arXiv subjects

Xin Guo

Publications and source records attributed to Xin Guo.

At least 199 records · Page 11Linked to original sources

arXiv2018

Smile detection in the wild based on transfer learning

Smile detection from unconstrained facial images is a specialized and challenging problem. As one of the most informative expressions, smiles convey basic underlying emotions, such as happiness and satisfaction, which lead to multiple applications, e.g., human behavior analysis and interactive controlling. Compared to the size of databases for face recognition, far less labeled data is available for training smile detection systems. To leverage the large amount of labeled data from face recognition datasets and to alleviate overfitting on smile detection, an efficient transfer learning-based smile detection approach is proposed in this paper. Unlike previous works which use either hand-engineered features or train deep convolutional networks from scratch, a well-trained deep face recognition model is explored and fine-tuned for smile detection in the wild. Three different models are built as a result of fine-tuning the face recognition model with different inputs, including aligned, unaligned and grayscale images generated from the GENKI-4K dataset. Experiments show that the proposed approach achieves improved state-of-the-art performance. Robustness of the model to noise and blur artifacts is also evaluated in this paper.

cs.CV↗

arXiv2017

On risk-sensitive piecewise deterministic Markov decision processes

We consider a piecewise deterministic Markov decision process, where the expected exponential utility of total (nonnegative) cost is to be minimized. The cost rate, transition rate and post-jump distributions are under control. The state space is Borel, and the transition and cost rates are locally integrable along the drift. Under natural conditions, we establish the optimality equation, justify the value iteration algorithm, and show the existence of a deterministic stationary optimal policy. Applied to special cases, the obtained results already significantly improve some existing results in the literature on finite horizon and infinite horizon discounted risk-sensitive continuous-time Markov decision processes.

math.OC↗

arXiv2017

Ambiguity set and learning via Bregman and Wasserstein

Construction of ambiguity set in robust optimization relies on the choice of divergences between probability distributions. In distribution learning, choosing appropriate probability distributions based on observed data is critical for approximating the true distribution. To improve the performance of machine learning models, there has recently been interest in designing objective functions based on Lp-Wasserstein distance rather than the classical Kullback-Leibler (KL) divergence. In this paper, we derive concentration and asymptotic results using Bregman divergence. We propose a novel asymmetric statistical divergence called Wasserstein-Bregman divergence as a generalization of L2-Wasserstein distance. We discuss how these results can be applied to the construction of ambiguity set in robust optimization.

stat.ML↗

arXiv2017

Zero-sum continuous-time Markov pure jump game over a fixed duration

This paper considers a two-person zero-sum continuous-time Markov pure jump game in Borel state and action spaces over a fixed finite horizon. The main assumption on the model is the existence of a drift function, which bounds the reward rate. Under some regularity conditions, we show that the game has a value, and both of the players have their optimal policies.

math.OC↗

arXiv2017

Distributed learning with regularized least squares

We study distributed learning with the least squares regularization scheme in a reproducing kernel Hilbert space (RKHS). By a divide-and-conquer approach, the algorithm partitions a data set into disjoint data subsets, applies the least squares regularization scheme to each data subset to produce an output function, and then takes an average of the individual output functions as a final global estimator or predictor. We show with error bounds in expectation in both the $L^2$-metric and RKHS-metric that the global output function of this distributed learning is a good approximation to the algorithm processing the whole data in one single machine. Our error bounds are sharp and stated in a general setting without any eigenfunction assumption. The analysis is achieved by a novel second order decomposition of operator differences in our integral operator approach. Even for the classical least squares regularization scheme in the RKHS associated with a general kernel, we give the best learning rate in the literature.

cs.LG↗

arXiv2016

Note on discounted continuous-time Markov decision processes with a lower bounding function

In this paper, we consider the discounted continuous-time Markov decision process (CTMDP) with a lower bounding function. In this model, the negative part of each cost rate is bounded by the drift function, say $w$, whereas the positive part is allowed to be arbitrarily unbounded. Our focus is on the existence of a stationary optimal policy for the discounted CTMDP problems out of the more general class. Both constrained and unconstrained problems are considered. Our investigations are based on a useful transformation for nonhomogeneous Markov pure jump processes that has not yet been widely applied to the study of CTMDPs. This technique was not employed in previous literature, but it clarifies the roles of the imposed conditions in a rather transparent way. As a consequence, we withdraw and weaken several conditions commonly imposed in the literature.

math.OC↗

arXiv2016

Fluid-structure interaction simulation of floating structures interacting with complex, large-scale ocean waves and atmospheric turbulence

