arXiv subjects

Liang Wang

Publications and source records attributed to Liang Wang.

At least 613 records · Page 34Linked to original sources

arXiv2018

Spatially-resolved galaxy angular momentum

The total specific angular momentum j of a galaxy disk is matched with that of its dark matter halo, but the distributions are different, in that there is a lack of both low- and high-j baryons with respect to the CDM predictions. I illustrate how the probability density function PDF(j/j_mean) can inform us of a galaxy's morphology and evolutionary history with a spanning set of examples from present-day galaxies and a galaxy at z~1.5. The shape of PDF(j/j_mean) is correlated with photometric morphology, with disk-dominated galaxies having more symmetric PDF(j/j_mean) and bulge-dominated galaxies having a strongly-skewed PDF(j/j_mean). Galaxies with bigger bulges have more strongly-tailed PDF(j/j_mean), but disks of all sizes have a similar PDF(j/j_mean). In future, PDF(j/j_mean) will be useful as a kinematic decomposition tool.

astro-ph.GA↗

arXiv2018

Practical Constrained Optimization of Auction Mechanisms in E-Commerce Sponsored Search Advertising

Sponsored search in E-commerce platforms such as Amazon, Taobao and Tmall provides sellers an effective way to reach potential buyers with most relevant purpose. In this paper, we study the auction mechanism optimization problem in sponsored search on Alibaba's mobile E-commerce platform. Besides generating revenue, we are supposed to maintain an efficient marketplace with plenty of quality users, guarantee a reasonable return on investment (ROI) for advertisers, and meanwhile, facilitate a pleasant shopping experience for the users. These requirements essentially pose a constrained optimization problem. Directly optimizing over auction parameters yields a discontinuous, non-convex problem that denies effective solutions. One of our major contribution is a practical convex optimization formulation of the original problem. We devise a novel re-parametrization of auction mechanism with discrete sets of representative instances. To construct the optimization problem, we build an auction simulation system which estimates the resulted business indicators of the selected parameters by replaying the auctions recorded from real online requests. We summarized the experiments on real search traffics to analyze the effects of fidelity of auction simulation, the efficacy under various constraint targets and the influence of regularization. The experiment results show that with proper entropy regularization, we are able to maximize revenue while constraining other business indicators within given ranges.

cs.GT↗

arXiv2018

Pose-Based Two-Stream Relational Networks for Action Recognition in Videos

Recently, pose-based action recognition has gained more and more attention due to the better performance compared with traditional appearance-based methods. However, there still exist two problems to be further solved. First, existing pose-based methods generally recognize human actions with captured 3D human poses which are very difficult to obtain in real scenarios. Second, few pose-based methods model the action-related objects in recognizing human-object interaction actions in which objects play an important role. To solve the problems above, we propose a pose-based two-stream relational network (PSRN) for action recognition. In PSRN, one stream models the temporal dynamics of the targeted 2D human pose sequences which are directly extracted from raw videos, and the other stream models the action-related objects from a randomly sampled video frame. Most importantly, instead of fusing two-streams in the class score layer as before, we propose a pose-object relational network to model the relationship between human poses and action-related objects. We evaluate the proposed PSRN on two challenging benchmarks, i.e., Sub-JHMDB and PennAction. Experimental results show that our PSRN obtains the state-the-of-art performance on Sub-JHMDB (80.2%) and PennAction (98.1%). Our work opens a new door to action recognition by combining 2D human pose extracted from raw video and image appearance.

cs.CV↗

arXiv2018

Yuanfudao at SemEval-2018 Task 11: Three-way Attention and Relational Knowledge for Commonsense Machine Comprehension

