March 2023 arXiv papers — page 67

Xiaotian Weng, Yi Chen, Zhichao Zheng, Yanhui Gu

Lightweight neural networks for single-image super-resolution (SISR) tasks have made substantial breakthroughs in recent years. Compared to low-frequency information, high-frequency detail is much more difficult to reconstruct. Most SISR models allocate equal computational resources for low-frequency and high-frequency information, which leads to redundant p

Bowei He, Xu He, Yingxue Zhang, Ruiming Tang

Personalized recommender systems have been widely studied and deployed to reduce information overload and satisfy users' diverse needs. However, conventional recommendation models solely conduct a one-time training-test fashion and can hardly adapt to evolving demands, considering user preference shifts and ever-increasing users and items in the real world.

Henrik Schopmans, Patrick Reiser, Pascal Friederich

Machine learning techniques have successfully been used to extract structural information such as the crystal space group from powder X-ray diffractograms. However, training directly on simulated diffractograms from databases such as the ICSD is challenging due to its limited size, class-inhomogeneity, and bias toward certain structure types. We propose an a

Zhiqiang Kou, Yuheng Jia, Jing Wang, Boyu Shi

Label distribution (LD) uses the description degree to describe instances, which provides more fine-grained supervision information when learning with label ambiguity. Nevertheless, LD is unavailable in many real-world applications. To obtain LD, label enhancement (LE) has emerged to recover LD from logical label. Existing LE approach have the following prob

Cécile Bouette, Laura Luzzi, Ligong Wang

We extend previous results on covert communication over the additive white Gaussian noise channel to two other types of additive noise channels. The first is the Gaussian channel with memory, where the noise sequence is a Gaussian vector with an arbitrary invertible covariance matrix. We show that the fundamental limit for covert communication over such a ch

Chen Lan, Hao Yang, Yang Guo, Yan-Gang Miao

The essential singularity in Einstein's gravity can be avoidable if the preconditions of Penrose's theorem can be bypassed, i.e., if the strong energy condition is broken in the vicinity of a black hole center. The singularity mentioned here includes two aspects: (i) the divergence of curvature invariants, and (ii) the incompleteness of geodesics. Both aspec

Yannis Papanikolaou, Francesco Tuveri, Misa Ogura, Daniel O'Donovan

In this work we present a deep learning approach to conduct hypothesis-free, transcriptomics-based matching of drugs for diseases. Our proposed neural network architecture is trained on approved drug-disease indications, taking as input the relevant disease and drug differential gene expression profiles, and learns to identify novel indications. We assemble

Ali Reza, Sonia Naseer, F D Zaman, A H Kara

The reaction diffusion equation arises in physical situations in problems from population growth, genetics and physical sciences. We consider the generalised Fisher equation in cylindrical coordinates from Lie theory stand point. An invariance method is performed and the optimal set of nonequivalent symmetries is obtained. Finally, the conservation laws are

Linfeng Shi, Yan Li, Xi Zhu

As the rapid development of depth learning, object detection in aviatic remote sensing images has become increasingly popular in recent years. Most of the current Anchor Free detectors based on key point detection sampling directly regression and classification features, with the design of object loss function based on the horizontal bounding box. It is more

Jaume de Dios Pont, Helge Jørgen Samuelsen

In this paper we prove a uniform Fourier restriction estimate over the class of simple curves where the last coordinate function can be extended to a holomorphic function of bounded frequency in a sufficiently large disc. The proof is based on a decomposition scheme for this class of functions.

Xingjian Du, Zijie Wang, Xia Liang, Huidong Liang

Deep learning based methods have become a paradigm for cover song identification (CSI) in recent years, where the ByteCover systems have achieved state-of-the-art results on all the mainstream datasets of CSI. However, with the burgeon of short videos, many real-world applications require matching short music excerpts to full-length music tracks in the datab

Sven Nordebo

This paper presents a simple approach to combine the high-resolution narrowband features of some desired isolated line models together with the far wing behavior of the projection based strong collision (SC) method to line mixing which was introduced by Bulanin, Dokuchaev, Tonkov and Filippov. The method can be viewed in terms of a small diagonal perturbatio

Elias Kyritsis, Charalampos Daoutis, Andreas Zezas, Konstantinos Kouroumpatzakis

We use the Random Forest (RF) algorithm to develop a tool for automated activity classification of galaxies into 5 different classes: Star-forming (SF), AGN, LINER, Composite, and Passive. We train the algorithm on a combination of mid-IR (WISE) and optical photometric data while the true labels (activity classes) are based on emission line ratios. Our class

