arXiv subjects

Xingyu Li

Publications and source records attributed to Xingyu Li.

152 records · Page 9Linked to original sources

arXiv2020

Discriminative Pattern Mining for Breast Cancer Histopathology Image Classification via Fully Convolutional Autoencoder

Accurate diagnosis of breast cancer in histopathology images is challenging due to the heterogeneity of cancer cell growth as well as of a variety of benign breast tissue proliferative lesions. In this paper, we propose a practical and self-interpretable invasive cancer diagnosis solution. With minimum annotation information, the proposed method mines contrast patterns between normal and malignant images in unsupervised manner and generates a probability map of abnormalities to verify its reasoning. Particularly, a fully convolutional autoencoder is used to learn the dominant structural patterns among normal image patches. Patches that do not share the characteristics of this normal population are detected and analyzed by one-class support vector machine and 1-layer neural network. We apply the proposed method to a public breast cancer image set. Our results, in consultation with a senior pathologist, demonstrate that the proposed method outperforms existing methods. The obtained probability map could benefit the pathology practice by providing visualized verification data and potentially leads to a better understanding of data-driven diagnosis solutions.

cs.CV↗

arXiv2019

Generalized logarithmic Hardy-Littlewood-Sobolev inequality

This paper is devoted to logarithmic Hardy-Littlewood-Sobolev inequalities in the two-dimensional Euclidean space, in presence of an external potential with logarithmic growth. The coupling with the potential introduces a new parameter, with two regimes. The attractive regime reflects the standard logarithmic Hardy-Littlewood-Sobolev inequality. The second regime corresponds to a reverse inequality, with the opposite sign in the convolution term, that allows us to bound the free energy of a drift-diffusion-Poisson system from below. Our method is based on an extension of an entropy method proposed by E. Carlen, J. Carrillo and M. Loss, and on a nonlinear diffusion equation.

math.AP↗

arXiv2019

Asymptotic behavior of Nernst-Planck equation

This paper is devoted to the Nernst-Planck system of equations with an external potential of confinement. The main result is concerned with the asymptotic behaviour of the solution of the Cauchy problem. We will prove that the optimal exponential rate of convergence of the solution to the unique stationary solution is determined by the spectral gap of the linearized problem around the minimizer of the free energy. The key issue is to consider an adapted notion of scalar product.

math.AP↗

arXiv2019

End-to-End Open-Domain Question Answering with BERTserini

We demonstrate an end-to-end question answering system that integrates BERT with the open-source Anserini information retrieval toolkit. In contrast to most question answering and reading comprehension models today, which operate over small amounts of input text, our system integrates best practices from IR with a BERT-based reader to identify answers from a large corpus of Wikipedia articles in an end-to-end fashion. We report large improvements over previous results on a standard benchmark test collection, showing that fine-tuning pretrained BERT with SQuAD is sufficient to achieve high accuracy in identifying answer spans.

cs.CL↗

arXiv2019

An Agent-Based Approach for Optimizing Modular Vehicle Fleet Operation

Modularity in military vehicle designs enables on-base assembly, disassembly, and reconfiguration of vehicles, which can be beneficial in promoting fleet adaptability and life cycle cost savings. To properly manage the fleet operation and to control the resupply, demand prediction, and scheduling process, this paper illustrates an agent-based approach customized for highly modularized military vehicle fleets and studies the feasibility and flexibility of modularity for various mission scenarios. Given deterministic field demands with operation stochasticity, we compare the performance of a modular fleet to a conventional fleet in equivalent operation strategies and also compare fleet performance driven by heuristic rules and optimization. Several indicators are selected to quantify the fleet performance, including operation costs, total resupplied resources, and fleet readiness. When the model is implemented for military Joint Tactical Transport System (JTTS) mission, our results indicate that fleet modularity can reduce total resource supplies without significant losses in fleet readiness. The benefits of fleet modularity can also be amplified through a real-time optimized operation strategy. To highlight the feasibility of fleet modularity, a parametric study is performed to show the impacts from working capacity on modular fleet performance. Finally, we provide practical suggestions of modular vehicle designs based on the analysis and other possible usage.

eess.SY↗

arXiv2019

Analysis of the Synergy between Modularity and Autonomy in an Artificial Intelligence Based Fleet Competition

A novel approach is provided for evaluating the benefits and burdens from vehicle modularity in fleets/units through the analysis of a game theoretical model of the competition between autonomous vehicle fleets in an attacker-defender game. We present an approach to obtain the heuristic operational strategies through fitting a decision tree on high-fidelity simulation results of an intelligent agent-based model. A multi-stage game theoretical model is also created for decision making considering military resources and impacts of past decisions. Nash equilibria of the operational strategy are revealed, and their characteristics are explored. The benefits of fleet modularity are also analyzed by comparing the results of the decision making process under diverse operational situations.

cs.AI↗

arXiv2019

Analysis of Fleet Modularity in an Artificial Intelligence-Based Attacker-Defender Game

Because combat environments change over time and technology upgrades are widespread for ground vehicles, a large number of vehicles and equipment become quickly obsolete. A possible solution for the U.S. Army is to develop fleets of modular military vehicles, which are built by interchangeable substantial components also known as modules. One of the typical characteristics of module is their ease of assembly and disassembly through simple means such as plug-in/pull-out actions, which allows for real-time fleet reconfiguration to meet dynamic demands. Moreover, military demands are time-varying and highly stochastic because commanders keep reacting to enemy's actions. To capture these characteristics, we formulated an intelligent agent-based model to imitate decision making process during fleet operation, which combines real-time optimization with artificial intelligence. The agents are capable of inferring enemy's future move based on historical data and optimize dispatch/operation decisions accordingly. We implement our model to simulate an attacker-defender game between two adversarial and intelligent players, representing the commanders from modularized fleet and conventional fleet respectively. Given the same level of combat resources and intelligence, we highlight the tactical advantages of fleet modularity in terms of win rate, unpredictability and suffered damage.

cs.AI↗

arXiv2018

Phi-entropies for Fokker-Planck and kinetic Fokker-Planck equations

This paper is devoted to $ϕ$-entropies applied to Fokker-Planck and kinetic Fokker-Planck equations in the whole space, with confinement. The so-called $ϕ$-entropies are Lyapunov functionals which typically interpolate between Gibbs entropies and L2 estimates. We review some of their properties in the case of diffusion equations of Fokker-Planck type, give new and simplified proofs, and then adapt these methods to a kinetic Fokker-Planck equation acting on a phase space with positions and velocities. At kinetic level, since the diffusion only acts on the velocity variable, the transport operator plays an essential role in the relaxation process. Here we adopt the H1 point of view and establish a sharp decay rate. Rather than giving general but quantitatively vague estimates, our goal here is to consider simple cases, benchmark available methods and obtain sharp estimates on a key example. Some $ϕ$-entropies give rise to improved entropy -- entropy production inequalities and, as a consequence, to faster decay rates for entropy estimates of solutions to non-degenerate diffusion equations. Our main result is to prove that faster entropy decay also holds at kinetic level and that optimal decay rates are achieved only in asymptotic regimes.

math.AP↗