This paper describes our system for SemEval-2018 Task 11: Machine Comprehension using Commonsense Knowledge. We use Three-way Attentive Networks (TriAN) to model interactions between the passage, question and answers. To incorporate commonsense knowledge, we augment the input with relation embedding from the graph of general knowledge ConceptNet (Speer et al., 2017). As a result, our system achieves state-of-the-art performance with 83.95% accuracy on the official test data. Code is publicly available at https://github.com/intfloat/commonsense-rc

cs.CL↗

arXiv2018

Giant planets around FGK stars form probably through core accretion

We present a statistical study of the planet-metallicity (P-M) correlation, by comparing the 744 stars with candidate planets (SWPs) in the Kepler field which have been observed with LAMOST, and a sample of distance-independent, fake "twin" stars in the Kepler field with no planet reported (CKSNPs) yet. With the well-defined and carefully-selected large samples, we find for the first time a turn-off P-M correlation of Delta [Fe/H]_(SWPs-SNPs), which in average increases from ~0.00+-0.03 dex to 0.06+-0.03 dex, and to 0.12+-0.03 for stars with Earth, Neptune, Jupiter-sized planets successively, and then declines to ~-0.01+-0.03 dex for more massive planets or brown dwarfs. Moreover, the percentage of those systems with positive Delta[Fe/H] has the same turn-off pattern. We also find FG-type stars follow this general trend, but K-type stars are different. Moderate metal enhancement (~0.1-0.2 dex) for K-type stars with planets of radii between 2 to 4 Earth radius as compared to CKSNPs is observed, which indicates much higher metallicities are required for Super-Earths, Neptune-sized planets to form around K-type stars. We point out that the P-M correlation is actually metallicity-dependent, i.e., the correlation is positive at solar and super-solar metallicities, and negative at subsolar metallicities. No steady increase of Delta[Fe/H] against planet sizes is observed for rocky planets, excluding the pollution scenario as a major mechanism for the P-M correlation. All these clues suggest that giant planets probably form differently from rocky planets or more massive planets/brown dwarfs, and the core-accretion scenario is highly favoured, and high metallicity is a prerequisite for massive planets to form.

astro-ph.EP↗

arXiv2018

Introducing galactic structure finder: the multiple stellar kinematic structures of a simulated Milky Way mass galaxy

We present the first results of applying Gaussian Mixture Models in the stellar kinematic space of normalized angular momentum and binding energy on NIHAO high resolution galaxies to separate the stars into multiple components. We exemplify this method using a simulated Milky Way analogue, whose stellar component hosts: thin and thick discs, classical and pseudo bulges, and a stellar halo. The properties of these stellar structures are in good agreement with observational expectations in terms of sizes, shapes and rotational support. Interestingly, the two kinematic discs show surface mass density profiles more centrally concentrated than exponentials, while the bulges and the stellar halo are purely exponential. We trace back in time the Lagrangian mass of each component separately to study their formation history. Between z~3 and the end of halo virialization, z~1.3, all components lose a fraction of their angular momentum. The classical bulge loses the most (~95%) and the thin disc the least (~60%). Both bulges formed their stars in-situ at high redshift, while the thin disc formed ~98% in-situ, but with a constant SFR~1.5M$_{\rm\odot}$/yr$^{\rm-1}$ over the last ~11 Gyr. Accreted stars (6% of total stellar mass) are mainly incorporated to the thick disc or the stellar halo, which formed ex-situ 8% and 45% of their respective masses. Our analysis pipeline is freely available at https://github.com/aobr/gsf.

astro-ph.GA↗

arXiv2018

Revisiting the stellar mass -- angular momentum -- morphology relation: extension to higher bulge fraction, and the effect of bulge type