Silvia Gallucci, Marta Bonato, Martina Benini, Marta Parazzini

Since the spread of the wearable systems and the implementation of the forthcoming 5G in many devices, the question about the assessment of the exposure in wearable typical usage to millimeter waves is crucial and timely. For such frequencies, the power absorption becomes strongly superficial and involves only the most superficial tissue of the human body, i

M. Husak, A. Budkova, T. Pycha, A. Laposa

The paper solves the model of the miniature Power supply based on the piezoelectric cantilever. The aim of the future is to further hybrid integration and use of nanotechnology. Contents of the article belongs to the category of renewable energy sources with environment energy conversion into electrical energy. The work is focused on the use in small tempera

N. T. Bechler, J. Masell

Magnetic textures are promising candidates for unconventional computing due to their non-linear dynamics. We propose to investigate the rich variety of seemingly trivial lamellar magnetic phases, e.g., helical, spiral, stripy phase, or other one-dimensional soliton lattices. These are the natural stray field-free ground states of almost every magnet. The ord

Gyungchoon Go, Daehyeon An, Hyun-Woo Lee, Se Kwon Kim

We theoretically investigate the transport of magnon orbitals in a honeycomb antiferromagnet. We find that the magnon orbital Berry curvature is finite even without spin-orbit coupling and thus the resultant magnon orbital Hall effect is an intrinsic property of the honeycomb antiferromagnet rooted only in the exchange interaction and the lattice structure.

Yuxuan Han, Zhibo Wang, Feng Xu

Modeling non-Lambertian effects such as facial specularity leads to a more realistic 3D Morphable Face Model. Existing works build parametric models for diffuse and specular albedo using Light Stage data. However, only diffuse and specular albedo cannot determine the full BRDF. In addition, the requirement of Light Stage data is hard to fulfill for the resea

Yang Yu, Danruo Deng, Furui Liu, Yueming Jin

Semi-supervised learning (SSL) methods assume that labeled data, unlabeled data and test data are from the same distribution. Open-set semi-supervised learning (Open-set SSL) considers a more practical scenario, where unlabeled data and test data contain new categories (outliers) not observed in labeled data (inliers). Most previous works focused on outlier

Silvan Käser, Markus Meuwly

The rise of machine learning has greatly influenced the field of computational chemistry, and that of atomistic molecular dynamics simulations in particular. One of its most exciting prospects is the development of accurate, full-dimensional potential energy surfaces (PESs) for molecules and clusters, which, however, often require thousands to tens of thousa

Yajing Zheng, Jiyuan Zhang, Rui Zhao, Jianhao Ding

SpikeCV is a new open-source computer vision platform for the spike camera, which is a neuromorphic visual sensor that has developed rapidly in recent years. In the spike camera, each pixel position directly accumulates the light intensity and asynchronously fires spikes. The output binary spikes can reach a frequency of 40,000 Hz. As a new type of visual ex

Fankai Xie, Tenglong Lu, Ze Yu, Yaxian Wang

Using a comprehensive structure search and high-throughput first-principles calculations of 1483 compounds, this study presents the phase diagram of Lu-H-N. The formation energy landscape of Lu-H-N was derived and utilized to assess the thermodynamic stability of compounds. Results indicate that there are no stable Lu-H-N ternary structures in this system, b

Alice Barbora Tumpach, Stephen C. Preston

In many applications, one is interested in the shape of an object, like the contour of a bone or the trajectory of joints of a tennis player, irrespective of the way these shapes are parameterized. However for analysis of these shape spaces, it is sometimes useful to have a parameterization at hand, in particular if one is interested in deforming shapes. The

Weijia Wu, Yuzhong Zhao, Mike Zheng Shou, Hong Zhou

Collecting and annotating images with pixel-wise labels is time-consuming and laborious. In contrast, synthetic data can be freely available using a generative model (e.g., DALL-E, Stable Diffusion). In this paper, we show that it is possible to automatically obtain accurate semantic masks of synthetic images generated by the Off-the-shelf Stable Diffusion m

Tomasz Denkiewicz, Vincenzo Salzano, Mariusz P. Dabrowski

We apply the full set of most update dynamical and geometrical data in cosmology to the nonextensive Barrow entropic holographic dark energy. We show that the data point towards an extensive Gibbs-like entropic behaviour for the cosmological horizons, which is the extreme case of the Barrow entropy, with the entropy parameter being $\Delta > 0.86$, close to

