arXiv research

Study on weakly G-slim complexes and non-positive immersions for group presentations.

Slim curves on 3-sphere help spherical CR uniformizations.

In this paper we consider sparse and identifiable linear latent variable (factor) and linear Bayesian network models for parsimonious analysis of multivariate data. We propose a computationally efficient method for joint parameter and model inference, and model comparison. It consists of a fully Bayesian hierarchy for …

We introduce Supersparse Linear Integer Models (SLIM) as a tool to create scoring systems for binary classification. We derive theoretical bounds on the true risk of SLIM scoring systems, and present experimental results to show that SLIM scoring systems are accurate, sparse, and interpretable classification models.

SLIM model tackles graph classification by resolving part-interaction dilemmas.

SLIM efficiently solves overidentified models in a scalable manner.

We prove that every slim double Lie groupoid with proper core action is completely determined by a factorization of a certain canonically defined "diagonal" Lie groupoid.

Scoring systems are linear classification models that only require users to add, subtract and multiply a few small numbers in order to make a prediction. These models are in widespread use by the medical community, but are difficult to learn from data because they need to be accurate and sparse, have coprime integer co…

We describe unicorn paths in the arc graph and show that they form 1-slim triangles and are invariant under taking subpaths. We deduce that all arc graphs are 7-hyperbolic. Considering the same paths in the arc and curve graph, this also shows that all curve graphs are 17-hyperbolic, including closed surfaces.

Scoring systems are classification models that only require users to add, subtract and multiply a few meaningful numbers to make a prediction. These models are often used because they are practical and interpretable. In this paper, we introduce an off-the-shelf tool to create scoring systems that both accurate and inte…

Bi-Lipschitz rigidity theorem for dense subgroups of algebraic groups.

Unified framework compresses GANs up to 47x with minimal quality loss.

Let $G$ be a Garside group with Garside element $Δ$, and let $Δ^m$ be the minimal positive central power of $Δ$. An element $g\in G$ is said to be 'periodic' if some power of it is a power of $Δ$. In this paper, we study periodic elements in Garside groups and their conjugacy classes. We show that the periodicity of an…

New method for mixed memberships using symmetrized Laplacian inverse matrix.

We model leverage as stochastic but independent of return shocks and of volatility and perform likelihood-based inference via the recently developed iterated filtering algorithm using S&P500 data, contributing new evidence to the still slim empirical support for random leverage variation.

Uniform bound on geodesic images for surfaces using bicorn curves.

Simplified LSTM models improve sentiment analysis on Twitter debate data.

Linear models like EASE and SLIM are competitive in recommendation, and this work explores their theoretical relationship.

The redundancy is widely recognized in Convolutional Neural Networks (CNNs), which enables to remove unimportant filters from convolutional layers so as to slim the network with acceptable performance drop. Inspired by the linear and combinational properties of convolution, we seek to make some filters increasingly clo…

The paper uses model-based trees to create interpretable surrogate models for complex machine learning models.

SLIM model predicts social network polarization using signed links.

ResRep prunes CNNs without losing accuracy by separating remembering and forgetting.

Convolutional Neural Network (CNN) based Deep Learning (DL) has achieved great progress in many real-life applications. Meanwhile, due to the complex model structures against strict latency and memory restriction, the implementation of CNN models on the resource-limited platforms is becoming more challenging. This work…

A new AI framework reduces costs and improves performance.

We prove an explicit equivalence between various hyperbolic type properties for quasi-geodesics in CAT(0) spaces. Specifically, we prove that for X a CAT(0) space and $γ$ a quasi-geodesic, the following four statements are equivalent and moreover the quantifiers in the equivalences are explicit: (i) $γ$ is S-Slim, (ii)…

This paper proposes the adaptation of Support Vector Data Description (SVDD) to the multiple kernel case (MK-SVDD), based on SimpleMKL. It also introduces a variant called Slim-MK-SVDD that is able to produce a tighter frontier around the data. For the sake of comparison, the equivalent methods are also developed for O…

We present a method that trains large capacity neural networks with significantly improved accuracy and lower dynamic computational cost. We achieve this by gating the deep-learning architecture on a fine-grained-level. Individual convolutional maps are turned on/off conditionally on features in the network. To achieve…

In machine learning and data mining, linear models have been widely used to model the response as parametric linear functions of the predictors. To relax such stringent assumptions made by parametric linear models, additive models consider the response to be a summation of unknown transformations applied on the predict…

New method improves online nonparametric estimators with minimal extra computation.

We introduce a simple generalization of rational bubble models which removes the fundamental problem discovered by [Lux and Sornette, 1999] that the distribution of returns is a power law with exponent less than 1, in contradiction with empirical data. The idea is that the price fluctuations associated with bubbles mus…

In this paper, we present some theoretical work to explain why simple gradient descent methods are so successful in solving non-convex optimization problems in learning large-scale neural networks (NN). After introducing a mathematical tool called canonical space, we have proved that the objective functions in learning…

Investors face constraints in Heston's model; optimal allocation differs from naive capped strategy.

A method models continuous-time glucose distributions in children with diabetes.

We investigate a long-debated question, which is how to create predictive models of recidivism that are sufficiently accurate, transparent, and interpretable to use for decision-making. This question is complicated as these models are used to support different decisions, from sentencing, to determining release on proba…

New theory controls compression change probability without prior knowledge.

New condition prevents hyperbolic spaces from matching curve complexes.

Study on complex line fields on almost-complex manifolds, proving existence conditions.

Homotopy types of curve and arc complexes are studied.

This research explores complex-valued neural networks and their implementation.

Paper introduces fat CW complexes including all closed manifolds.

The paper discusses $q$-deformations of the Aomoto complex.

Study calculates global sections on complex curves.

The paper studies lifts of complex structures on a manifold.

In this paper, we first provide an updated survey of the geometry of complex Cartan spaces. New characterizations for some particular classes of complex Cartan spaces are pointed out, e.g. Landsberg-Cartan, strongly Berwald-Cartan and others. We introduce the Cartan-Randers spaces which offer examples of Berwald-Cartan…

Study $L^2$ Hilbert complexes on complex manifolds.

The paper defines and constructs almost complex blow-ups on 4D almost complex manifolds.

Research shows arc complex is not quasi-isometric to sphere complex.

Study Hodge-de Rham numbers for almost complex 4-manifolds, extending properties from complex surfaces.