSC Class: 14F20 (Primary) 11G10; 14K15 (Secondary)

Rational points on hyperelliptic Atkin-Lehner quotients of modular curves and their coverings

Authors: Nikola Adžaga, Shiva Chidambaram, Timo Keller, Oana Padurariu

Abstract: We complete the computation of all $\mathbb{Q}$-rational points on all the $64$ maximal Atkin-Lehner quotients $X_0(N)^*$ such that the quotient is hyperelliptic. To achieve this, we use a combination of various methods, namely the classical Chabauty--Coleman, elliptic curve Chabauty, quadratic Chabauty, and the bielliptic quadratic Chabauty method combined with the Mordell-Weil sieve. Additionall… ▽ More We complete the computation of all $\mathbb{Q}$-rational points on all the $64$ maximal Atkin-Lehner quotients $X_0(N)^*$ such that the quotient is hyperelliptic. To achieve this, we use a combination of various methods, namely the classical Chabauty--Coleman, elliptic curve Chabauty, quadratic Chabauty, and the bielliptic quadratic Chabauty method combined with the Mordell-Weil sieve. Additionally, for square-free levels $N$, we classify all $\mathbb{Q}$-rational points as cusps, CM points (including their CM field and $j$-invariants) and exceptional ones. We further indicate how to use this to compute the $\mathbb{Q}$-rational points on all of their modular coverings. △ Less

Submitted 23 September, 2022; v1 submitted 10 March, 2022; originally announced March 2022.

Comments: 29 pages, 6 tables; accepted by ANTS-XV and hence for publication in Research in Number Theory

MSC Class: 14G05 (Primary) 11G30; 11G18 (Secondary)

Modeling Category-Selective Cortical Regions with Topographic Variational Autoencoders

Authors: T. Anderson Keller, Qinghe Gao, Max Welling

Abstract: Category-selectivity in the brain describes the observation that certain spatially localized areas of the cerebral cortex tend to respond robustly and selectively to stimuli from specific limited categories. One of the most well known examples of category-selectivity is the Fusiform Face Area (FFA), an area of the inferior temporal cortex in primates which responds preferentially to images of face… ▽ More Category-selectivity in the brain describes the observation that certain spatially localized areas of the cerebral cortex tend to respond robustly and selectively to stimuli from specific limited categories. One of the most well known examples of category-selectivity is the Fusiform Face Area (FFA), an area of the inferior temporal cortex in primates which responds preferentially to images of faces when compared with objects or other generic stimuli. In this work, we leverage the newly introduced Topographic Variational Autoencoder to model the emergence of such localized category-selectivity in an unsupervised manner. Experimentally, we demonstrate our model yields spatially dense neural clusters selective to faces, bodies, and places through visualized maps of Cohen's d metric. We compare our model with related supervised approaches, namely the Topographic Deep Artificial Neural Network (TDANN) of Lee et al., and discuss both theoretical and empirical similarities. Finally, we show preliminary results suggesting that our model yields a nested spatial hierarchy of increasingly abstract categories, analogous to observations from the human ventral temporal cortex. △ Less

Submitted 18 December, 2021; v1 submitted 25 October, 2021; originally announced October 2021.

Self-Pinning Transition of a Tonks-Girardeau Gas in a Bose-Einstein Condensate

Authors: Tim Keller, Thomás Fogarty, Thomas Busch

Abstract: We show that a Tonks-Girardeau (TG) gas that is immersed in a Bose-Einstein condensate can undergo a transition to a crystal-like Mott state with regular spacing between the atoms without any externally imposed lattice potential. We characterize this phase transition as a function of the interspecies interaction and temperature of the TG gas, and show how it can be measured via accessible observab… ▽ More We show that a Tonks-Girardeau (TG) gas that is immersed in a Bose-Einstein condensate can undergo a transition to a crystal-like Mott state with regular spacing between the atoms without any externally imposed lattice potential. We characterize this phase transition as a function of the interspecies interaction and temperature of the TG gas, and show how it can be measured via accessible observables in cold atom experiments. We also develop an effective model that accurately describes the system in the pinned insulator state and which allows us to derive the critical temperature of the transition. △ Less

Submitted 1 February, 2022; v1 submitted 2 September, 2021; originally announced September 2021.