Tom Hirschowitz, Ambroise Lafont

We prove a general congruence result for bisimilarity in higher-order languages, which generalises previous work to languages specified by a labelled transition system in which programs may occur as labels, and which may rely on operations on terms other than capture-avoiding substitution. This is typically the case for PCF, $\lambda$-calculus with delimited

Javier Gamazo Tejero, Martin S. Zinkernagel, Sebastian Wolf, Raphael Sznitman

Annotating new datasets for machine learning tasks is tedious, time-consuming, and costly. For segmentation applications, the burden is particularly high as manual delineations of relevant image content are often extremely expensive or can only be done by experts with domain-specific knowledge. Thanks to developments in transfer learning and training with we

Ping-Rui Tsai, Yen-Ting Chou, Nathan-Christopher Wang, Hui-Ling Chen

Music, enchanting and poetic, permeates every corner of human civilization. Although music is not unfamiliar to people, our understanding of its essence remains limited, and there is still no universally accepted scientific description. This is primarily due to music being regarded as a product of both reason and emotion, making it difficult to define. In th

Zhaoyi Xu

Dark matter halos are common in galaxies and globular clusters, making it possible to use dark matter density profiles to construct transverse wormholes. In reference [1], an approximate solution of the traversable wormhole satisfying the isotropic pressure condition is constructed by using the dark matter density profile. However, approximate solutions limi

Tina Yao, Nicole St. Clair, Gabriel F. Miller, Adam L. Dorfman

Purpose: To develop and evaluate an end-to-end deep learning pipeline for segmentation and analysis of cardiac magnetic resonance images to provide core-lab processing for a multi-centre registry of Fontan patients. Materials and Methods: This retrospective study used training (n = 175), validation (n = 25) and testing (n = 50) cardiac magnetic resonance ima

Jintao Guo, Na Wang, Lei Qi, Yinghuan Shi

Domain generalization (DG) aims to learn a model that generalizes well to unseen target domains utilizing multiple source domains without re-training. Most existing DG works are based on convolutional neural networks (CNNs). However, the local operation of the convolution kernel makes the model focus too much on local representations (e.g., texture), which i

Mayukh Bagchi, Subhashis Das

In this paper, we introduce and illustrate the novel phenomenon of Conceptual Entanglement which emerges due to the representational manifoldness immanent while incrementally modelling domain ontologies step-by-step across the following five levels: perception, labelling, semantic alignment, hierarchical modelling and intensional definition. In turn, we prop

Jing Zhang, Chuanwen Li, Jianzgong Qi, Jiayuan He

Federated learning, which allows multiple client devices in a network to jointly train a machine learning model without direct exposure of clients' data, is an emerging distributed learning technique due to its nature of privacy preservation. However, it has been found that models trained with federated learning usually have worse performance than their coun

L. Gehrig, E. I. Vorobyov

We compute the evolution and rotational periods of young stars, using the MESA code, starting from a stellar seed, and take protostellar accretion, stellar winds, and the magnetic star-disk interaction into account. Furthermore, we add a certain fraction of the energy of accreted material into the stellar interior as additional heat and combine the resulting

Eugene Strahov

Let $G$ be a finite group with $k$ conjugacy classes, and $S(\infty)$ be the infinite symmetric group, i.e. the group of finite permutations of $\left\{1,2,3,\ldots\right\}$. Then the wreath product $G_{\infty}=G\sim S(\infty)$ of $G$ with $S(\infty)$ (called the big wreath product) can be defined. The group $G_{\infty}$ is a generalization of the infinite s

Manuel Parra-Royon, Austin Shen, Tristan Reynolds, Parthasarathy Venkataraman

The Widefield ASKAP L-band Legacy All-sky Blind surveY (WALLABY) is a neutral hydrogen survey (HI) that is running on the Australian SKA Pathfinder (ASKAP), a precursor telescope for the Square Kilometre Array (SKA). The goal of WALLABY is to use ASKAP's powerful wide-field phased array feed technology to observe three quarters of the entire sky at the 21 cm

Saeed Saremi, Rupesh Kumar Srivastava, Francis Bach

We consider the problem of generative modeling based on smoothing an unknown density of interest in $\mathbb{R}^d$ using factorial kernels with $M$ independent Gaussian channels with equal noise levels introduced by Saremi and Srivastava (2022). First, we fully characterize the time complexity of learning the resulting smoothed density in $\mathbb{R}^{Md}$,

