arXiv subjects

Xiang Li

Publications and source records attributed to Xiang Li.

At least 1,279 records · Page 71Linked to original sources

arXiv2020

Adversarial Multimodal Representation Learning for Click-Through Rate Prediction

For better user experience and business effectiveness, Click-Through Rate (CTR) prediction has been one of the most important tasks in E-commerce. Although extensive CTR prediction models have been proposed, learning good representation of items from multimodal features is still less investigated, considering an item in E-commerce usually contains multiple heterogeneous modalities. Previous works either concatenate the multiple modality features, that is equivalent to giving a fixed importance weight to each modality; or learn dynamic weights of different modalities for different items through technique like attention mechanism. However, a problem is that there usually exists common redundant information across multiple modalities. The dynamic weights of different modalities computed by using the redundant information may not correctly reflect the different importance of each modality. To address this, we explore the complementarity and redundancy of modalities by considering modality-specific and modality-invariant features differently. We propose a novel Multimodal Adversarial Representation Network (MARN) for the CTR prediction task. A multimodal attention network first calculates the weights of multiple modalities for each item according to its modality-specific features. Then a multimodal adversarial network learns modality-invariant representations where a double-discriminators strategy is introduced. Finally, we achieve the multimodal item representations by combining both modality-specific and modality-invariant representations. We conduct extensive experiments on both public and industrial datasets, and the proposed method consistently achieves remarkable improvements to the state-of-the-art methods. Moreover, the approach has been deployed in an operational E-commerce system and online A/B testing further demonstrates the effectiveness.

cs.IR↗

arXiv2020

Local polynomial regression for pooled response data

We propose local polynomial estimators for the conditional mean of a continuous response when only pooled response data are collected under different pooling designs. Asymptotic properties of these estimators are investigated and compared. Extensive simulation studies are carried out to compare finite sample performance of the proposed estimators under various model settings and pooling strategies. We apply the proposed local polynomial regression methods to two real-life applications to illustrate practical implementation and performance of the estimators for the mean function.

stat.ME↗

arXiv2020

Boresight Alignment of DArk Matter Particle Explorer

The DArk Matter Particle Explorer (DAMPE) can measure $γ$-rays in the energy range from a few GeV to about 10 TeV. The direction of each $γ$-ray is reconstructed with respect to the reference system of the DAMPE payload. In this paper, we adopt a maximum likelihood method and use the $γ$-ray data centered around several bright point-like sources to measure and correct the angular deviation from the real celestial coordinate system, the so called ``boresight alignment'' of the DAMPE payload. As a check, we also estimate the boresight alignment for some sets of simulation data with artificial orientation and obtain consistent results. The time-dependent boresight alignment analysis does not show evidence for significant variation of the parameters.

astro-ph.IM↗

arXiv2019

Materials Relevant to Realizing a Field-Effect Transistor based on Spin-Orbit Torques

Spin-orbit torque is a promising mechanism for writing magnetic memories, while field-effect transistors are the gold-standard device for logic operation. The spin-orbit torque field effect transistor (SOTFET) is a proposed device that couples a spin-orbit-torque-controlled ferromagnet to a semiconducting transistor channel via the transduction in a magnetoelectric multiferroic. This allows the SOTFET to operate as both a memory and a logic device, but its realization depends on the choice of appropriate materials. In this report, we discuss and parametrize the types of materials that can lead to a SOTFET heterostructure.

cond-mat.mtrl-sci↗

arXiv2019

Safe Sample Screening for Robust Support Vector Machine

Robust support vector machine (RSVM) has been shown to perform remarkably well to improve the generalization performance of support vector machine under the noisy environment. Unfortunately, in order to handle the non-convexity induced by ramp loss in RSVM, existing RSVM solvers often adopt the DC programming framework which is computationally inefficient for running multiple outer loops. This hinders the application of RSVM to large-scale problems. Safe sample screening that allows for the exclusion of training samples prior to or early in the training process is an effective method to greatly reduce computational time. However, existing safe sample screening algorithms are limited to convex optimization problems while RSVM is a non-convex problem. To address this challenge, in this paper, we propose two safe sample screening rules for RSVM based on the framework of concave-convex procedure (CCCP). Specifically, we provide screening rule for the inner solver of CCCP and another rule for propagating screened samples between two successive solvers of CCCP. To the best of our knowledge, this is the first work of safe sample screening to a non-convex optimization problem. More importantly, we provide the security guarantee to our sample screening rules to RSVM. Experimental results on a variety of benchmark datasets verify that our safe sample screening rules can significantly reduce the computational time.