Comments: 5 pages, 3 figures; 5 supplementary pages, 3 supplementary figures

Journal ref: Phys. Rev. Lett. 128, 053401 (2022)

Topographic VAEs learn Equivariant Capsules

Authors: T. Anderson Keller, Max Welling

Abstract: In this work we seek to bridge the concepts of topographic organization and equivariance in neural networks. To accomplish this, we introduce the Topographic VAE: a novel method for efficiently training deep generative models with topographically organized latent variables. We show that such a model indeed learns to organize its activations according to salient characteristics such as digit class,… ▽ More In this work we seek to bridge the concepts of topographic organization and equivariance in neural networks. To accomplish this, we introduce the Topographic VAE: a novel method for efficiently training deep generative models with topographically organized latent variables. We show that such a model indeed learns to organize its activations according to salient characteristics such as digit class, width, and style on MNIST. Furthermore, through topographic organization over time (i.e. temporal coherence), we demonstrate how predefined latent space transformation operators can be encouraged for observed transformed input sequences -- a primitive form of unsupervised learned equivariance. We demonstrate that this model successfully learns sets of approximately equivariant features (i.e. "capsules") directly from sequences and achieves higher likelihood on correspondingly transforming test sequences. Equivariance is verified quantitatively by measuring the approximate commutativity of the inference network and the sequence transformations. Finally, we demonstrate approximate equivariance to complex transformations, expanding upon the capabilities of existing group equivariant neural networks. △ Less

Submitted 9 January, 2022; v1 submitted 3 September, 2021; originally announced September 2021.

Influence of Contacts and Applied Voltage on a Structure of a Single GaN Nanowire

Authors: Sergey Lazarev, Luca Gelisio, Young Yong Kim, Bi Zhaoxia, Ali Nowzari, Ivan A. Zaluzhnyy, Ruslan Khubbutdinov, Dmitry Dzhigaev, Arno Jeromin, Thomas Keller, Michael Sprung, Anders Mikkelsen, Lars Samuelson, Ivan A. Vartanyants

Abstract: Semiconductor nanowires (NWs) have a broad range of applications for nano- and optoelectronics. The strain field of gallium nitride (GaN) NWs could be significantly changed when contacts are applied to them to form a final device, especially considering the piezoelectric properties of GaN. Investigation of influence of the metallic contacts on the structure of the NWs is of high importance for the… ▽ More Semiconductor nanowires (NWs) have a broad range of applications for nano- and optoelectronics. The strain field of gallium nitride (GaN) NWs could be significantly changed when contacts are applied to them to form a final device, especially considering the piezoelectric properties of GaN. Investigation of influence of the metallic contacts on the structure of the NWs is of high importance for their applications in real devices. We have studied a series of different type of contacts and influence of the applied voltage bias on the contacted GaN NWs with the length of about 3 to 4 micrometers and with two different diameters of 200 nm and 350 nm. It was demonstrated that the NWs with the diameter of 200 nm are bend already by the interaction with the substrate. For all GaN NWs, significant structural changes were revealed after the contacts deposition. The results of our research may contribute to the future optoelectronic applications of the GaN nanowires. △ Less

Submitted 23 August, 2021; originally announced August 2021.

Comments: 18 pages, 7 figures, 14 references

Hidden Charge Order in an Iron Oxide Square-Lattice Compound

Authors: Jung-Hwa Kim, Darren C. Peets, Manfred Reehuis, Peter Adler, Andrey Maljuk, Tobias Ritschel, Morgan C. Allison, Jochen Geck, Jose R. L. Mardegan, Pablo J. Bereciartua Perez, Sonia Francoual, Andrew C. Walters, Thomas Keller, Paula M. Abdala, Philip Pattison, Pinder Dosanjh, Bernhard Keimer

Abstract: Since the discovery of charge disproportionation in the FeO$_2$ square-lattice compound Sr$_3$Fe$_2$O$_7$ by Mössbauer spectroscopy more than fifty years ago, the spatial ordering pattern of the disproportionated charges has remained "hidden" to conventional diffraction probes, despite numerous x-ray and neutron scattering studies. We have used neutron Larmor diffraction and Fe K-edge resonant x-r… ▽ More Since the discovery of charge disproportionation in the FeO$_2$ square-lattice compound Sr$_3$Fe$_2$O$_7$ by Mössbauer spectroscopy more than fifty years ago, the spatial ordering pattern of the disproportionated charges has remained "hidden" to conventional diffraction probes, despite numerous x-ray and neutron scattering studies. We have used neutron Larmor diffraction and Fe K-edge resonant x-ray scattering to demonstrate checkerboard charge order in the FeO$_2$ planes that vanishes at a sharp second-order phase transition upon heating above 332 K. Stacking disorder of the checkerboard pattern due to frustrated interlayer interactions broadens the corresponding superstructure reflections and greatly reduces their amplitude, thus explaining the difficulty to detect them by conventional probes. We discuss implications of these findings for research on "hidden order" in other materials. △ Less

Submitted 5 August, 2021; originally announced August 2021.

