arXiv subjects

Han Li

Publications and source records attributed to Han Li.

At least 361 records · Page 20Linked to original sources

arXiv2020

Explicit Result on Equivalence of Rational Quadratic Forms Avoiding Primes

Given a pair of regular quadratic forms over $\mathbb{Q}$ which are in the same genus and a finite set of primes $P$, we show that there is an effective way to determine a rational equivalence between these two quadratic forms which are integral over every prime in $P$. This answers one of the principal questions posed by Conway and Sloane in their book {\em Sphere packings, lattices and groups}, Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], Vol 290, Springer-Verlag, New York, 1999; page 402.

math.NT↗

arXiv2020

Bounding Maps for Universal Lesion Detection

Universal Lesion Detection (ULD) in computed tomography plays an essential role in computer-aided diagnosis systems. Many detection approaches achieve excellent results for ULD using possible bounding boxes (or anchors) as proposals. However, empirical evidence shows that using anchor-based proposals leads to a high false-positive (FP) rate. In this paper, we propose a box-to-map method to represent a bounding box with three soft continuous maps with bounds in x-, y- and xy- directions. The bounding maps (BMs) are used in two-stage anchor-based ULD frameworks to reduce the FP rate. In the 1 st stage of the region proposal network, we replace the sharp binary ground-truth label of anchors with the corresponding xy-direction BM hence the positive anchors are now graded. In the 2 nd stage, we add a branch that takes our continuous BMs in x- and y- directions for extra supervision of detailed locations. Our method, when embedded into three state-of-the-art two-stage anchor-based detection methods, brings a free detection accuracy improvement (e.g., a 1.68% to 3.85% boost of sensitivity at 4 FPs) without extra inference time.

cs.CV↗

arXiv2020

Routing of valley photons in a WS2 monolayer via delocalized Bloch modes of in-plane inversion-symmetry broken photonic crystal slabs

The valleys of two-dimensional transition metal dichalcogenides (TMDCs) offer a new degree of freedom for information processing. To take advantage of this valley degree of freedom, on one hand, it is feasible to control valleys by utilizing different external stimuli like optical and electric fields. On the other hand, nanostructures are also used to separate the valleys by near field coupling. However, for both above methods, either required low-temperature environment or low degree of coherence properties limit their further applications. Here, we demonstrate all-dielectric photonic crystal (PhC) slabs without in-plane inversion symmetry (C2 symmetry) could separate and route valley photons in a WS2 monolayer at room temperature. Coupling with circularly polarized photonic Bloch modes of such PhC slabs, valley photons emitted by a WS2 monolayer are routed directionally and efficiently separated in the far field. In addition, the far-field emission is directionally enhanced and with long-distance spatial coherence property.

physics.optics↗

arXiv2020

Dynamic Knapsack Optimization Towards Efficient Multi-Channel Sequential Advertising

In E-commerce, advertising is essential for merchants to reach their target users. The typical objective is to maximize the advertiser's cumulative revenue over a period of time under a budget constraint. In real applications, an advertisement (ad) usually needs to be exposed to the same user multiple times until the user finally contributes revenue (e.g., places an order). However, existing advertising systems mainly focus on the immediate revenue with single ad exposures, ignoring the contribution of each exposure to the final conversion, thus usually falls into suboptimal solutions. In this paper, we formulate the sequential advertising strategy optimization as a dynamic knapsack problem. We propose a theoretically guaranteed bilevel optimization framework, which significantly reduces the solution space of the original optimization space while ensuring the solution quality. To improve the exploration efficiency of reinforcement learning, we also devise an effective action space reduction approach. Extensive offline and online experiments show the superior performance of our approaches over state-of-the-art baselines in terms of cumulative revenue.

cs.LG↗

arXiv2020

Learning Optimal Tree Models Under Beam Search

Retrieving relevant targets from an extremely large target set under computational limits is a common challenge for information retrieval and recommendation systems. Tree models, which formulate targets as leaves of a tree with trainable node-wise scorers, have attracted a lot of interests in tackling this challenge due to their logarithmic computational complexity in both training and testing. Tree-based deep models (TDMs) and probabilistic label trees (PLTs) are two representative kinds of them. Though achieving many practical successes, existing tree models suffer from the training-testing discrepancy, where the retrieval performance deterioration caused by beam search in testing is not considered in training. This leads to an intrinsic gap between the most relevant targets and those retrieved by beam search with even the optimally trained node-wise scorers. We take a first step towards understanding and analyzing this problem theoretically, and develop the concept of Bayes optimality under beam search and calibration under beam search as general analyzing tools for this purpose. Moreover, to eliminate the discrepancy, we propose a novel algorithm for learning optimal tree models under beam search. Experiments on both synthetic and real data verify the rationality of our theoretical analysis and demonstrate the superiority of our algorithm compared to state-of-the-art methods.