Yongjin Jeon, Youngtack Oh, Doyoung Jeong, Hyunguk Choi

In recent years, monitoring the world wide area with satellite images has been emerged as an important issue. Site monitoring task can be divided into two independent tasks; 1) Change Detection and 2) Anomaly Event Detection. Unlike to change detection research is actively conducted based on the numerous datasets(\eg LEVIR-CD, WHU-CD, S2Looking, xView2 and e

Shuyi Shi, Kuan-Rong Hao, Xing-Yu Ma, Qing-Bo Yan

${\gamma}$-GeSe is a new type of layered bulk material that was recently successfully synthesized. By means of density functional theory first-principles calculations, we systematically studied the physical properties of two-dimensional (2D) few-layer ${\gamma}$-GeSe. It is found that few-layer ${\gamma}$-GeSe are semiconductors with band gaps decreasing wit

Huishi Luo, Fuzhen Zhuang, Ruobing Xie, Hengshu Zhu

Causal inference has recently garnered significant interest among recommender system (RS) researchers due to its ability to dissect cause-and-effect relationships and its broad applicability across multiple fields. It offers a framework to model the causality in recommender systems like confounding effects and deal with counterfactual problems such as offlin

Akio Kawasaki, Takumi Kobayashi, Akiko Nishiyama, Takehiko Tanabe

We report on the observation of the $4f^{14}6s^{2}~^1S_0- 4f^{13}5d6s^{2}(J=2)$ transition at 431 nm in $^{171}$Yb by depleting atoms in a magneto-optical trap formed by the $6s^{2}~^1S_0-6s6p~^3P_1$ intercombination transition. The absolute frequency of the transition to the $F=3/2$ state is determined to be $695~171~054~858.1(8.2)$~kHz against physical rea

Étienne Fouvry, Emmanuel Kowalski, Philippe Michel

We initiate the study of certain families of $L$-functions attached to characters of subgroups of higher-rank tori, and of their average at the central point. In particular, we evaluate the average of the values $L(\demi,\chi^a)L(\demi,\chi^b)$ for arbitrary integers~$a$ and~$b$ when $\chi$ varies over Dirichlet characters to a prime modulus.

Nicolò Cangiotti, Maicol Caponi, Alberto Maione, Enzo Vitillaro

Classical results concerning Klein-Gordon-Maxwell type systems are shortly reviewed and generalized to the setting of mixed local-nonlocal operators, where the nonlocal one is allowed to be nonpositive definite according to a real parameter. In this paper, we provide a range of parameter values to ensure the existence of solitary (standing) waves, obtained a

Juan Romero, Nick Luiken, Matteo Ravasi

4D seismic inversion is the leading method to quantitatively monitor fluid flow dynamics in the subsurface, with applications ranging from enhanced oil recovery to subsurface CO2 storage. The process of inverting seismic data for reservoir properties is, however, a notoriously ill-posed inverse problem due to the band-limited and noisy nature of seismic data

Fang Hu, Xuexue Sun, Ke Qing, Fenxi Xiao

Although deep learning (DL) shows powerful potential in cell segmentation tasks, it suffers from poor generalization as DL-based methods originally simplified cell segmentation in detecting cell membrane boundary, lacking prominent cellular structures to position overall differentiating. Moreover, the scarcity of annotated cell images limits the performance

Ming Shen, Jie Ma, Shuai Wang, Yogarshi Vyas

Opinion summarization provides an important solution for summarizing opinions expressed among a large number of reviews. However, generating aspect-specific and general summaries is challenging due to the lack of annotated data. In this work, we propose two simple yet effective unsupervised approaches to generate both aspect-specific and general opinion summ

Yoshio Komori, Guoguo Yang, Kevin Burrage

This paper deals with the expectation of monomials with respect to the stochastic area integral $A_{1,2}(t,t+h)=\int_{t}^{t+h}\int_{t}^{s}{\rm d} W_{1}(r){\rm d} W_{2}(s) -\int_{t}^{t+h}\int_{t}^{s}{\rm d} W_{2}(r){\rm d} W_{1}(s)$ and the increments of two Wiener processes, $\Delta{W}_{i}(t,t+h)=W_{i}(t+h)-W_{i}(t),\ i=1,2$. In a monomial, if the exponent o

Athanassios Tzouvaras

We adjust the notion of typicality originated with Russell, which was introduced and studied in a previous paper for general first-order structures, to make it expressible in the language of set theory. The adopted definition of the class ${\rm NT}$ of nontypical sets comes out as a natural strengthening of Russell's initial definition, which employs propert