Comments: PRL in press, 12 pages including Supplementary, CIF files available as ancillary files

Journal ref: Phys. Rev. Lett. 127, 097203 (2021)

Chirality of Bloch domain walls in exchange biased CoO/Co bilayer seen by waveguide-enhanced neutron spin-flip scattering

Authors: Yu. Khaydukov, D. Lenk, V. Zdravkov, R. Morari, T. Keller, A. S. Sidorenko, L. R. Tagirov, R. Tidecks, S. Horn, B. Keimer

Abstract: Magnetic state of exchanged biased CoO(20nm)/Co($d_F$) bilayer ($d_F$=5-20nm) was studied by means of polarized neutron reflectometry. By spacing of CoO/Co bilayer and Al$_2$O$_3$ substrate with Nb(20nm) layer we created waveguide structure which allowed us to significantly enhance intensity of spin-flip (SF) scattering in the position of optical resonances. For the trained sample with thinnest Co… ▽ More Magnetic state of exchanged biased CoO(20nm)/Co($d_F$) bilayer ($d_F$=5-20nm) was studied by means of polarized neutron reflectometry. By spacing of CoO/Co bilayer and Al$_2$O$_3$ substrate with Nb(20nm) layer we created waveguide structure which allowed us to significantly enhance intensity of spin-flip (SF) scattering in the position of optical resonances. For the trained sample with thinnest Co(5nm) we detected strong SF scattering at the resonance position (up to 30\% of incoming intensity) speaking about high non-collinearity of the system. As $d_F$ increases, the intensity of SF scattering linearly decreases. At the same time we observed asymmetry of up-down and down-up scattering channels at the resonance positions. We attribute this asymmetry to the Zeeman splitting of neutrons energies with different initial polarization taking place in high external field. Analysis, however, shows that the applied in the PNR experiment external field is not enough to quantitatively explain the observed asymmetry for the samples with $d_F > $ 5nm and we have to postulate presence of additional magnetic field produced by sample. We attribute this additional field to the stray field produced by chiral Bloch domain walls. The chirality of the domain walls can be explained by Dzyaloshinskii-Moriya interaction arising at the CoO/Co interface. Our results can be useful for designing of spintronic devices using exchange bias effect. △ Less

Submitted 11 July, 2021; originally announced July 2021.

Journal ref: Phys. Rev. B 104, 174445(2021)

Exact verification of the strong BSD conjecture for some absolutely simple abelian surfaces

Authors: Timo Keller, Michael Stoll

Abstract: Let $X$ be one of the $28$ Atkin-Lehner quotients of a curve $X_0(N)$ such that $X$ has genus $2$ and its Jacobian variety $J$ is absolutely simple. We show that the Shafarevich-Tate group of $J/\mathbb{Q}$ is trivial. This verifies the strong BSD conjecture for $J$. Let $X$ be one of the $28$ Atkin-Lehner quotients of a curve $X_0(N)$ such that $X$ has genus $2$ and its Jacobian variety $J$ is absolutely simple. We show that the Shafarevich-Tate group of $J/\mathbb{Q}$ is trivial. This verifies the strong BSD conjecture for $J$. △ Less

Submitted 1 July, 2021; originally announced July 2021.

Comments: 7 pages, details will be presented in a forthcoming article

As easy as APC: overcoming missing data and class imbalance in time series with self-supervised learning

Authors: Fiorella Wever, T. Anderson Keller, Laura Symul, Victor Garcia

Abstract: High levels of missing data and strong class imbalance are ubiquitous challenges that are often presented simultaneously in real-world time series data. Existing methods approach these problems separately, frequently making significant assumptions about the underlying data generation process in order to lessen the impact of missing information. In this work, we instead demonstrate how a general se… ▽ More High levels of missing data and strong class imbalance are ubiquitous challenges that are often presented simultaneously in real-world time series data. Existing methods approach these problems separately, frequently making significant assumptions about the underlying data generation process in order to lessen the impact of missing information. In this work, we instead demonstrate how a general self-supervised training method, namely Autoregressive Predictive Coding (APC), can be leveraged to overcome both missing data and class imbalance simultaneously without strong assumptions. Specifically, on a synthetic dataset, we show that standard baselines are substantially improved upon through the use of APC, yielding the greatest gains in the combined setting of high missingness and severe class imbalance. We further apply APC on two real-world medical time-series datasets, and show that APC improves the classification performance in all settings, ultimately achieving state-of-the-art AUPRC results on the Physionet benchmark. △ Less

Submitted 26 January, 2022; v1 submitted 29 June, 2021; originally announced June 2021.