cs.LG↗

arXiv2019

Multi-label Detection and Classification of Red Blood Cells in Microscopic Images

Cell detection and cell type classification from biomedical images play an important role for high-throughput imaging and various clinical application. While classification of single cell sample can be performed with standard computer vision and machine learning methods, analysis of multi-label samples (region containing congregating cells) is more challenging, as separation of individual cells can be difficult (e.g. touching cells) or even impossible (e.g. overlapping cells). As multi-instance images are common in analyzing Red Blood Cell (RBC) for Sickle Cell Disease (SCD) diagnosis, we develop and implement a multi-instance cell detection and classification framework to address this challenge. The framework firstly trains a region proposal model based on Region-based Convolutional Network (RCNN) to obtain bounding-boxes of regions potentially containing single or multiple cells from input microscopic images, which are extracted as image patches. High-level image features are then calculated from image patches through a pre-trained Convolutional Neural Network (CNN) with ResNet-50 structure. Using these image features inputs, six networks are then trained to make multi-label prediction of whether a given patch contains cells belonging to a specific cell type. As the six networks are trained with image patches consisting of both individual cells and touching/overlapping cells, they can effectively recognize cell types that are presented in multi-instance image samples. Finally, for the purpose of SCD testing, we train another machine learning classifier to predict whether the given image patch contains abnormal cell type based on outputs from the six networks. Testing result of the proposed framework shows that it can achieve good performance in automatic cell detection and classification.

eess.IV↗

arXiv2019

Intermolecular forces and correlations mediated by a phonon bath

Inspired by the possibility to experimentally manipulate and enhance chemical reactivity in helium nanodroplets, we investigate the effective interaction and the resulting correlations between two diatomic molecules immersed in a bath of bosons. By analogy with the bipolaron, we introduce the \emph{biangulon} quasiparticle describing two rotating molecules that align with respect to each other due to the effective attractive interaction mediated by the excitations of the bath. We study this system in different parameter regimes and apply several theoretical approaches to describe its properties. Using a Born-Oppenheimer approximation, we investigate the dependence of the effective intermolecular interaction on the rotational state of the two molecules. In the strong-coupling regime, a product-state ansatz shows that the molecules tend to have a strong alignment in the ground state. To investigate the system in the weak-coupling regime, we apply a one-phonon excitation variational ansatz, which allows us to access the energy spectrum. In comparison to the angulon quasiparticle, the biangulon shows shifted angulon instabilities and an additional spectral instability, where resonant angular momentum transfer between the molecules and the bath takes place. These features are proposed as an experimentally observable signature for the formation of the biangulon quasiparticle. Finally, by using products of single angulon and bare impurity wave functions as basis states, we introduce a diagonalization scheme that allows us to describe the transition from two separated angulons to a biangulon as a function of the distance between the two molecules.

cond-mat.mes-hall↗

arXiv2019

Image-free real-time classification of fast moving objects using 'learned' spatial light modulation and a single-pixel detector

Objects classification generally relies on image acquisition and analysis. Real-time classification of high-speed moving objects is challenging, as both high temporal resolution in image acquisition and low computational complexity in objects classification algorithms are required. Here we propose and experimentally demonstrate an approach for real-time moving objects classification without image acquisition. As objects classification algorithms rely on the feature information of objects, we propose to use spatial light modulation to acquire the feature information directly rather than performing image acquisition followed by features extraction. A convolutional neural network is designed and trained to learn the spatial features of the target objects. The trained network can generate structured patterns for spatial light modulation. Using the resulting structured patterns for spatial light modulation, the feature information of target objects can be compressively encoded into a short light intensity sequence. The resulting one-dimensional signal is collected by a single-pixel detector and fed to the convolutional neural network for objects classification. As experimentally demonstrated, the proposed approach can achieve accurate and real-time classification of fast moving objects. The proposed method has potential applications in the fields where fast moving objects classification in real time and for long duration is required.