Chung Kwan Lai, Bhan Lam, Dongyuan Shi, Woon-Seng Gan

The remote microphone technique (RMT) is often used in active noise control (ANC) applications to overcome design constraints in microphone placements by estimating the acoustic pressure at inconvenient locations using a pre-calibrated observation filter (OF), albeit limited to stationary primary acoustic fields. While the OF estimation in varying primary fi

Xuhang Jiang, Li Lin Yang

In this paper, we explore the recursive structure of Baikov representations for Feynman integrals. We demonstrate that the various Baikov representations for all sectors of an integral family can be organized in a tree-like structure. Using this structure, we show that the symbol letters of one-loop Feynman integrals can be written in terms of minors of a ma

Frédéric Chapoton

This article deals with a relationship between derived categories of modules over some partially ordered sets and triangulated categories arising from quasi-homogeneous isolated singularities. It produces heuristics for the existence of derived equivalences between posets, using the geometric category as an auxiliary intermediate. The notion of Weight plays

Carmen Amarra, Alice Devillers, Cheryl E. Praeger

More than $30$ years ago, Delandtsheer and Doyen showed that the automorphism group of a block-transitive $2$-design, with blocks of size $k$, could leave invariant a nontrivial point-partition, but only if the number of points was bounded in terms of $k$. Since then examples have been found where there are two nontrivial point partitions, either forming a c

Usman Nazir, Murtaza Taj, Momin Uppal, Sara Khalid

Industrial air pollution has a direct health impact and is a major contributor to climate change. Small scale industries particularly bull-trench brick kilns are one of the key sources of air pollution in South Asia often creating hazardous levels of smog that is injurious to human health. To mitigate the climate and health impact of the kiln industry, fine-

Paul-Emile Paradan

The main focus of this work is the study of several cones relating the eigenvalues or singular values of a matrix to those of its off-diagonal blocks.

Salman Ashraf

In this paper, we estimate an operator norm of dilation operators on block spaces ($\mathfrak{B}_{r,\alpha}(\mathbb{Q}_p)$) over $p$-adic field. With this estimate, we establish the boundedness of $p$-adic Hardy-Hilbert type integral operator on $\mathfrak{B}_{r,\alpha}(\mathbb{Q}_p)$. Moreover as application to our result, we obtain the $p$-adic Hilbert ine

Zhilong Wang, Junfei Cai, An Chen, Yanqiang Han

The development of modern civil industry, energy and information technology is inseparable from the rapid explorations of new materials, which are hampered by months to years of painstaking attempts, resulting in only a small fraction of materials being determined in a vast chemical space. Artificial intelligence (AI)-based methods are promising to address t

Fabien Lauer

This paper extends standard results from learning theory with independent data to sequences of dependent data. Contrary to most of the literature, we do not rely on mixing arguments or sequential measures of complexity and derive uniform risk bounds with classical proof patterns and capacity measures. In particular, we show that the standard classification r

Yang Zhao, Jianwen Xie, Ping Li

Numerous research efforts have been made to stabilize the training of the Generative Adversarial Networks (GANs), such as through regularization and architecture design. However, we identify the instability can also arise from the fragile balance at the early stage of adversarial learning. This paper proposes the CoopInit, a simple yet effective cooperative

Gustavo Penha, Enrico Palumbo, Maryam Aziz, Alice Wang

An important goal of online platforms is to enable content discovery, i.e. allow users to find a catalog entity they were not familiar with. A pre-requisite to discover an entity, e.g. a book, with a search engine is that the entity is retrievable, i.e. there are queries for which the system will surface such entity in the top results. However, machine-learn

Shima Imani, Harsh Shrivastava

We frequently encounter multiple series that are temporally correlated in our surroundings, such as EEG data to examine alterations in brain activity or sensors to monitor body movements. Segmentation of multivariate time series data is a technique for identifying meaningful patterns or changes in the time series that can signal a shift in the system's behav

Xiangchen Cheng, Wei Tang, Ming Yang, Li Jin

Signal-free intersections are a representative application of smart and connected vehicle technologies. Although extensive results have been developed for trajectory planning and autonomous driving, the formulation and evaluation of vehicle sequencing have not been well understood.In this paper, we consider theoretical guarantees of macroscopic performance (

Zhong-Jun Yang, Tai-Fu Feng, Xing-Gang Wu

We introduce four supersymmetric (SUSY) axion models in which the strong CP problem and the $\mu$ problem are solved with the help of the Peccei-Quinn mechanism and the Kim-Nilles mechanism, respectively. The axion physics enriches the SUSY model by introducing axion as a dark matter candidate and, therefore, the lightest supersymmetric particle (LSP) could