Comments: Accepted to the NeurIPS 2021 Workshop on Self-Supervised Learning: Theory and Practice

Quadratic Chabauty for Atkin-Lehner Quotients of Modular Curves of Prime Level and Genus 4, 5, 6

Authors: Nikola Adžaga, Vishal Arul, Lea Beneish, Mingjie Chen, Shiva Chidambaram, Timo Keller, Boya Wen

Abstract: We use the method of quadratic Chabauty on the quotients $X_0^+(N)$ of modular curves $X_0(N)$ by their Fricke involutions to provably compute all the rational points of these curves for prime levels $N$ of genus four, five, and six. We find that the only such curves with exceptional rational points are of levels $137$ and $311$. In particular there are no exceptional rational points on those curv… ▽ More We use the method of quadratic Chabauty on the quotients $X_0^+(N)$ of modular curves $X_0(N)$ by their Fricke involutions to provably compute all the rational points of these curves for prime levels $N$ of genus four, five, and six. We find that the only such curves with exceptional rational points are of levels $137$ and $311$. In particular there are no exceptional rational points on those curves of genus five and six. More precisely, we determine the rational points on the curves $X_0^+(N)$ for $N=137,173,199,251,311,157,181,227,263,163,197,211,223,269,271,359$. △ Less

Submitted 11 May, 2021; originally announced May 2021.

Comments: 25 pages. Comments are welcome

MSC Class: 14G05 (Primary) 11G30; 11G18 (Secondary)

A Conceptual Framework for Establishing Trust in Real World Intelligent Systems

Authors: Michael Guckert, Nils Gumpfer, Jennifer Hannig, Till Keller, Neil Urquhart

Abstract: Intelligent information systems that contain emergent elements often encounter trust problems because results do not get sufficiently explained and the procedure itself can not be fully retraced. This is caused by a control flow depending either on stochastic elements or on the structure and relevance of the input data. Trust in such algorithms can be established by letting users interact with the… ▽ More Intelligent information systems that contain emergent elements often encounter trust problems because results do not get sufficiently explained and the procedure itself can not be fully retraced. This is caused by a control flow depending either on stochastic elements or on the structure and relevance of the input data. Trust in such algorithms can be established by letting users interact with the system so that they can explore results and find patterns that can be compared with their expected solution. Reflecting features and patterns of human understanding of a domain against algorithmic results can create awareness of such patterns and may increase the trust that a user has in the solution. If expectations are not met, close inspection can be used to decide whether a solution conforms to the expectations or whether it goes beyond the expected. By either accepting or rejecting a solution, the user's set of expectations evolves and a learning process for the users is established. In this paper we present a conceptual framework that reflects and supports this process. The framework is the result of an analysis of two exemplary case studies from two different disciplines with information systems that assist experts in their complex tasks. △ Less

Submitted 12 April, 2021; originally announced April 2021.

Adiabatic critical quantum metrology cannot reach the Heisenberg limit even when shortcuts to adiabaticity are applied

Authors: Karol Gietka, Friederike Metz, Tim Keller, **g Li

Abstract: We show that the quantum Fisher information attained in an adiabatic approach to critical quantum metrology cannot lead to the Heisenberg limit of precision and therefore regular quantum metrology under optimal settings is always superior. Furthermore, we argue that even though shortcuts to adiabaticity can arbitrarily decrease the time of preparing critical ground states, they cannot be used to a… ▽ More We show that the quantum Fisher information attained in an adiabatic approach to critical quantum metrology cannot lead to the Heisenberg limit of precision and therefore regular quantum metrology under optimal settings is always superior. Furthermore, we argue that even though shortcuts to adiabaticity can arbitrarily decrease the time of preparing critical ground states, they cannot be used to achieve or overcome the Heisenberg limit for quantum parameter estimation in adiabatic critical quantum metrology. As case studies, we explore the application of counter-diabatic driving to the Landau-Zener model and the quantum Rabi model. △ Less

Submitted 24 June, 2021; v1 submitted 23 March, 2021; originally announced March 2021.