We present the relation between stellar specific angular momentum $j_*$, stellar mass $M_*$, and bulge-to-total light ratio $β$ for THINGS, CALIFA and Romanowsky \& Fall datasets, exploring the existence of a fundamental plane between these parameters as first suggested by Obreschkow \& Glazebrook. Our best-fit $M_*-j_*$ relation yields a slope of $α= 1.03 \pm 0.11$ with a trivariate fit including $β$. When ignoring the effect of $β$, the exponent $α= 0.56 \pm 0.06$ is consistent with $α= 2/3$ predicted for dark matter halos. There is a linear $β- j_*/M_*$ relation for $β\lesssim 0.4$, exhibiting a general trend of increasing $β$ with decreasing $j_*/M_*$. Galaxies with $β\gtrsim 0.4$ have higher $j_*$ than predicted by the relation. Pseudobulge galaxies have preferentially lower $β$ for a given $j_*/M_*$ than galaxies that contain classical bulges. Pseudobulge galaxies follow a well-defined track in $β- j_*/M_*$ space, consistent with Obreschkow \& Glazebrook, while galaxies with classical bulges do not. These results are consistent with the hypothesis that while growth in either bulge type is linked to a decrease in $j_*/M_*$, the mechanisms that build pseudobulges seem to be less efficient at increasing bulge mass per decrease in specific angular momentum than those that build classical bulges.

astro-ph.GA↗

arXiv2018

Data Analytics Service Composition and Deployment on Edge Devices

Data analytics on edge devices has gained rapid growth in research, industry, and different aspects of our daily life. This topic still faces many challenges such as limited computation resource on edge devices. In this paper, we further identify two main challenges: the composition and deployment of data analytics services on edge devices. We present the Zoo system to address these two challenge: on one hand, it provides simple and concise domain-specific language to enable easy and and type-safe composition of different data analytics services; on the other, it utilises multiple deployment backends, including Docker container, JavaScript, and MirageOS, to accommodate the heterogeneous edge deployment environment. We show the expressiveness of Zoo with a use case, and thoroughly compare the performance of different deployment backends in evaluation.

cs.SE↗

arXiv2018

Privacy-Preserving Personal Model Training

Many current Internet services rely on inferences from models trained on user data. Commonly, both the training and inference tasks are carried out using cloud resources fed by personal data collected at scale from users. Holding and using such large collections of personal data in the cloud creates privacy risks to the data subjects, but is currently required for users to benefit from such services. We explore how to provide for model training and inference in a system where computation is pushed to the data in preference to moving data to the cloud, obviating many current privacy risks. Specifically, we take an initial model learnt from a small set of users and retrain it locally using data from a single user. We evaluate on two tasks: one supervised learning task, using a neural network to recognise users' current activity from accelerometer traces; and one unsupervised learning task, identifying topics in a large set of documents. In both cases the accuracy is improved. We also analyse the robustness of our approach against adversarial attacks, as well as its feasibility by presenting a performance evaluation on a representative resource-constrained device (a Raspberry Pi).

cs.LG↗

arXiv2018

Cluster synchronization in complex network of coupled chaotic circuits: an experimental study

By a small-size complex network of coupled chaotic Hindmarsh-Rose circuits, we study experimentally the stability of network synchronization to the removal of shortcut links. It is shown that the removal of a single shortcut link may destroy either completely or partially the network synchronization. Interestingly, when the network is partially desynchronized, it is found that the oscillators can be organized into different groups, with oscillators within each group being highly synchronized but are not for oscillators from different groups, showing the intriguing phenomenon of cluster synchronization. The experimental results are analyzed by the method of eigenvalue analysis, which implies that the formation of cluster synchronization is crucially dependent on the network symmetries. Our study demonstrates the observability of cluster synchronization in realistic systems, and indicates the feasibility of controlling network synchronization by adjusting network topology.

nlin.CD↗

arXiv2018

Optimizing Sponsored Search Ranking Strategy by Deep Reinforcement Learning

Sponsored search is an indispensable business model and a major revenue contributor of almost all the search engines. From the advertisers' side, participating in ranking the search results by paying for the sponsored search advertisement to attract more awareness and purchase facilitates their commercial goal. From the users' side, presenting personalized advertisement reflecting their propensity would make their online search experience more satisfactory. Sponsored search platforms rank the advertisements by a ranking function to determine the list of advertisements to show and the charging price for the advertisers. Hence, it is crucial to find a good ranking function which can simultaneously satisfy the platform, the users and the advertisers. Moreover, advertisements showing positions under different queries from different users may associate with advertisement candidates of different bid price distributions and click probability distributions, which requires the ranking functions to be optimized adaptively to the traffic characteristics. In this work, we proposed a generic framework to optimize the ranking functions by deep reinforcement learning methods. The framework is composed of two parts: an offline learning part which initializes the ranking functions by learning from a simulated advertising environment, allowing adequate exploration of the ranking function parameter space without hurting the performance of the commercial platform. An online learning part which further optimizes the ranking functions by adapting to the online data distribution. Experimental results on a large-scale sponsored search platform confirm the effectiveness of the proposed method.