Sandi Klavžar, Gašper Domen Romih

The cut method has been proved to be extremely useful in chemical graph theory. In this paper the cut method is extended to hypergraphs. More precisely, the method is developed for the Wiener index of $k$-uniform partial cube-hypergraphs. The method is applied to cube-hypergraphs and hypertrees. Extensions of the method to hypergraphs arising in chemistry wh

K. Venkatasubbareddy, A. Sankaranarayanan

We study the behavior of the shifted convolution sum involving fourth power of the Fourier coefficients of holomorphic cusp forms with a weight function to be the $k$-full kernel function for any fixed integer $k\geq2$.

Yulong Tian, Fnu Suya, Anshuman Suri, Fengyuan Xu

Transfer learning is a popular method for tuning pretrained (upstream) models for different downstream tasks using limited data and computational resources. We study how an adversary with control over an upstream model used in transfer learning can conduct property inference attacks on a victim's tuned downstream model. For example, to infer the presence of

Muyao Niu, Zhuoxiao Li, Zhihang Zhong, Yinqiang Zheng

Seeing-in-the-dark is one of the most important and challenging computer vision tasks due to its wide applications and extreme complexities of in-the-wild scenarios. Existing arts can be mainly divided into two threads: 1) RGB-dependent methods restore information using degraded RGB inputs only (\eg, low-light enhancement), 2) RGB-independent methods transla

Yepeng Ding, Hiroyuki Sato, Maro G. Machizawa

Data aggregation has been widely implemented as an infrastructure of data-driven systems. However, a centralized data aggregation model requires a set of strong trust assumptions to ensure security and privacy. In recent years, decentralized data aggregation has become realizable based on distributed ledger technology. Nevertheless, the lack of appropriate c

Eduardo C. Garrido-Merchán, Cristina González-Barthe, María Coronado Vaca

In recent years there has been a growing demand from financial agents, especially from particular and institutional investors, for companies to report on climate-related financial risks. A vast amount of information, in text format, can be expected to be disclosed in the short term by firms in order to identify these types of risks in their financial and non

Santiago Jimenez Leudo, Kunal Garg, Ricardo G. Sanfelice, Alvaro A. Cardenas

The design of safe-critical control algorithms for systems under Denial-of-Service (DoS) attacks on the system output is studied in this work. We aim to address scenarios where attack-mitigation approaches are not feasible, and the system needs to maintain safety under adversarial attacks. We propose an attack-recovery strategy by designing a switching obser

Changxing Zhou, Hua Feng, Fuyan Bian

With Very Large Telescope (VLT) Multi Unit Spectroscopic Explorer (MUSE) observations, we detected highly variable helium emission lines from the optical counterpart of the supersoft ultraluminous X-ray source (ULX) NGC 247 ULX-1. No Balmer lines can be seen in the source spectrum. This is the first evidence for the presence of a helium donor star in ULXs, c

Zigang Geng, Chunyu Wang, Yixuan Wei, Ze Liu

Human pose is typically represented by a coordinate vector of body joints or their heatmap embeddings. While easy for data processing, unrealistic pose estimates are admitted due to the lack of dependency modeling between the body joints. In this paper, we present a structured representation, named Pose as Compositional Tokens (PCT), to explore the joint dep

Yang Shen, Xuhao Sun, Xiu-Shen Wei

We propose Equiangular Basis Vectors (EBVs) for classification tasks. In deep neural networks, models usually end with a k-way fully connected layer with softmax to handle different classification tasks. The learning objective of these methods can be summarized as mapping the learned feature representations to the samples' label space. While in metric learni

Kiri Sakahara, Takashi Sato

The present paper attempts to modify the way of constructing a measure in the Alternative Set Theory setting originally devised by Martin Kalina. Introducing a system of cuts of rational numbers extended with some special ones, it is proved that the measure which is nondecreasing, nonnegative and "depending on the way of measurement" as same as Kalina's, but

Wen Huo, Yasutaka Shimizu

In this paper, we discuss the convergence rate of empirical processes of Gaussian processes for a large class of function families. Our main goal is to show that the tail of the uniform norm of the empirical processes can be dominated by polynomials. We put forward the properties of Hermite polynomials which play a crucial role in the proof of main theorems.