Comments: 22 pages, 6 figures

Journal ref: Quantum 5, 489 (2021)

Single Alloy Nanoparticle X-Ray Imaging during a Catalytic Reaction

Authors: Young Yong Kim, Thomas F. Keller, Tiago J. Goncalves, Manuel Abuin, Henning Runge, Luca Gelisio, Jerome Carnis, Vedran Vonk, Philipp N. Plessow, Ivan A. Vartanyants, Andreas Stierle

Abstract: The imaging of active nanoparticles represents a milestone in decoding heterogeneous catalysts dynamics. We report the facet resolved, surface strain state of a single PtRh alloy nanoparticle on SrTiO3 determined by coherent x-ray diffraction imaging under catalytic reaction conditions. Density functional theory calculations allow us to correlate the facet surface strain state to its reaction envi… ▽ More The imaging of active nanoparticles represents a milestone in decoding heterogeneous catalysts dynamics. We report the facet resolved, surface strain state of a single PtRh alloy nanoparticle on SrTiO3 determined by coherent x-ray diffraction imaging under catalytic reaction conditions. Density functional theory calculations allow us to correlate the facet surface strain state to its reaction environment dependent chemical composition. We find that the initially Pt terminated nanoparticle surface gets Rh enriched under CO oxidation reaction conditions. The local composition is facet orientation dependent and the Rh enrichment is non-reversible under subsequent CO reduction. Tracking facet resolved strain and composition under operando conditions is crucial for a rational design of more efficient heterogeneous catalysts with tailored activity, selectivity and lifetime. △ Less

Submitted 2 March, 2021; originally announced March 2021.

Comments: 15 pages, 4 figures, 32 references

Journal ref: Science Advances 7, eabh0757 (2021)

Tunable spin-flop transition in artificial ferrimagnets

Authors: N. O. Antropov, E. A. Kravtsov, M. V. Makarova, V. V. Proglyado, T. Keller, I. A. Subbotin, E. M. Pashaev, G. V. Prutskov, A. L. Vasiliev, Yu. M. Chesnokov, N. G. Bebenin, V. V. Ustinov, B. Keimer, Yu. N. Khaydukov

Abstract: Spin-flop transition (SFT) consists in a jump-like reversal of antiferromagnetic magnetic moments into a non-collinear state when the magnetic field increases above the critical value. Potentially the SFT can be utilized in many applications of a rapidly develo** antiferromagnetic spintronics. However, the difficulty of using them in conventional antiferromagnets lies in (a) too large switching… ▽ More Spin-flop transition (SFT) consists in a jump-like reversal of antiferromagnetic magnetic moments into a non-collinear state when the magnetic field increases above the critical value. Potentially the SFT can be utilized in many applications of a rapidly develo** antiferromagnetic spintronics. However, the difficulty of using them in conventional antiferromagnets lies in (a) too large switching magnetic fields (b) the need for presence of a magnetic anisotropy, and (c) requirement to apply magnetic field along the correspondent anisotropy axis. In this work we propose to use artificial ferrimagnets in which the spin-flop transition occurs without anisotropy and the transition field can be lowered by adjusting exchange coupling in the structure. This is proved by experiment on artificial Fe-Gd ferrimagnets where usage of Pd spacers allowed us to suppress the transition field by two orders of magnitude. △ Less

Submitted 1 February, 2021; v1 submitted 15 January, 2021; originally announced January 2021.

Comments: article, 5 pages, 3 figures

Journal ref: Phys. Rev. B 104, 054414 (2021)

The Prime Graphs of Some Classes of Finite Groups

Authors: Chris Florez, Jonathan Higgins, Kyle Huang, Thomas Michael Keller, Dawei Shen, Yong Yang

Abstract: In this paper we study prime graphs of finite groups. The prime graph of a finite group $G$, also known as the Gruenberg-Kegel graph, is the graph with vertex set {primes dividing $|G|$} and an edge $p$-$q$ if and only if there exists an element of order $pq$ in $G$. In finite group theory, studying the prime graph of a group has been an important topic for the past almost half century. Only recen… ▽ More In this paper we study prime graphs of finite groups. The prime graph of a finite group $G$, also known as the Gruenberg-Kegel graph, is the graph with vertex set {primes dividing $|G|$} and an edge $p$-$q$ if and only if there exists an element of order $pq$ in $G$. In finite group theory, studying the prime graph of a group has been an important topic for the past almost half century. Only recently prime graphs of solvable groups have been characterized in graph theoretical terms only. In this paper, we continue this line of research and give complete characterizations of several classes of groups, including groups of square-free order, metanilpotent groups, groups of cube-free order, and, for any $n\in \mathbb{N}$, solvable groups of $n^\text{th}$-power-free order. We also explore the prime graphs of groups whose composition factors are cyclic or $A_5$ and draw connections to a conjecture of Maslova. We then propose an algorithm that recovers the prime graph from a dual prime graph. △ Less

Submitted 2 January, 2022; v1 submitted 1 January, 2021; originally announced January 2021.