cs.IR↗

arXiv2018

Electron Physics in 3D Two-Fluid Ten-Moment Modeling of Ganymede's Magnetosphere

We studied the role of electron physics in 3D two-fluid 10-moment simulation of the Ganymede's magnetosphere. The model captures non-ideal physics like the Hall effect, the electron inertia, and anisotropic, non-gyrotropic pressure effects. A series of analyses were carried out: 1) The resulting magnetic field topology and electron and ion convection patterns were investigated. The magnetic fields were shown to agree reasonably well with in-situ measurements by the Galileo satellite. 2) The physics of collisionless magnetic reconnection were carefully examined in terms of the current sheet formation and decomposition of generalized Ohm's law. The importance of pressure anisotropy and non-gyrotropy in supporting the reconnection electric field is confirmed. 3) We compared surface "brightness" morphology, represented by surface electron and ion pressure contours, with oxygen emission observed by the Hubble Space Telescope (HST). The correlation between the observed emission morphology and spatial variability in electron/ion pressure was demonstrated. Potential extension to multi-ion species in the context of Ganymede and other magnetospheric systems is also discussed.

physics.space-ph↗

arXiv2018

A Hierarchical Contextual Attention-based GRU Network for Sequential Recommendation

Sequential recommendation is one of fundamental tasks for Web applications. Previous methods are mostly based on Markov chains with a strong Markov assumption. Recently, recurrent neural networks (RNNs) are getting more and more popular and has demonstrated its effectiveness in many tasks. The last hidden state is usually applied as the sequence's representation to make recommendation. Benefit from the natural characteristics of RNN, the hidden state is a combination of long-term dependency and short-term interest to some degrees. However, the monotonic temporal dependency of RNN impairs the user's short-term interest. Consequently, the hidden state is not sufficient to reflect the user's final interest. In this work, to deal with this problem, we propose a Hierarchical Contextual Attention-based GRU (HCA-GRU) network. The first level of HCA-GRU is conducted on the input. We construct a contextual input by using several recent inputs based on the attention mechanism. This can model the complicated correlations among recent items and strengthen the hidden state. The second level is executed on the hidden state. We fuse the current hidden state and a contextual hidden state built by the attention mechanism, which leads to a more suitable user's overall interest. Experiments on two real-world datasets show that HCA-GRU can effectively generate the personalized ranking list and achieve significant improvement.

cs.IR↗

arXiv2017

Learning Semantic Concepts and Order for Image and Sentence Matching

Image and sentence matching has made great progress recently, but it remains challenging due to the large visual-semantic discrepancy. This mainly arises from that the representation of pixel-level image usually lacks of high-level semantic information as in its matched sentence. In this work, we propose a semantic-enhanced image and sentence matching model, which can improve the image representation by learning semantic concepts and then organizing them in a correct semantic order. Given an image, we first use a multi-regional multi-label CNN to predict its semantic concepts, including objects, properties, actions, etc. Then, considering that different orders of semantic concepts lead to diverse semantic meanings, we use a context-gated sentence generation scheme for semantic order learning. It simultaneously uses the image global context containing concept relations as reference and the groundtruth semantic order in the matched sentence as supervision. After obtaining the improved image representation, we learn the sentence representation with a conventional LSTM, and then jointly perform image and sentence matching and sentence generation for model learning. Extensive experiments demonstrate the effectiveness of our learned semantic concepts and order, by achieving the state-of-the-art results on two public benchmark datasets.