Max Peter Ronecker, Yuan Zhu

Currently decision making is one of the biggest challenges in autonomous driving. This paper introduces a method for safely navigating an autonomous vehicle in highway scenarios by combining deep Q-Networks and insight from control theory. A Deep Q-Network is trained in simulation to serve as a central decision-making unit by proposing targets for a trajecto

Zhuotao Tian, Jiequan Cui, Li Jiang, Xiaojuan Qi

Semantic segmentation is still a challenging task for parsing diverse contexts in different scenes, thus the fixed classifier might not be able to well address varying feature distributions during testing. Different from the mainstream literature where the efficacy of strong backbones and effective decoder heads has been well studied, in this paper, addition

Nitish Shukla

Identifying defect patterns in a wafer map during manufacturing is crucial to find the root cause of the underlying issue and provides valuable insights on improving yield in the foundry. Currently used methods use deep neural networks to identify the defects. These methods are generally very huge and have significant inference time. They also require GPU su

Karol Gietka

It is commonly agreed that the electromagnetic vacuum is not empty but filled with virtual photons. This leads to effects like Lamb shift and spontaneous emission. Here we argue that if the vacuum has virtual photons it might mean that it is very weakly squeezed and therefore the electromagnetic field is not in its ground state (vacuum) but in an excited dar

Rui Yang, Lin Song, Yixiao Ge, Xiu Li

Box-supervised instance segmentation has gained much attention as it requires only simple box annotations instead of costly mask or polygon annotations. However, existing box-supervised instance segmentation models mainly focus on mask-based frameworks. We propose a new end-to-end training technique, termed BoxSnake, to achieve effective polygonal instance s

Haotian Liu, Guang Chen, Sanqing Qu, Yanping Zhang

Event cameras have the ability to record continuous and detailed trajectories of objects with high temporal resolution, thereby providing intuitive motion cues for optical flow estimation. Nevertheless, most existing learning-based approaches for event optical flow estimation directly remould the paradigm of conventional images by representing the consecutiv

Lei Zhang, Yue Zhao

We study an inverse source scattering problem for the Schr\"odinger equation with a quadratic nonlinearity. In general, uniqueness of inverse source problems can not be guaranteed at a fixed energy. Therefore, additional information is required for the source in order to obtain a unique solution. By adding reference point sources, we show that a general sour

Maksim V. Kukushkin

In this paper we study non-selfadjoint operators using the methods of the spectral theory. The main challenge is to represent a complete description of an operator belonging to the Schatten-von Neumann class having used the order of the Hermitian real component. The latter fundamental result is advantageous since many theoretical statements based upon it and

Silverio Rosa, Delfim F. M. Torres

In this article, we develop a simple mathematical GNU Octave/MATLAB code that is easy to modify for the simulation of mathematical models governed by fractional-order differential equations, and for the resolution of fractional-order optimal control problems through Pontryagin's maximum principle (indirect approach to optimal control). For this purpose, a fr

Yao Zhu, Yuefeng Chen, Xiaodan Li, Rong Zhang

With the development of deep learning technology, the facial manipulation system has become powerful and easy to use. Such systems can modify the attributes of the given facial images, such as hair color, gender, and age. Malicious applications of such systems pose a serious threat to individuals' privacy and reputation. Existing studies have proposed variou

Muhammad Anwar Ma'sum, Mahardhika Pratama, Edwin Lughofer, Weiping Ding

This paper proposes an assessor-guided learning strategy for continual learning where an assessor guides the learning process of a base learner by controlling the direction and pace of the learning process thus allowing an efficient learning of new environments while protecting against the catastrophic interference problem. The assessor is trained in a meta-

Shuailei Ma, Yuefeng Wang, Ying Wei, Peihao Chen

Open World Object Detection (OWOD) is a novel computer vision task with a considerable challenge, bridging the gap between classic object detection (OD) benchmarks and real-world object detection. In addition to detecting and classifying seen/known objects, OWOD algorithms are expected to detect unseen/unknown objects and incrementally learn them. The natura

Yong Zhong, Zhuoyu Chen, Su-Di Chen, Ke-Jun Xu

The effect of Lifshitz transition on thermodynamics and superconductivity in hole-doped cuprates has been heavily debated but remains an open question. In particular, an observed peak of electronic specific heat is proposed to originate from fluctuations of a putative quantum critical point p* (e.g. the termination of pseudogap at zero temperature), which is