Comparison of different Tate conjectures

Authors: Veronika Ertl, Timo Keller, Yanshuai Qin

Abstract: For an abelian variety $A$ over a finitely generated field $K$ of characteristic $p > 0$, we prove that the algebraic rank of $A$ is at most a suitably defined analytic rank. Moreover, we prove that equality, i.e., the BSD rank conjecture, holds for $A/K$ if and only if a suitably defined Tate--Shafarevich group of $A/K$ (1) has finite $\ell$-primary component for some/all $\ell \neq p$, or (2) fi… ▽ More For an abelian variety $A$ over a finitely generated field $K$ of characteristic $p > 0$, we prove that the algebraic rank of $A$ is at most a suitably defined analytic rank. Moreover, we prove that equality, i.e., the BSD rank conjecture, holds for $A/K$ if and only if a suitably defined Tate--Shafarevich group of $A/K$ (1) has finite $\ell$-primary component for some/all $\ell \neq p$, or (2) finite prime-to-$p$ part, or (3) has $p$-primary part of finite exponent, or (4) is of finite exponent. There is an algorithm to verify those conditions for concretely given $A/K$. △ Less

Submitted 24 June, 2024; v1 submitted 2 December, 2020; originally announced December 2020.

MSC Class: 11G40 (Primary) 11G05; 11G10; 14G10 (Secondary)

Minimal Prime Graphs of Solvable Groups

Authors: Chris Florez, Jonathan Higgins, Kyle Huang, Thomas Michael Keller, Dawei Shen

Abstract: We explore graph theoretical properties of minimal prime graphs of finite solvable groups. In finite group theory studying the prime graph of a group has been an important topic for the past almost half century. Recently prime graphs of solvable groups have been characterized in graph theoretical terms only. This now allows the study of these graphs with methods from graph theory only. Minimal pri… ▽ More We explore graph theoretical properties of minimal prime graphs of finite solvable groups. In finite group theory studying the prime graph of a group has been an important topic for the past almost half century. Recently prime graphs of solvable groups have been characterized in graph theoretical terms only. This now allows the study of these graphs with methods from graph theory only. Minimal prime graphs turn out to be of particular interest, and in this paper we pursue this further by exploring, among other things, diameters, Hamiltonian cycles and the property of being self-complementary for minimal prime graphs. We also study a new, but closely related notion of minimality for prime graphs and look into counting minimal prime graphs. △ Less

Submitted 20 November, 2020; originally announced November 2020.

The Adjacency Spectra of Some Families of Minimally Connected Prime Graphs

Authors: Chris Florez, Jonathan Higgins, Kyle Huang, Thomas Michael Keller, Dawei Shen

Abstract: In finite group theory, studying the prime graph of a group has been an important topic for almost the past half-century. Recently, prime graphs of solvable groups have been characterized in graph theoretical terms only. This now allows the study of these graphs without any knowledge of the group theoretical background. In this paper we study prime graphs from a linear algebra angle and focus on t… ▽ More In finite group theory, studying the prime graph of a group has been an important topic for almost the past half-century. Recently, prime graphs of solvable groups have been characterized in graph theoretical terms only. This now allows the study of these graphs without any knowledge of the group theoretical background. In this paper we study prime graphs from a linear algebra angle and focus on the class of minimally connected prime graphs introduced in earlier work on the subject. As our main results, we determine the determinants of the adjacency matrices and the spectra of some important families of these graphs. △ Less

Submitted 17 November, 2020; originally announced November 2020.

Comments: 23 pages, 4 figures

MSC Class: 15A18; 05C25

Self Normalizing Flows

Authors: T. Anderson Keller, Jorn W. T. Peters, Priyank Jaini, Emiel Hoogeboom, Patrick Forré, Max Welling

Abstract: Efficient gradient computation of the Jacobian determinant term is a core problem in many machine learning settings, and especially so in the normalizing flow framework. Most proposed flow models therefore either restrict to a function class with easy evaluation of the Jacobian determinant, or an efficient estimator thereof. However, these restrictions limit the performance of such density models,… ▽ More Efficient gradient computation of the Jacobian determinant term is a core problem in many machine learning settings, and especially so in the normalizing flow framework. Most proposed flow models therefore either restrict to a function class with easy evaluation of the Jacobian determinant, or an efficient estimator thereof. However, these restrictions limit the performance of such density models, frequently requiring significant depth to reach desired performance levels. In this work, we propose Self Normalizing Flows, a flexible framework for training normalizing flows by replacing expensive terms in the gradient by learned approximate inverses at each layer. This reduces the computational complexity of each layer's exact update from $\mathcal{O}(D^3)$ to $\mathcal{O}(D^2)$, allowing for the training of flow architectures which were otherwise computationally infeasible, while also providing efficient sampling. We show experimentally that such models are remarkably stable and optimize to similar data likelihood values as their exact gradient counterparts, while training more quickly and surpassing the performance of functionally constrained counterparts. △ Less

Submitted 9 June, 2021; v1 submitted 14 November, 2020; originally announced November 2020.

