arXiv subjects

Le Li

Publications and source records attributed to Le Li.

27 records · Page 2Linked to original sources

arXiv2019

Facial Expression Restoration Based on Improved Graph Convolutional Networks

Facial expression analysis in the wild is challenging when the facial image is with low resolution or partial occlusion. Considering the correlations among different facial local regions under different facial expressions, this paper proposes a novel facial expression restoration method based on generative adversarial network by integrating an improved graph convolutional network (IGCN) and region relation modeling block (RRMB). Unlike conventional graph convolutional networks taking vectors as input features, IGCN can use tensors of face patches as inputs. It is better to retain the structure information of face patches. The proposed RRMB is designed to address facial generative tasks including inpainting and super-resolution with facial action units detection, which aims to restore facial expression as the ground-truth. Extensive experiments conducted on BP4D and DISFA benchmarks demonstrate the effectiveness of our proposed method through quantitative and qualitative evaluations.

cs.CV↗

arXiv2019

Sequential Learning of Principal Curves: Summarizing Data Streams on the Fly

When confronted with massive data streams, summarizing data with dimension reduction methods such as PCA raises theoretical and algorithmic pitfalls. Principal curves act as a nonlinear generalization of PCA and the present paper proposes a novel algorithm to automatically and sequentially learn principal curves from data streams. We show that our procedure is supported by regret bounds with optimal sublinear remainder terms. A greedy local search implementation (called \texttt{slpc}, for Sequential Learning Principal Curves) that incorporates both sleeping experts and multi-armed bandit ingredients is presented, along with its regret computation and performance on synthetic and real-life data.

stat.ML↗

arXiv2018

A Quasi-Bayesian Perspective to Online Clustering

When faced with high frequency streams of data, clustering raises theoretical and algorithmic pitfalls. We introduce a new and adaptive online clustering algorithm relying on a quasi-Bayesian approach, with a dynamic (i.e., time-dependent) estimation of the (unknown and changing) number of clusters. We prove that our approach is supported by minimax regret bounds. We also provide an RJMCMC-flavored implementation (called PACBO, see https://cran.r-project.org/web/packages/PACBO/index.html) for which we give a convergence guarantee. Finally, numerical experiments illustrate the potential of our procedure.

stat.ML↗

arXiv2018

Diversification, economies of scope, and exports growth of Chinese firms

In the 1990s, China started a process of structural reforms and of trade liberalization, which was followed by the accession to the World Trade Organization (WTO) in 2001. In this paper, we analyze trade patterns of Chinese firms for the period 2000-2006, characterized by a notable increase in exports volumes. Theoretically, in a more open economy, firms are expected to move from the production of a set of less-competitive products towards more internationally competitive ones, which implies specialization. We study several stylized facts on the distribution of Chinese firms trade and growth rates, and we analyze whether firms have diversified or specialized their trade patterns between 2000 and 2006. We show that Chinese export patterns are very heterogeneous, that the volatility of growth rates depends on the level of exports, and that volatility is stronger after trade liberalization. Both, diversification in products and destinations have a positive impact on trade growth, but diversification of destinations has a stronger effect. We conclude that the success of Chinese exports is not only due to an increase in the intensive margin, related to the existence of economies of scale, but also due to an increase in the extensive margin, related to the existence of economies of scope.

econ.GN↗

arXiv2014

Strong shape dependence of the Morin transition in alpha-Fe2O3 single-crystalline nanostructures

