arXiv research

Enhanced X-ray polarimetry with deep learning for better exposure times.

The paper develops residue currents for cohesive modules and proves a generalized Poincaré-Lelong formula.

In this paper we propose the use of continuous residual modules for graph kernels in Graph Neural Networks. We show how both discrete and continuous residual layers allow for more robust training, being that continuous residual layers are those which are applied by integrating through an Ordinary Differential Equation …

We show that Residual Networks (ResNet) is equivalent to boosting feature representation, without any modification to the underlying ResNet training algorithm. A regret bound based on Online Gradient Boosting theory is proved and suggests that ResNet could achieve Online Gradient Boosting regret bounds through neural n…

This paper studies deep learning methodologies for portfolio optimization in the US equities market. We present a novel residual switching network that can automatically sense changes in market regimes and switch between momentum and reversal predictors accordingly. The residual switching network architecture combines …

In this paper, we study the interplay between modules and sub-objects in holomorphic Poisson geometry. In particular, we define a new notion of "residue" for a Poisson module, analogous to the Poincaré residue of a meromorphic volume form. Of particular interest is the interaction between the residues of the canonical …

Neuro-symbolic agent learns systematic generalisation from formal instructions.

Sources of variability in experimentally derived data include measurement error in addition to the physical phenomena of interest. This measurement error is a combination of systematic components, originating from the measuring instrument, and random measurement errors. Several novel biological technologies, such as ma…

Paper tackles division difficulty, proposing new methods to improve accuracy.

CRC improves multivariate forecasting accuracy without risking performance degradation.

DEPTS learns to forecast periodic time series with improved accuracy.

Proposes a neural network method to correct residual distortions in coordinate transformations.

PRISMA uses PDE residuals for fast, robust, and accurate inference.

Consider the space $R_Δ$ of rational functions of several variables with poles on a fixed arrangement $Δ$ of hyperplanes. We obtain a decomposition of $R_Δ$ as a module over the ring of differential operators with constant coefficients. We generalize to the space $R_Δ$ the notions of principal part and of residue, and …

We equip a knot $K$ with a set of colored bonds, that is, colored intervals properly embedded into $\mathbb{R}^3 \setminus K$. Such a construction can be viewed as a structure that topologically models a closed protein chain including any type of bridges connecting the backbone residues. We introduce an invariant of su…

Study of cubic skein modules in 3-sphere and arbitrary 3-manifolds.

Time-aware deep learning methods improve spatial downscaling of atmospheric pollutants.

Image super-resolution is a challenging task and has attracted increasing attention in research and industrial communities. In this paper, we propose a novel end-to-end Attention-based DenseNet with Residual Deconvolution named as ADRD. In our ADRD, a weighted dense block, in which the current layer receives weighted f…

We present a novel high frequency residual learning framework, which leads to a highly efficient multi-scale network (MSNet) architecture for mobile and embedded vision problems. The architecture utilizes two networks: a low resolution network to efficiently approximate low frequency components and a high resolution ne…

A new method suppresses echo more effectively with lower latency.

In this work, we introduce the Global Planar Convolution module as a building-block for fully-convolutional networks that aggregates global information and, therefore, enhances the context perception capabilities of segmentation networks in the context of brain tumor segmentation. We implement two baseline architecture…

Residual networks' depth is mathematically equivalent to expanding an implicit ensemble size.

The classical abelian invariants of a knot are the Alexander module, which is the first homology group of the the unique infinite cyclic covering space of S^3-K, considered as a module over the (commutative) Laurent polynomial ring, and the Blanchfield linking pairing defined on this module. From the perspective of the…

We propose a novel exponentially-modified Gaussian (EMG) mixture residual model. The EMG mixture is well suited to model residuals that are contaminated by a distribution with positive support. This is in contrast to commonly used robust residual models, like the Huber loss or $\ell_1$, which assume a symmetric contami…

The paper establishes a Poisson Poincaré-Dulac theorem for Poisson-flat connections.

Introduces a novel spatial attention module for convolutional networks.

We have developed a statistical technique to test the model assumption of binary regime switching extension of the geometric Brownian motion (GBM) model by proposing a new discriminating statistics. Given a time series data, we have identified an admissible class of the regime switching candidate models for the statist…

26 different concrete representations of the space of vector valued distributions on a smooth manifold of dimension n are presented systematically, most of them new. In the particular case of representations as module homomorphisms acting on sections of the dual bundle resp. on n-forms, the continuity of these homomorp…

Develops EFT for ResNets, revealing limitations of kernel-only approach.

This work introduces a new loss function to improve the efficiency of optimization-based PDE solvers.

We identify and study two common failure modes for early training in deep ReLU nets. For each we give a rigorous proof of when it occurs and how to avoid it, for fully connected and residual architectures. The first failure mode, exploding/vanishing mean activation length, can be avoided by initializing weights from a …

Extends hyperparameter transfer across model sizes and modules, improving training speed.

A framework evaluates the impact of different modules in graph contrastive learning.

Bias correction needed after deep learning regression training.

Skip connections are an essential component of current state-of-the-art deep neural networks (DNNs) such as ResNet, WideResNet, DenseNet, and ResNeXt. Despite their huge success in building deeper and more powerful DNNs, we identify a surprising security weakness of skip connections in this paper. Use of skip connectio…

This work proposes a novel autoencoder for fusing visible and infrared images.

In this work, we investigate the value of employing deep learning for the task of wireless signal modulation recognition. Recently in [1], a framework has been introduced by generating a dataset using GNU radio that mimics the imperfections in a real wireless channel, and uses 10 different modulation types. Further, a …

Despite recent advances, large scale visual artifacts are still a common occurrence in images generated by GANs. Previous work has focused on improving the generator's capability to accurately imitate the data distribution $p_{data}$. In this paper, we instead explore methods that enable GANs to actively avoid errors b…

Recent work in fairness in machine learning has proposed adjusting for fairness by equalizing accuracy metrics across groups and has also studied how datasets affected by historical prejudices may lead to unfair decision policies. We connect these lines of work and study the residual unfairness that arises when a fairn…

Proposes RaT to mitigate bias in student-teacher estimation.

Field theory explains optimal scaling in ResNets for signal propagation.

Given a prime, alternating link diagram, we build a special cover of the link complement whose degree is bounded by a factorial function of the crossing number. It follows that a subgroup of the link group of that index embeds into right-angled Artin and Coxeter groups. Corollaries of this result include a quantificati…

New computations show symplectic groups and mapping class groups have different properties regarding torsion.

Let $X$ be a topological space. We consider certain generalized configuration spaces of points on $X$, obtained from the cartesian product $X^n$ by removing some intersections of diagonals. We give a systematic framework for studying the cohomology of such spaces using what we call "tcdga models" for the cochains on $X…

There is a growing interest in learning data representations that work well for many different types of problems and data. In this paper, we look in particular at the task of learning a single visual representation that can be successfully utilized in the analysis of very different types of images, from dog breeds to s…

Framework identifies discrepancies in physics models, improving sensor accuracy.

We propose a new architecture called Memory-Augmented Encoder-Solver (MAES) that enables transfer learning to solve complex working memory tasks adapted from cognitive psychology. It uses dual recurrent neural network controllers, inside the encoder and solver, respectively, that interface with a shared memory module a…

Let g be a complex, simple Lie algebra with Cartan subalgebra h and Weyl group W. We construct a one-parameter family of flat connections D on h with values in any finite-dimensional h-module V and simple poles on the root hyperplanes. The corresponding monodromy representation of the braid group B of type g is a defor…