We develop a numerical method for simulating coupled interactions of complex floating structures with large-scale ocean waves and atmospheric turbulence. We employ an efficient large-scale model to develop offshore wind and wave environmental conditions, which are then incorporated into a high resolution two-phase flow solver with fluid-structure interaction (FSI). The large-scale wind-wave interaction model is based on the two-fluid dynamically-coupled approach of Yang and Shen (2011), which employs a high-order spectral method for simulating the water motion and a viscous solver with undulatory boundaries for the air motion. The two-phase flow FSI solver, developed by Calderer, Kang, and Sotiropoulos (2014), is based on the level set method and is capable of simulating the coupled dynamic interaction of arbitrarily complex bodies with airflow and waves. The large-scale wave field solver is coupled with the near-field FSI solver by feeding into the latter waves via the pressure-forcing method of Guo and Shen (2009), which has been extended herein for the level set method. We validate the model for both simple wave trains and three-dimensional directional waves and compare the results with experimental and theoretical solutions. Finally, we demonstrate the capabilities of the new computational framework by carrying out large-eddy simulation of a floating offshore wind turbine interacting with realistic ocean wind and waves.

physics.flu-dyn↗

arXiv2016

Decomposition and Analysis of Technological domains for better understanding of Technological Structure

Patents represent one of the most complete sources of information related to technological change. This paper presents three months of research on U.S. patents in the field of patent analysis. The methodology consists of using search terms to locate the most representative international and US patent classes and determines the overlap of those classes to arrive at the final set of patents and using the prediction model developed by Benson and Magee to calculate the technological improvement rate for the technological domains. My research focused on the Biochemical Pharmacology technological area and selecting relevant patents for technological domains and sub-domains within this area. The goal is to better understand structure of technology domain and understand how fast the domains and their sub-domains progress. The method I used is developed by Benson and Magee which is called the Classification Overlap Method1, it provides a reliable and largely automated way to break the patent database into understandable technological domains where progress can be measured.

cs.DL↗

arXiv2016

Observational evidence for the evolution of nuclear metallicity and star formation rate as the merger stage

We investigate the evolution of nuclear gas-phase oxygen abundance and star formation rate (SFR) of local far-infrared selected star-forming galaxies along the merger sequence, as traced by their optical morphologies. The sample was drawn from a cross-correlation analysis of the IRAS Point Source Catalog Redshift Survey and 1 Jy ultraluminous infrared galaxies sample with the Sloan Digital Sky Survey Data Release 7 database. The investigation is done by comparing our sample to a control sample matched in the normalized redshift distribution in two diagnostics, which are the nuclear gas-phase metallicity vs. stellar mass and the nuclear SFR vs. stellar mass diagrams. Galaxies with different morphological types show different mass-metallicity relations (MZR). Compared to the MZR defined by the control sample, isolated spirals have comparable metallicities with the control sample at a given stellar mass. Spirals in pairs and interacting galaxies with projected separations $r_{p} >$ 20 kpc show mild metallicity dilution of 0.02-0.03 dex. Interacting galaxies with $r_{p} <$ 20 kpc, pre-mergers and advanced mergers are under-abundant by ~0.06, ~0.05 and ~0.04 dex, respectively. This shows an evolutionary trend that the metallicity is increasingly depressed as the merging proceeds and it is diluted most dramatically when two galaxies are closely interacting. Afterwards, the interstellar medium (ISM) is enriched when the galaxies coalesce. This is the first time that such ISM enrichment at the final coalescence stage is observed, which demonstrates the importance of supernova explosion in affecting the nuclear metallicity. Meanwhile the central SFR enhancement relative to the control sample evolves simultaneously with the nuclear gas-phase oxygen abundance. Our results support the predictions from numerical simulations.

astro-ph.GA↗

arXiv2015

Dynamics of Order Positions and Related Queues in a Limit Order Book

Order positions are key variables in algorithmic trading. This paper studies the limiting behavior of order positions and related queues in a limit order book. In addition to the fluid and diffusion limits for the processes, fluctuations of order positions and related queues around their fluid limits are analyzed. As a corollary, explicit analytical expressions for various quantities of interests in a limit order book are derived.

q-fin.TR↗

arXiv2014

Martingale Problem under Nonlinear Expectations

We formulate and solve the martingale problem in a nonlinear expectation space. Unlike the classical work of Stroock and Varadhan (1969) where the linear operator in the associated PDE is naturally defined from the corresponding diffusion process, the main difficulty in the nonlinear setting is to identify an appropriate class of nonlinear operators for the associated fully nonlinear PDEs. Based on the analysis of the martingale problem, we introduce the notion of weak solution for stochastic differential equations under nonlinear expectations and obtain an existence theorem under the Hölder continuity condition of the coefficients. The approach to establish the existence of weak solutions generalizes the classical Girsanov transformation method in that it no longer requires the two (probability) measures to be absolutely continuous.