stat.ML↗

arXiv2020

Universal Thermodynamics in the Kitaev Fractional Liquid

In the Kitaev honeycomb model, the quantum spin fractionalizes into itinerant Majorana and gauge flux spontaneously upon cooling, leading to rich experimental ramifications at finite temperature and an upsurge of research interest. In this work, we employ the exponential tensor renormalization group approach to explore the Kitaev model under various perturbations, including the external fields, Heisenberg, and the off-diagonal couplings that are common in the Kitaev materials. Through large-scale manybody calculations, we find a Kitaev fractional liquid at intermediate temperature that is robust against perturbations. The fractional liquid exhibits universal thermodynamic behaviors, including the fractional thermal entropy, metallic specific heat, and an intermediate-temperature Curie law of magnetic susceptibility. The emergent universal susceptibility behavior, with a modified Curie constant, can be ascribed to the strongly fluctuating $\mathbb{Z}_2$ fluxes as well as the extremely short-ranged and bond-directional spin correlations. With this insight, we revisit the susceptibility measurements of Na$_2$IrO$_3$ and $α$-RuCl$_3$, and find evident signatures of finite-temperature fractionalization and ferromagnetic Kitaev couplings. Moreover, the peculiar spin correlation in the fractional liquid corresponds to a stripy structure factor which rotates in the extended Brillouin zone as the spin component changes. Therefore, our findings encourage future experimental exploration of fractional liquid in the Kitaev materials by thermodynamic measurements and spin-resolved structure factor probes.

cond-mat.str-el↗

arXiv2020

Learning to Accelerate Heuristic Searching for Large-Scale Maximum Weighted b-Matching Problems in Online Advertising

Bipartite b-matching is fundamental in algorithm design, and has been widely applied into economic markets, labor markets, etc. These practical problems usually exhibit two distinct features: large-scale and dynamic, which requires the matching algorithm to be repeatedly executed at regular intervals. However, existing exact and approximate algorithms usually fail in such settings due to either requiring intolerable running time or too much computation resource. To address this issue, we propose \texttt{NeuSearcher} which leverages the knowledge learned from previously instances to solve new problem instances. Specifically, we design a multichannel graph neural network to predict the threshold of the matched edges weights, by which the search region could be significantly reduced. We further propose a parallel heuristic search algorithm to iteratively improve the solution quality until convergence. Experiments on both open and industrial datasets demonstrate that \texttt{NeuSearcher} can speed up 2 to 3 times while achieving exactly the same matching solution compared with the state-of-the-art approximation approaches.

cs.DC↗

arXiv2020

Kosterlitz-Thouless Melting of Magnetic Order in the Triangular Quantum Ising Material TmMgGaO$_4$

Frustrated magnets host the promises of material realizations of new paradigm of quantum matter, while direct comparison of unbiased model calculations with experimental measurements is still very challenging. Here, we design and implement a protocol of employing many-body computation methodologies for accurate model calculation -- both equilibrium and dynamical properties -- of a frustrated rare-earth magnet TmMgGaO$_4$ (TMGO), which perfectly explains the corresponding experimental findings. Our results confirm TMGO is an ideal realization of triangular-lattice Ising model with an intrinsic transverse field. The magnetic order of TMGO is predicted to melt through two successive Kosterlitz-Thouless (KT) phase transitions, with a floating KT phase in between. The dynamical spectra calculated suggest remnant images of a vanishing magnetic stripe order that represent vortex-antivortex pairs, resembling rotons in a superfluid helium film. TMGO therefore constitutes a rare quantum magnet for realizing KT physics and we further propose experimental detections of its intriguing properties.

cond-mat.str-el↗

arXiv2019

A Statistical Explanation of the Timing Attack on QC-MDPC Code Crypto-system

The McEliece cryptosystem based on quasi-cyclic moderate-density parity-check (QC-MDPC) codes is first purposed in 2013\cite{QCMDPC} and is considered a promising contender in the post-quantum era. Understanding its security is hence essential. Till now, the most effective attacks are the reaction attack\cite{Reaction} and the timing attack\cite{Timing}. Both of these attacks rely on the decoding performance to recover the private key. The reaction attack relies on the decoding failure rate and the timing attack relies on the iterations during decoding. However, the mechanics behind these attacks remain elusive. In this paper, a mathematical model is proposed to explain both attacks by connecting the spectrum of private key and first-layer performance of the decoder.