eess.IV↗

arXiv2019

Implications on the origin of cosmic rays in light of 10 TV spectral softenings

Precise measurements of the energy spectra of cosmic rays (CRs) show various kinds of features deviating from single power-laws, which give very interesting and important implications on their origin and propagation. Previous measurements from a few balloon and space experiments indicate the existence of spectral softenings around 10 TV for protons (and probably also for Helium nuclei). Very recently, the DArk Matter Particle Explorer (DAMPE) measurement about the proton spectrum clearly reveals such a softening with a high significance. Here we study the implications of these new measurements, as well as the groundbased indirect measurements, on the origin of CRs. We find that a single component of CRs fails to fit the spectral softening and the air shower experiment data simultaneously. In the framework of multiple components, we discuss two possible scenarios, the multiple source population scenario and the background plus nearby source scenario. Both scenarios give reasonable fits to the wide-band data from TeV to 100 PeV energies. Considering the anisotropy observations, the nearby source model is favored.

astro-ph.HE↗

arXiv2019

Linear quantum measurement theory of matter-wave interferometry

The theory of linear quantum measurement has been developed for analysing the sensitivities of experimental devices that measure extremely weak signals, such as gravitational waves. It has successfully contributed to the theoretical understanding of laser interferometer gravitational-wave detectors (used by LIGO, VIRGO and KaGRA) and helped many important experimental upgrades. In this work, we establish a linear quantum measurement theory for another kind of measurement device--- matter wave interferometers, which has been widely discussed as an important platform for many high-precision experiments. This theory allows us to account for both atom and light fluctuations, and leads to a detailed analysis of back-action in matter-wave interferometry (action of light back onto the atoms) and its effect on dynamics and measurement noise. From this analysis, we obtain a Standard Quantum Limit (SQL) for matter-wave interferometry. A comparison between the LIGO detector and matter wave interferometer is also given from the perspective of quantum measurement.

quant-ph↗

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

Understanding the Teaching Styles by an Attention based Multi-task Cross-media Dimensional modelling

Teaching style plays an influential role in helping students to achieve academic success. In this paper, we explore a new problem of effectively understanding teachers' teaching styles. Specifically, we study 1) how to quantitatively characterize various teachers' teaching styles for various teachers and 2) how to model the subtle relationship between cross-media teaching related data (speech, facial expressions and body motions, content et al.) and teaching styles. Using the adjectives selected from more than 10,000 feedback questionnaires provided by an educational enterprise, a novel concept called Teaching Style Semantic Space (TSSS) is developed based on the pleasure-arousal dimensional theory to describe teaching styles quantitatively and comprehensively. Then a multi-task deep learning based model, Attention-based Multi-path Multi-task Deep Neural Network (AMMDNN), is proposed to accurately and robustly capture the internal correlations between cross-media features and TSSS. Based on the benchmark dataset, we further develop a comprehensive data set including 4,541 full-annotated cross-modality teaching classes. Our experimental results demonstrate that the proposed AMMDNN outperforms (+0.0842 in terms of the concordance correlation coefficient (CCC) on average) baseline methods. To further demonstrate the advantages of the proposed TSSS and our model, several interesting case studies are carried out, such as teaching styles comparison among different teachers and courses, and leveraging the proposed method for teaching quality analysis.

cs.MM↗

arXiv2019

Generative adversarial networks (GAN) based efficient sampling of chemical space for inverse design of inorganic materials

A major challenge in materials design is how to efficiently search the vast chemical design space to find the materials with desired properties. One effective strategy is to develop sampling algorithms that can exploit both explicit chemical knowledge and implicit composition rules embodied in the large materials database. Here, we propose a generative machine learning model (MatGAN) based on a generative adversarial network (GAN) for efficient generation of new hypothetical inorganic materials. Trained with materials from the ICSD database, our GAN model can generate hypothetical materials not existing in the training dataset, reaching a novelty of 92.53% when generating 2 million samples. The percentage of chemically valid (charge neutral and electronegativity balanced) samples out of all generated ones reaches 84.5% by our GAN when trained with materials from ICSD even though no such chemical rules are explicitly enforced in our GAN model, indicating its capability to learn implicit chemical composition rules. Our algorithm could be used to speed up inverse design or computational screening of inorganic materials.