cs.CV↗

arXiv2017

User-centric Composable Services: A New Generation of Personal Data Analytics

Machine Learning (ML) techniques, such as Neural Network, are widely used in today's applications. However, there is still a big gap between the current ML systems and users' requirements. ML systems focus on improving the performance of models in training, while individual users cares more about response time and expressiveness of the tool. Many existing research and product begin to move computation towards edge devices. Based on the numerical computing system Owl, we propose to build the Zoo system to support construction, compose, and deployment of ML models on edge and local devices.

cs.LG↗

arXiv2017

Probabilistic Synchronous Parallel

Most machine learning and deep neural network algorithms rely on certain iterative algorithms to optimise their utility/cost functions, e.g. Stochastic Gradient Descent. In distributed learning, the networked nodes have to work collaboratively to update the model parameters, and the way how they proceed is referred to as synchronous parallel design (or barrier control). Synchronous parallel protocol is the building block of any distributed learning framework, and its design has direct impact on the performance and scalability of the system. In this paper, we propose a new barrier control technique - Probabilistic Synchronous Parallel (PSP). Com- paring to the previous Bulk Synchronous Parallel (BSP), Stale Synchronous Parallel (SSP), and (Asynchronous Parallel) ASP, the proposed solution e ectively improves both the convergence speed and the scalability of the SGD algorithm by introducing a sampling primitive into the system. Moreover, we also show that the sampling primitive can be applied atop of the existing barrier control mechanisms to derive fully distributed PSP-based synchronous parallel. We not only provide a thorough theoretical analysis1 on the convergence of PSP-based SGD algorithm, but also implement a full-featured distributed learning framework called Actor and perform intensive evaluation atop of it.

cs.DC↗

arXiv2017

Towards Better Understanding of Bitcoin Unreachable Peers

The bitcoin peer-to-peer network has drawn significant attention from researchers, but so far has mostly focused on publicly visible portions of the network, i.e., publicly reachable peers. This mostly ignores the hidden parts of the network: unreachable Bitcoin peers behind NATs and firewalls. In this paper, we characterize Bitcoin peers that might be behind NATs or firewalls from different perspectives. Using a special-purpose measurement tool we conduct a large scale measurement study of the Bitcoin network, and discover several previously unreported usage patterns: a small number of peers are involved in the propagation of 89% of all bitcoin transactions, public cloud services are being used for Bitcoin network probing and crawling, a large amount of transactions are generated from only two mobile applications. We also empirically evaluate a method that uses timing information to re-identify the peer that created a transaction against unreachable peers. We find this method very accurate for peers that use the latest version of the Bitcoin Core client.

cs.NI↗

arXiv2017

Optimal Reserve Price for Online Ads Trading Based on Inventory Identification

The online ads trading platform plays a crucial role in connecting publishers and advertisers and generates tremendous value in facilitating the convenience of our lives. It has been evolving into a more and more complicated structure. In this paper, we consider the problem of maximizing the revenue for the seller side via utilizing proper reserve price for the auctions in a dynamical way. Predicting the optimal reserve price for each auction in the repeated auction marketplaces is a non-trivial problem. However, we were able to come up with an efficient method of improving the seller revenue by mainly focusing on adjusting the reserve price for those high-value inventories. Previously, no dedicated work has been performed from this perspective. Inspired by Paul and Michael, our model first identifies the value of the inventory by predicting the top bid price bucket using a cascade of classifiers. The cascade is essential in significantly reducing the false positive rate of a single classifier. Based on the output of the first step, we build another cluster of classifiers to predict the price separations between the top two bids. We showed that although the high-value auctions are only a small portion of all the traffic, successfully identifying them and setting correct reserve price would result in a significant revenue lift. Moreover, our optimization is compatible with all other reserve price models in the system and does not impact their performance. In other words, when combined with other models, the enhancement on exchange revenue will be aggregated. Simulations on randomly sampled Yahoo ads exchange (YAXR) data showed stable and expected lift after applying our model.

cs.GT↗