cs.CR↗

arXiv2019

Joint Optimization of Tree-based Index and Deep Model for Recommender Systems

Large-scale industrial recommender systems are usually confronted with computational problems due to the enormous corpus size. To retrieve and recommend the most relevant items to users under response time limits, resorting to an efficient index structure is an effective and practical solution. The previous work Tree-based Deep Model (TDM) \cite{zhu2018learning} greatly improves recommendation accuracy using tree index. By indexing items in a tree hierarchy and training a user-node preference prediction model satisfying a max-heap like property in the tree, TDM provides logarithmic computational complexity w.r.t. the corpus size, enabling the use of arbitrary advanced models in candidate retrieval and recommendation. In tree-based recommendation methods, the quality of both the tree index and the user-node preference prediction model determines the recommendation accuracy for the most part. We argue that the learning of tree index and preference model has interdependence. Our purpose, in this paper, is to develop a method to jointly learn the index structure and user preference prediction model. In our proposed joint optimization framework, the learning of index and user preference prediction model are carried out under a unified performance measure. Besides, we come up with a novel hierarchical user preference representation utilizing the tree index hierarchy. Experimental evaluations with two large-scale real-world datasets show that the proposed method improves recommendation accuracy significantly. Online A/B test results at a display advertising platform also demonstrate the effectiveness of the proposed method in production environments.

cs.IR↗

arXiv2019

Encoding CT Anatomy Knowledge for Unpaired Chest X-ray Image Decomposition

Although chest X-ray (CXR) offers a 2D projection with overlapped anatomies, it is widely used for clinical diagnosis. There is clinical evidence supporting that decomposing an X-ray image into different components (e.g., bone, lung and soft tissue) improves diagnostic value. We hereby propose a decomposition generative adversarial network (DecGAN) to anatomically decompose a CXR image but with unpaired data. We leverage the anatomy knowledge embedded in CT, which features a 3D volume with clearly visible anatomies. Our key idea is to embed CT priori decomposition knowledge into the latent space of unpaired CXR autoencoder. Specifically, we train DecGAN with a decomposition loss, adversarial losses, cycle-consistency losses and a mask loss to guarantee that the decomposed results of the latent space preserve realistic body structures. Extensive experiments demonstrate that DecGAN provides superior unsupervised CXR bone suppression results and the feasibility of modulating CXR components by latent space disentanglement. Furthermore, we illustrate the diagnostic value of DecGAN and demonstrate that it outperforms the state-of-the-art approaches in terms of predicting 11 out of 14 common lung diseases.

eess.IV↗

arXiv2019

Low-temperature electroluminescence excitation mapping of excitons and trions in short channel monochiral carbon nanotube device

Single walled carbon nanotubes as emerging quantum-light sources may fill a technological gap in silicon photonics due to their potential use as near infrared, electrically driven, classical or non-classical emitters. Unlike in photoluminescence, where nanotubes are excited with light, electrical excitation of single tubes is challenging and heavily influenced by device fabrication, architecture and biasing conditions. Here we present electroluminescence spectroscopy data of ultra short channel devices made from (9,8) carbon nanotubes emitting in the telecom band. Emissions are stable under current biasing and no quenching is observed down to 10 nm gap size. Low-temperature electroluminescence spectroscopy data also reported exhibits cold emission and linewidths down to 2 meV at 4 K. Electroluminescence excitation maps give evidence that carrier recombination is the mechanism for light generation in short channels. Excitonic and trionic emissions can be switched on and off by gate voltage and corresponding emission efficiency maps were compiled. Insights are gained into the influence of acoustic phonons on the linewidth, absence of intensity saturation and exciton exciton annihilation, environmental effects like dielectric screening and strain on the emission wavelength, and conditions to suppress hysteresis and establish optimum operation conditions.

physics.app-ph↗

arXiv2019

Learning Adaptive Display Exposure for Real-Time Advertising

In E-commerce advertising, where product recommendations and product ads are presented to users simultaneously, the traditional setting is to display ads at fixed positions. However, under such a setting, the advertising system loses the flexibility to control the number and positions of ads, resulting in sub-optimal platform revenue and user experience. Consequently, major e-commerce platforms (e.g., Taobao.com) have begun to consider more flexible ways to display ads. In this paper, we investigate the problem of advertising with adaptive exposure: can we dynamically determine the number and positions of ads for each user visit under certain business constraints so that the platform revenue can be increased? More specifically, we consider two types of constraints: request-level constraint ensures user experience for each user visit, and platform-level constraint controls the overall platform monetization rate. We model this problem as a Constrained Markov Decision Process with per-state constraint (psCMDP) and propose a constrained two-level reinforcement learning approach to decompose the original problem into two relatively independent sub-problems. To accelerate policy learning, we also devise a constrained hindsight experience replay mechanism. Experimental evaluations on industry-scale real-world datasets demonstrate the merits of our approach in both obtaining higher revenue under the constraints and the effectiveness of the constrained hindsight experience replay mechanism.