cs.LG↗

arXiv2019

An End-to-End Deep RL Framework for Task Arrangement in Crowdsourcing Platforms

In this paper, we propose a Deep Reinforcement Learning (RL) framework for task arrangement, which is a critical problem for the success of crowdsourcing platforms. Previous works conduct the personalized recommendation of tasks to workers via supervised learning methods. However, the majority of them only consider the benefit of either workers or requesters independently. In addition, they cannot handle the dynamic environment and may produce sub-optimal results. To address these issues, we utilize Deep Q-Network (DQN), an RL-based method combined with a neural network to estimate the expected long-term return of recommending a task. DQN inherently considers the immediate and future reward simultaneously and can be updated in real-time to deal with evolving data and dynamic changes. Furthermore, we design two DQNs that capture the benefit of both workers and requesters and maximize the profit of the platform. To learn value functions in DQN effectively, we also propose novel state representations, carefully design the computation of Q values, and predict transition probabilities and future states. Experiments on synthetic and real datasets demonstrate the superior performance of our framework.

cs.LG↗

arXiv2019

A General Early-Stopping Module for Crowdsourced Ranking

Crowdsourcing can be used to determine a total order for an object set (e.g., the top-10 NBA players) based on crowd opinions. This ranking problem is often decomposed into a set of microtasks (e.g., pairwise comparisons). These microtasks are passed to a large number of workers and their answers are aggregated to infer the ranking. The number of microtasks depends on the budget allocated for the problem. Intuitively, the higher the number of microtask answers, the more accurate the ranking becomes. However, it is often hard to decide the budget required for an accurate ranking. We study how a ranking process can be terminated early, and yet achieve a high-quality ranking and great savings in the budget. We use statistical tools to estimate the quality of the ranking result at any stage of the crowdsourcing process and terminate the process as soon as the desired quality is achieved. Our proposed early-stopping module can be seamlessly integrated with most existing inference algorithms and task assignment methods. We conduct extensive experiments and show that our early-stopping module is better than other existing general stopping criteria. We also implement a prototype system to demonstrate the usability and effectiveness of our approach in practice.

cs.DB↗

arXiv2019

Gene-Patterns: Should Architecture be Customized for Each Application?

Providing architectural support is crucial for newly arising applications to achieve high performance and high system efficiency. Currently there is a trend in designing accelerators for special applications, while arguably a debate is sparked whether we should customize architecture for each application. In this study, we introduce what we refer to as Gene-Patterns, which are the base patterns of diverse applications. We present a Recursive Reduce methodology to identify the hotspots, and a HOtspot Trace Suite (HOTS) is provided for the research community. We first extract the hotspot patterns, and then, remove the redundancy to obtain the base patterns. We find that although the number of applications is huge and ever-increasing, the amount of base patterns is relatively small, due to the similarity among the patterns of diverse applications. The similarity stems not only from the algorithms but also from the data structures. We build the Periodic Table of Memory Access Patterns (PT-MAP), where the indifference curves are analogous to the energy levels in physics, and memory performance optimization is essentially an energy level transition. We find that inefficiency results from the mismatch between some of the base patterns and the micro-architecture of modern processors. We have identified the key micro-architecture demands of the base patterns. The Gene-Pattern concept, methodology, and toolkit will facilitate the design of both hardware and software for the matching between architectures and applications.

cs.DC↗

arXiv2019

Prediction stability as a criterion in active learning

Recent breakthroughs made by deep learning rely heavily on large number of annotated samples. To overcome this shortcoming, active learning is a possible solution. Beside the previous active learning algorithms that only adopted information after training, we propose a new class of method based on the information during training, named sequential-based method. An specific criterion of active learning called prediction stability is proposed to prove the feasibility of sequential-based methods. Experiments are made on CIFAR-10 and CIFAR-100, and the results indicates that prediction stability is effective and works well on fewer-labeled datasets. Prediction stability reaches the accuracy of traditional acquisition functions like entropy on CIFAR-10, and notably outperforms them on CIFAR-100.

cs.CV↗

arXiv2019

New Insight about the Effective Restoration of U_A(1) Symmetry

The effective restoration of the U_{A}(1) symmetry is revisited by implementing the functional renormalization group approach combining with the 2+1 flavor Polyakov-loop quark-meson model. A temperature-dependent 't Hooft term is taken to imitate the restoration of the U_{A}(1) symmetry. Order parameters, meson spectrum and mixing angles, especially the pressure and the entropy density of the system are calculated to explore the effects of different U_{A}(1) symmetry restoration patterns. We show then that the temperature for the restoration of the U_{A}(1) symmetry is much higher than that for the chiral symmetry SU_{A}(3).

hep-ph↗