math.PR↗

arXiv2013

Electrical Tuning of Surface Plasmon Polariton Propagation in Graphene-Nanowire Hybrid Structure

We demonstrate a dynamic surface plasmonic modulation of graphene-nanowire hybrid structures in visible light range, which was thought to be a tough task for graphene based field effect transistor modulator previously. Static modulation depth of as high as 0.07 dB/μm has been achieved experimentally. Carefully simulation indicates the strongly focused electromagnetic field and dramatically enhanced electric field at the interface between a silver NW and a graphene sheet are key roles for bringing the optical response of the device to the visible range. Furthermore, the modulation behaviors near the Dirac point of monolayer graphene and the singularity of gap-induced bilayer graphene are investigated.

cond-mat.mes-hall↗

arXiv2013

A Note on $G$- Optimal Stopping Problems

We consider a class of discretionary stopping problems within the $G$-framework. We first establish the well-definedness of the stopping problem under the $G$-expectation, by showing the quasi-continuity of the stopped process. We then prove a verification theorem for $G$-optimal stopping problem. One corollary is a direct proof for the well-known fact that the $G$-optimal stopping problem is the same as the classical optimal stopping problem with appropriate parameters, when the payoff function is concave or convex.

math.PR↗

arXiv2013

Impulse Control of Multi-dimensional Jump Diffusions in Finite Time Horizon

This paper analyzes a class of impulse control problems for multi-dimensional jump diffusions in the finite time horizon. Following the basic mathematical setup from Stroock and Varadhan \cite{StroockVaradhan06}, this paper first establishes rigorously an appropriate form of Dynamic Programming Principle (DPP). It then shows that the value function is a viscosity solution for the associated Hamilton-Jacobi-Belleman (HJB) equation involving integro-differential operators. Finally, under additional assumptions that the jumps are of infinite activity but are of finite variation and that the diffusion is uniformly elliptic, it proves that the value function is the unique viscosity solution and has $W_{loc}^{(2,1),p}$ regularity for $1< p< \infty$

math.OC↗

arXiv2012

Towards a Mathematical Foundation of Immunology and Amino Acid Chains

We attempt to set a mathematical foundation of immunology and amino acid chains. To measure the similarities of these chains, a kernel on strings is defined using only the sequence of the chains and a good amino acid substitution matrix (e.g. BLOSUM62). The kernel is used in learning machines to predict binding affinities of peptides to human leukocyte antigens DR (HLA-DR) molecules. On both fixed allele (Nielsen and Lund 2009) and pan-allele (Nielsen et.al. 2010) benchmark databases, our algorithm achieves the state-of-the-art performance. The kernel is also used to define a distance on an HLA-DR allele set based on which a clustering analysis precisely recovers the serotype classifications assigned by WHO (Nielsen and Lund 2009, and Marsh et.al. 2010). These results suggest that our kernel relates well the chain structure of both peptides and HLA-DR molecules to their biological functions, and that it offers a simple, powerful and promising methodology to immunology and amino acid chain studies.

stat.ML↗

arXiv2011

The economic default time and the Arcsine law

This paper develops a structural credit risk model to characterize the difference between the economic and recorded default times for a firm. Recorded default occurs when default is recorded in the legal system. The economic default time is the last time when the firm is able to pay off its debt prior to the legal default time. It has been empirically documented that these two times are distinct (see Guo, Jarrow, and Lin (2008)). In our model, the probability distribution for the time span between economic and recorded defaults follows a mixture of Arcsine Laws, which is consistent with the results contained in Guo, Jarrow, and Lin. In addition, we show that the classical structural model is a limiting case of our model as the time period between debt repayment dates goes to zero. As a corollary, we show how the firm value process's parameters can be estimated using the tail index and correlation structure of the firm's return.

q-fin.RM↗

arXiv2009

Impulse Control of Multidimensional Jump Diffusions

This paper studies regularity property of the value function for an infinite-horizon discounted cost impulse control problem, where the underlying controlled process is a multidimensional jump diffusion with possibly `infinite-activity' jumps. Surprisingly, despite these jumps, we obtain the same degree of regularity as for the diffusion case, at least when the jump satisfies certain integrability conditions.

math.OC↗

arXiv2008

Graphical models for correlated defaults

A simple graphical model for correlated defaults is proposed, with explicit formulas for the loss distribution. Algebraic geometry techniques are employed to show that this model is well posed for default dependence: it represents any given marginal distribution for single firms and pairwise correlation matrix. These techniques also provide a calibration algorithm based on maximum likelihood estimation. Finally, the model is compared with standard normal copula model in terms of tails of the loss distribution and implied correlation smile.

q-fin.CP↗