Silicon Flexes Muscles: Giant Electrochemical Actuation in a Nanoporous Silicon-Polypyrrole Hybrid Material

Authors: Manuel Brinker, Guido Dittrich, Claudia Richert, Pirmin Lakner, Tobias Krekeler, Thomas F. Keller, Norbert Huber, Patrick Huber

Abstract: The absence of piezoelectricity in silicon makes direct electro-mechanical applications of this mainstream semiconductor impossible. Integrated electrical control of the silicon mechanics, however, would open up new perspectives for on-chip actuorics. Here, we combine wafer-scale nanoporosity in single-crystalline silicon with polymerization of an artificial muscle material inside pore space to sy… ▽ More The absence of piezoelectricity in silicon makes direct electro-mechanical applications of this mainstream semiconductor impossible. Integrated electrical control of the silicon mechanics, however, would open up new perspectives for on-chip actuorics. Here, we combine wafer-scale nanoporosity in single-crystalline silicon with polymerization of an artificial muscle material inside pore space to synthesize a composite that shows macroscopic electrostrain in aqueous electrolyte. The voltage-strain coupling is 3 orders of magnitude larger than the best-performing ceramics in terms of piezoelectric actuation. We trace this huge electroactuation to the concerted action of 100 billions of nanopores per square centimetre cross-section and to potential-dependent pressures of up to 150 atmospheres at the single-pore scale. The exceptionally small operation voltages (0.4-0.9 V) along with the sustainable and biocompatible base materials make this hybrid promising for bio-actuator applications. △ Less

Submitted 8 October, 2020; originally announced October 2020.

Comments: Supplementary data as ancillary file available

Journal ref: Science Advances 6, eaba1483 (2020)

A Feshbach engine in the Thomas-Fermi regime

Authors: Tim Keller, Thomás Fogarty, **g Li, Thomas Busch

Abstract: Bose-Einstein condensates can be used to produce work by tuning the strength of the interparticle interactions with the help of Feshbach resonances. In inhomogeneous potentials, these interaction ramps change the volume of the trapped gas allowing one to create a thermodynamic cycle known as the Feshbach engine. However, in order to obtain a large power output, the engine strokes must be performed… ▽ More Bose-Einstein condensates can be used to produce work by tuning the strength of the interparticle interactions with the help of Feshbach resonances. In inhomogeneous potentials, these interaction ramps change the volume of the trapped gas allowing one to create a thermodynamic cycle known as the Feshbach engine. However, in order to obtain a large power output, the engine strokes must be performed on a short timescale, which is in contrast with the fact that the efficiency of the engine is reduced by irreversible work if the strokes are done in a non-adiabatic fashion. Here we investigate how such an engine can be run in the Thomas-Fermi regime and present a shortcut to adiabaticity that minimizes the irreversible work and allows for efficient engine operation. △ Less

Submitted 1 September, 2020; v1 submitted 14 May, 2020; originally announced May 2020.

Comments: 8 pages, 7 figures

Journal ref: Phys. Rev. Research 2, 033335 (2020)

Dimension of Restricted Classes of Interval Orders

Authors: Mitchel T. Keller, Ann N. Trenk, Stephen J. Young

Abstract: Rabinovitch showed in 1978 that the interval orders having a representation consisting of only closed unit intervals have order dimension at most 3. This article shows that the same dimension bound applies to two other classes of posets: those having a representation consisting of unit intervals (but with a mixture of open and closed intervals allowed) and those having a representation consisting… ▽ More Rabinovitch showed in 1978 that the interval orders having a representation consisting of only closed unit intervals have order dimension at most 3. This article shows that the same dimension bound applies to two other classes of posets: those having a representation consisting of unit intervals (but with a mixture of open and closed intervals allowed) and those having a representation consisting of closed intervals with lengths in $\{0,1\}$. △ Less

Submitted 3 January, 2022; v1 submitted 17 April, 2020; originally announced April 2020.