Single-crystalline alpha-Fe2O3 nanorings (short nanotubes) and nanotubes were synthesized by a hydrothermal method. High-resolution transmission electron microscope and selected-area electron diffraction confirm that the axial directions of both nanorings and nanotubes are parallel to the crystalline c-axis. What is intriguing is that the Morin transition occurs at about 210 K in the short nanotubes with a mean tube length of about 115 nm and a mean outer diameter of 169 nm while it disappears in the nanotubes with a mean tube length of about 317 nm and a mean outer diameter of 148 nm. Detailed analyses of magnetization data, x-ray diffraction spectra, and room-temperature Mossbauer spectra demonstrate that this very strong shape dependence of the Morin transition is intrinsic to hematite. We can quantitatively explain this intriguing shape dependence in terms of opposite signs of the surface magnetic anisotropy constants in the surface planes parallel and perpendicular to the c-axis (that is, K_parallel = -0.37 erg/cm^2 and K_perp = 0.42 erg/cm^{2}).

cond-mat.mes-hall↗

arXiv2014

Document Clustering Based On Max-Correntropy Non-Negative Matrix Factorization

Nonnegative matrix factorization (NMF) has been successfully applied to many areas for classification and clustering. Commonly-used NMF algorithms mainly target on minimizing the $l_2$ distance or Kullback-Leibler (KL) divergence, which may not be suitable for nonlinear case. In this paper, we propose a new decomposition method by maximizing the correntropy between the original and the product of two low-rank matrices for document clustering. This method also allows us to learn the new basis vectors of the semantic feature space from the data. To our knowledge, we haven't seen any work has been done by maximizing correntropy in NMF to cluster high dimensional document data. Our experiment results show the supremacy of our proposed method over other variants of NMF algorithm on Reuters21578 and TDT2 databasets.

cs.IR↗

arXiv2014

Location Aided Energy Balancing Strategy in Green Cellular Networks

Most cellular network communication strategies are focused on data traffic scenarios rather than energy balance and efficient utilization. Thus mobile users in hot cells may suffer from low throughput due to energy loading imbalance problem. In state of art cellular network technologies, relay stations extend cell coverage and enhance signal strength for mobile users. However, busy traffic makes the relay stations in hot area run out of energy quickly. In this paper, we propose an energy balancing strategy in which the mobile nodes are able to dynamically select and hand over to the relay station with the highest potential energy capacity to resume communication. Key to the strategy is that each relay station merely maintains two parameters that contains the trend of its previous energy consumption and then predicts its future quantity of energy, which is defined as the relay station potential energy capacity. Then each mobile node can select the relay station with the highest potential energy capacity. Simulations demonstrate that our approach significantly increase the aggregate throughput and the average life time of relay stations in cellular network environment.

cs.NI↗

arXiv2014

Graph Regularized Non-negative Matrix Factorization By Maximizing Correntropy

Non-negative matrix factorization (NMF) has proved effective in many clustering and classification tasks. The classic ways to measure the errors between the original and the reconstructed matrix are $l_2$ distance or Kullback-Leibler (KL) divergence. However, nonlinear cases are not properly handled when we use these error measures. As a consequence, alternative measures based on nonlinear kernels, such as correntropy, are proposed. However, the current correntropy-based NMF only targets on the low-level features without considering the intrinsic geometrical distribution of data. In this paper, we propose a new NMF algorithm that preserves local invariance by adding graph regularization into the process of max-correntropy-based matrix factorization. Meanwhile, each feature can learn corresponding kernel from the data. The experiment results of Caltech101 and Caltech256 show the benefits of such combination against other NMF algorithms for the unsupervised image clustering.

cs.CV↗

arXiv2013

Adaptive Learning of Region-based pLSA Model for Total Scene Annotation

In this paper, we present a region-based pLSA model to accomplish the task of total scene annotation. To be more specific, we not only properly generate a list of tags for each image, but also localizing each region with its corresponding tag. We integrate advantages of different existing region-based works: employ efficient and powerful JSEG algorithm for segmentation so that each region can easily express meaningful object information; the introduction of pLSA model can help better capturing semantic information behind the low-level features. Moreover, we also propose an adaptive padding mechanism to automatically choose the optimal padding strategy for each region, which directly increases the overall system performance. Finally we conduct 3 experiments to verify our ideas on Corel database and demonstrate the effectiveness and accuracy of our system.

cs.CV↗