Yiwei Li, Shaoxiong Feng, Bin Sun, Kan Li

With the development of deep learning, advanced dialogue generation methods usually require a greater amount of computational resources. One promising approach to obtaining a high-performance and lightweight model is knowledge distillation, which relies heavily on the pre-trained powerful teacher. Collaborative learning, also known as online knowledge distil

Dhruv Kohli, Gal Mishne, Alexander Cloninger

Given a set of overlapping local views (patches) of a dataset, we consider the problem of finding a rigid alignment of the views that minimizes a $2$-norm based alignment error. In general, the views are noisy and a perfect alignment may not exist. In this work, we characterize the non-degeneracy of an alignment in the noisy setting based on the kernel and p

Joe Hirose

When considering a real log canonical threshold (RLCT) that gives a Bayesian generalization error, in general, papers replace a mean error function with a relatively simple polynomial whose RLCT corresponds to that of the mean error function, and obtain its RLCT by resolving its singularities through an algebraic operation called blow-up. Though it is known

Donghoon Jang, Jiyun Park

An almost complex torus manifold is a $2n$-dimensional compact connected almost complex manifold equipped with an effective action of a real $n$-dimensional torus $T^n \simeq (S^1)^n$ that has fixed points. For an almost complex torus manifold, there is a labeled directed graph which contains information on weights at the fixed points and isotropy spheres. L

Alexandre Magueresse, Santiago Badia

Physics-informed neural networks (PINNs) and their variants have recently emerged as alternatives to traditional partial differential equation (PDE) solvers, but little literature has focused on devising accurate numerical integration methods for neural networks (NNs), which is essential for getting accurate solutions. In this work, we propose adaptive quadr

Hao Ai, Zidong cao, Yan-pei Cao, Ying Shan

Depth estimation from a monocular 360{\deg} image is a burgeoning problem owing to its holistic sensing of a scene. Recently, some methods, \eg, OmniFusion, have applied the tangent projection (TP) to represent a 360{\deg}image and predicted depth values via patch-wise regressions, which are merged to get a depth map with equirectangular projection (ERP) for

Jiawei Wang, Weihong Lin, Chixiang Ma, Mingze Li

We present a new table structure recognition (TSR) approach, called TSRFormer, to robustly recognizing the structures of complex tables with geometrical distortions from various table images. Unlike previous methods, we formulate table separation line prediction as a line regression problem instead of an image segmentation problem and propose a new two-stage

Zhuo Cao, Hua-bai Li

Magnetic fields (B-fields) play an important role in molecular cloud fragmentation and star formation, but are very difficult to detect. The temporal correlation between the field strength (B) and gas density (n) of an isolated cloud has been suggested as an indication of the dynamical importance of B-fields relative to self-gravity. This temporal B-n relati

Kohki Kawabata, Shinichiro Yahagi

We construct a class of chiral fermionic CFTs from classical codes over finite fields whose order is a prime number. We exploit the relationship between classical codes and Euclidean lattices to provide the Neveu-Schwarz sector of fermionic CFTs. On the other hand, we construct the Ramond sector using the shadow theory of classical codes and Euclidean lattic

Maolin Che, Yimin Wei, Hong Yan

This paper develops fast and efficient algorithms for computing Tucker decomposition with a given multilinear rank. By combining random projection and the power scheme, we propose two efficient randomized versions for the truncated high-order singular value decomposition (T-HOSVD) and the sequentially T-HOSVD (ST-HOSVD), which are two common algorithms for a

Yuzheng Wang, Zhaoyu Chen, Dingkang Yang, Pinxue Guo

Adversarial Robustness Distillation (ARD) is a promising task to solve the issue of limited adversarial robustness of small capacity models while optimizing the expensive computational costs of Adversarial Training (AT). Despite the good robust performance, the existing ARD methods are still impractical to deploy in natural high-security scenes due to these

Luigi Riz, Cristiano Saltori, Elisa Ricci, Fabio Poiesi

Novel class discovery (NCD) for semantic segmentation is the task of learning a model that can segment unlabelled (novel) classes using only the supervision from labelled (base) classes. This problem has recently been pioneered for 2D image data, but no work exists for 3D point cloud data. In fact, the assumptions made for 2D are loosely applicable to 3D in

Yunzhuo Guo, Cheng Wang, Steven M. Wise, Zhengru Zhang

A finite difference numerical scheme is proposed and analyzed for the Cahn-Hilliard-Stokes system with Flory-Huggins energy functional. A convex splitting is applied to the chemical potential, which in turns leads to the implicit treatment for the singular logarithmic terms and the surface diffusion term, and an explicit update for the expansive concave term