Comments: 9 pages, 3 figures

MSC Class: 06A07

A lower bound for the number of odd-degree representations of a finite group

Authors: Nguyen Ngoc Hung, Thomas Michael Keller, Yong Yang

Abstract: Let $G$ be a finite group and $P$ a Sylow $2$-subgroup of $G$. We obtain both asymptotic and explicit bounds for the number of odd-degree irreducible complex representations of $G$ in terms of the size of the abelianization of $P$. To do so, we, on one hand, make use of the recent proof of the McKay conjecture for the prime 2 by Malle and Späth, and, on the other hand, prove lower bounds for the c… ▽ More Let $G$ be a finite group and $P$ a Sylow $2$-subgroup of $G$. We obtain both asymptotic and explicit bounds for the number of odd-degree irreducible complex representations of $G$ in terms of the size of the abelianization of $P$. To do so, we, on one hand, make use of the recent proof of the McKay conjecture for the prime 2 by Malle and Späth, and, on the other hand, prove lower bounds for the class number of the semidirect product of an odd-order group acting on an abelian $2$-group. △ Less

Submitted 26 August, 2020; v1 submitted 6 April, 2020; originally announced April 2020.

Comments: 17 pages, an asymptotic bound has been added to the newer version

MSC Class: 20C15; 20D10; 20D05

Proximity effect in [Nb(1.5nm)/Fe(x)]$_{10}$/Nb(50nm) superconducting/ferromagnet heterostructures

Authors: Yu. N. Khaydukov, S. Pütter, L. Guasco, R. Morari, G. Kim, T. Keller, A. S. Sidorenko, B. Keimer

Abstract: We have investigated the structural, magnetic and superconducting properties of [Nb(1.5nm)/Fe(x)]$_{10}$ superlattices deposited on a thick Nb(50nm) layer. Our investigation showed that the Nb(50nm) layer grows epitaxially at 800$^\circ$C on Al$_2$O$_3$(1$\bar{1}$02) substrate. Samples grown at this condition posses a high residual resistivity ratio of 15-20. By using neutron reflectometry we show… ▽ More We have investigated the structural, magnetic and superconducting properties of [Nb(1.5nm)/Fe(x)]$_{10}$ superlattices deposited on a thick Nb(50nm) layer. Our investigation showed that the Nb(50nm) layer grows epitaxially at 800$^\circ$C on Al$_2$O$_3$(1$\bar{1}$02) substrate. Samples grown at this condition posses a high residual resistivity ratio of 15-20. By using neutron reflectometry we show that Fe/Nb superlattices with $x<$ 4 nm form a depth-modulated FeNb alloy with concentration of iron varying between 60% and 90%. This alloy has properties of a weak ferromagnet. Proximity of this weak ferromagnetic layer to a thick superconductor leads to an intermediate phase that is characterized by co-existing superconducting and normal-state domains. By increasing the thickness of the Fe layer to $x$ = 4 nm the intermediate phase disappears. We attribute the intermediate state to proximity induced non-homogeneous superconductivity in the periodic Fe/Nb structure. △ Less

Submitted 28 November, 2019; v1 submitted 22 November, 2019; originally announced November 2019.

Journal ref: Beilstein J. Nanotechnol. 2020, 11,1254-1263

Tunable perpendicular exchange bias in oxide heterostructures

Authors: Gideok Kim, Yury Khaydukov, Martin Bluschke, Y. Eren Suyolcu, Georg Christiani, Kwanghyo Son, Christopher Dietl, Thomas Keller, Eugen Weschke, P. A. van Aken, Gennady Logvenov, Bernhard Keimer

Abstract: The exchange bias effect is an essential component of magnetic memory and spintronic devices. Whereas recent research has shown that anisotropies perpendicular to the device plane provide superior stability against thermal noise, it has proven remarkably difficult to realize perpendicular exchange bias in thin-film structures. Here we demonstrate a strong perpendicular exchange bias effect in hete… ▽ More The exchange bias effect is an essential component of magnetic memory and spintronic devices. Whereas recent research has shown that anisotropies perpendicular to the device plane provide superior stability against thermal noise, it has proven remarkably difficult to realize perpendicular exchange bias in thin-film structures. Here we demonstrate a strong perpendicular exchange bias effect in heterostructures of the quasi-two-dimensional canted antiferromagnet La$_2$CuO$_4$ and ferromagnetic (La,Sr)MnO$_3$ synthesized by ozone-assisted molecular beam epitaxy. The magnitude of this effect can be controlled via the do** level of the cuprate layers. Canted antiferromagnetism of layered oxides is thus a new and potentially powerful source of uniaxial anisotropy in magnetic devices. △ Less

Submitted 13 September, 2019; originally announced September 2019.

Comments: 7 pages, 5 figures

Journal ref: Phys. Rev. Materials 3, 084420 (2019)