cs.LG↗

arXiv2019

Learning to Advertise for Organic Traffic Maximization in E-Commerce Product Feeds

Most e-commerce product feeds provide blended results of advertised products and recommended products to consumers. The underlying advertising and recommendation platforms share similar if not exactly the same set of candidate products. Consumers' behaviors on the advertised results constitute part of the recommendation model's training data and therefore can influence the recommended results. We refer to this process as Leverage. Considering this mechanism, we propose a novel perspective that advertisers can strategically bid through the advertising platform to optimize their recommended organic traffic. By analyzing the real-world data, we first explain the principles of Leverage mechanism, i.e., the dynamic models of Leverage. Then we introduce a novel Leverage optimization problem and formulate it with a Markov Decision Process. To deal with the sample complexity challenge in model-free reinforcement learning, we propose a novel Hybrid Training Leverage Bidding (HTLB) algorithm which combines the real-world samples and the emulator-generated samples to boost the learning speed and stability. Our offline experiments as well as the results from the online deployment demonstrate the superior performance of our approach.

stat.ML↗

arXiv2019

Spectral-based Graph Convolutional Network for Directed Graphs

Graph convolutional networks(GCNs) have become the most popular approaches for graph data in these days because of their powerful ability to extract features from graph. GCNs approaches are divided into two categories, spectral-based and spatial-based. As the earliest convolutional networks for graph data, spectral-based GCNs have achieved impressive results in many graph related analytics tasks. However, spectral-based models cannot directly work on directed graphs. In this paper, we propose an improved spectral-based GCN for the directed graph by leveraging redefined Laplacians to improve its propagation model. Our approach can work directly on directed graph data in semi-supervised nodes classification tasks. Experiments on a number of directed graph datasets demonstrate that our approach outperforms the state-of-the-art methods.

cs.LG↗

arXiv2019

Nonparametric Statistical Inference and Imputation for Incomplete Categorical Data

Missingness in categorical data is a common problem in various real applications. Traditional approaches either utilize only the complete observations or impute the missing data by some ad hoc methods rather than the true conditional distribution of the missing data, thus losing or distorting the rich information in the partial observations. In this paper, we propose the Dirichlet Process Mixture of Collapsed Product-Multinomials (DPMCPM) to model the full data jointly and compute the model efficiently. By fitting an infinite mixture of product-multinomial distributions, DPMCPM is applicable for any categorical data regardless of the true distribution, which may contain complex association among variables. Under the framework of latent class analysis, we show that DPMCPM can model general missing mechanisms by creating an extra category to denote missingness, which implicitly integrates out the missing part with regard to their true conditional distribution. Through simulation studies and a real application, we demonstrate that DPMCPM outperforms existing approaches on statistical inference and imputation for incomplete categorical data of various missing mechanisms. DPMCPM is implemented as the R package \texttt{MMDai}, which is available from the Comprehensive R Archive Network at https://cran.r-project.org/web/packages/MMDai/index.html.

stat.ME↗

arXiv2019

Continuous Learning for Large-scale Personalized Domain Classification

Domain classification is the task of mapping spoken language utterances to one of the natural language understanding domains in intelligent personal digital assistants (IPDAs). This is a major component in mainstream IPDAs in industry. Apart from official domains, thousands of third-party domains are also created by external developers to enhance the capability of IPDAs. As more domains are developed rapidly, the question of how to continuously accommodate the new domains still remains challenging. Moreover, existing continual learning approaches do not address the problem of incorporating personalized information dynamically for better domain classification. In this paper, we propose CoNDA, a neural network based approach for domain classification that supports incremental learning of new classes. Empirical evaluation shows that CoNDA achieves high accuracy and outperforms baselines by a large margin on both incrementally added new domains and existing domains.

cs.LG↗

arXiv2019

On effective equidistribution for quotients of SL(d)

We prove the first case of polynomially effective equidistribution of closed orbits of semisimple groups with nontrivial centralizer. The proof relies on uniform spectral gap, builds on, and extends work of Einsiedler, Margulis, and Venkatesh.

math.DS↗