arXiv research

Extends algorithms to solve minimization problems faster.

SKI accelerates GP inference with sparse grids to handle higher dimensions.

It is shown that if $M$ is a strongly causal free of naked singularities space-time, then its causal structure is completely characterized by a partial order in the space of skies defined by means of a class non-negative Legendrian isotopies. It is also proved that such partial order is determined by the class of futur…

Optimal ski rental strategies with machine learning predictions.

A new GP inference method using simplices for high-dimensional data.

We introduce a new structured kernel interpolation (SKI) framework, which generalises and unifies inducing point methods for scalable Gaussian processes (GPs). SKI methods produce kernel approximations for fast computations through kernel interpolation. The SKI framework clarifies how the quality of an inducing point a…

Optimizes solving complex min-max problems with stochastic and nonconvex elements.

This paper analyzes error in SKI for Gaussian Processes, providing conditions for linear time inference.

Study examines diversification of mid-mountain ski tourism.

Grey-box model combines GP with motion data to analyze skiing forces.

A reconstruction theorem in terms of the topology and geometrical structures on the spaces of light rays and skies of a given space-time is discussed. This result can be seen as part of Penrose and Low's programme intending to describe the causal structure of a space-time $M$ in terms of the topological and geometrical…

Accelerator boosts energy efficiency for MANNs on FPGAs.

Efficiently maps indoor magnetic fields with SKI and D-SKI.

Develops accelerated fixed-point methods with delayed oracles for scientific computing.

Memory-augmented neural networks (MANNs) refer to a class of neural network models equipped with external memory (such as neural Turing machines and memory networks). These neural networks outperform conventional recurrent neural networks (RNNs) in terms of learning long-term dependency, allowing them to solve intrigui…

Project uses machine learning to identify skiers' techniques from power meter data.

SoftKI combines SKI and variational methods for scalable GP regression.

We give a shorter proof of the following theorem of Kathryn Mann \cite{M}: the identity component of the group of the compactly supported $C^r$ diffeomorphisms of $\R^n$ cannot admit a nontrivial $C^p$-action on $S^1$, provided $n\geq2$, $r\neq n+1$ and $p\geq2$. We also give a new proof of another theorem of Mann: any…

Recent work shows that inference for Gaussian processes can be performed efficiently using iterative methods that rely only on matrix-vector multiplications (MVMs). Structured Kernel Interpolation (SKI) exploits these techniques by deriving approximate kernels with very fast MVMs. Unfortunately, such strategies suffer …

New method solves root-finding problems with faster convergence.

Study of CB generating sets for infinite-type surfaces.

DAM with MRL improves relational reasoning in MANNs.

The paper proposes calibration to improve algorithm performance using machine learning predictions.

SKI speeds up Toeplitz Neural Networks by avoiding explicit decay bias and using frequency response.

New algorithm optimizes adaptive return level for Markowitz portfolios.

The paper examines properties of self-affine Sierpiński sponges using metric invariants.

Deep learning typically requires training a very capable architecture using large datasets. However, many important learning problems demand an ability to draw valid inferences from small size datasets, and such problems pose a particular challenge for deep learning. In this regard, various researches on "meta-learning…

Applying a theorem due to Belopol'ski and Birman, we show that the Laplace-Beltrami operator on 1-forms on ${\bf R}^n$ endowed with an asymptotically Euclidean metric has absolutely continuous spectrum equal to $[0, +\infty)$.

ARMIN improves memory efficiency and lightness in neural networks.

Augmenting a neural network with memory that can grow without growing the number of trained parameters is a recent powerful concept with many exciting applications. We propose a design of memory augmented neural networks (MANNs) called Labeled Memory Networks (LMNs) suited for tasks requiring online adaptation in class…

Classifies pure mapping class groups based on surface properties.

A three dimensional supergravity theory which generalizes the super IG theory of Witten and resembles the model discussed recently by Mann and Papadopoulos is displayed. The partition function is computed, and is shown to be a three-manifold invariant generalizing the Casson invariant.

The study examines when mapping class groups are quasi-isometric to graphs of curves.

No natural topological compactification for Fulton-MacPherson.

Common complex diseases are likely influenced by the interplay of hundreds, or even thousands, of genetic variants. Converging evidence shows that genetic variants with low marginal effects (LME) play an important role in disease development. Despite their potential significance, discovering LME genetic variants and as…

Historically tensor calculus emerged in an attempt to formalize Rie- mann's ideas. We show that tensor calculus can be based also on Lie's idea of a transformation group and this approach leads quite naturally to the concept of deformation of a transformation group and the Kodaira- Spencer map.

We calculate the integer cohomology ring and stable tangent bundle of a family of compact, 3-Sasakian 7-manifolds constructed by Boyer, Galicki, Mann, and Rees. Previously only the rational cohomology ring was known. The most important part of the cohomology ring is a torsion group that we describe explicitly and whose…

New analysis of stochastic approximation with non-expansive mappings.

We consider the interplay of point counts, singular cohomology, étale cohomology, eigenvalues of the Frobenius and the Grothendieck ring of varieties for two families of varieties: spaces of rational maps and moduli spaces of marked, degree $d$ rational curves in $\mathbb{P}^n$. We deduce as special cases algebro-geome…

In this paper, we study Jacobi operators associated to algebraic curvature maps (tensors) on lightlike submanifolds M. We investigate conditions for an induced Rie- mann curvature tensor to be an algebraic curvature tensor on M. We introduce the notion of lightlike Osserman submanifolds and an example of 2-degenerate O…

Optimal hashing embeddings reduce linear least squares solving time.

The set N of all null geodesics of a globally hyperbolic (d+1)-dimensional spacetime (M,g) is naturally a smooth (2d-1)-dimensional contact manifold. The sky of an event is the subset of N defined by all null geodesics through that event, and is an embedded Legendrian submanifold of N diffeomorphic to a (d-1)-dimension…

Perfect mapping class groups of specific surfaces have no proper subgroups.

We define a conformal reference frame, i.e., a special projection of the six-dimensional sky bundle of a Lorentzian manifold (or the five-dimensional twistor space) to a three-dimensional manifold. We construct an example, a conformal compactification, for Minkowski space. Based on the complex structure on the skies, w…

Classification of torus homeomorphisms on fine curve graph completed.

Study shows surfaces without certain curves have infinite orbit graph.

The spectral asymptotics for linear elasticity with mixed boundary conditions are shown to be old results.

New proof classifies homogeneous 3-Sasakian and quaternionic Kähler manifolds.