arXiv subjects

Jun Ma

Publications and source records attributed to Jun Ma.

At least 415 records · Page 23Linked to original sources

arXiv2018

RepeatNet: A Repeat Aware Neural Recommendation Machine for Session-based Recommendation

Recurrent neural networks for session-based recommendation have attracted a lot of attention recently because of their promising performance. repeat consumption is a common phenomenon in many recommendation scenarios (e.g., e-commerce, music, and TV program recommendations), where the same item is re-consumed repeatedly over time. However, no previous studies have emphasized repeat consumption with neural networks. An effective neural approach is needed to decide when to perform repeat recommendation. In this paper, we incorporate a repeat-explore mechanism into neural networks and propose a new model, called RepeatNet, with an encoder-decoder structure. RepeatNet integrates a regular neural recommendation approach in the decoder with a new repeat recommendation mechanism that can choose items from a user's history and recommends them at the right time. We report on extensive experiments on three benchmark datasets. RepeatNet outperforms state-of-the-art baselines on all three datasets in terms of MRR and Recall. Furthermore, as the dataset size and the repeat ratio increase, the improvements of RepeatNet over the baselines also increase, which demonstrates its advantage in handling repeat recommendation scenarios.

cs.IR↗

arXiv2018

The origin of high velocity stars from Gaia and LAMOST

Based on the second Gaia data (Gaia DR2) and spectroscopy from the LAMOST Data Release 5, we defined the high-velocity (HiVel) stars sample as those stars with $v_{\mathrm{gc}} > 0.85 v_{\mathrm{esc}}$, and derived the final sample of 24 HiVel stars with stellar astrometric parameters and radial velocities. Most of the HiVel stars are metal-poor and $α$-enhanced. In order to further explore the origin of these HiVel stars, we traced the backwards orbits of each HiVel star in the Galactic potential to derive probability parameters which are used to classify these HiVel stars. Of these, 5 stars are from the tidal debris of disrupted dwarf galaxy and 19 stars are runaway-star candidates which originate from the stellar disk.

astro-ph.GA↗

arXiv2018

Attentive Long Short-Term Preference Modeling for Personalized Product Search

E-commerce users may expect different products even for the same query, due to their diverse personal preferences. It is well-known that there are two types of preferences: long-term ones and short-term ones. The former refers to user' inherent purchasing bias and evolves slowly. By contrast, the latter reflects users' purchasing inclination in a relatively short period. They both affect users' current purchasing intentions. However, few research efforts have been dedicated to jointly model them for the personalized product search. To this end, we propose a novel Attentive Long Short-Term Preference model, dubbed as ALSTP, for personalized product search. Our model adopts the neural networks approach to learn and integrate the long- and short-term user preferences with the current query for the personalized product search. In particular, two attention networks are designed to distinguish which factors in the short-term as well as long-term user preferences are more relevant to the current query. This unique design enables our model to capture users' current search intentions more accurately. Our work is the first to apply attention mechanisms to integrate both long- and short-term user preferences with the given query for the personalized search. Extensive experiments over four Amazon product datasets show that our model significantly outperforms several state-of-the-art product search methods in terms of different evaluation metrics.

cs.IR↗

arXiv2018

The First Data Release of the Beijing-Arizona Sky Survey

The Beijing-Arizona Sky Survey (BASS) is a new wide-field legacy imaging survey in the northern Galactic cap using the 2.3m Bok telescope. The survey will cover about 5400 deg$^2$ in the $g$ and $r$ bands, and the expected 5$σ$ depths (corrected for the Galactic extinction) in the two bands are 24.0 and 23.4 mag, respectively. BASS started observations in January 2015, and has completed about 41% of the whole area as of July 2016. The first data release contains both calibrated images and photometric catalogs obtained in 2015 and 2016. The depths of single-epoch images in the two bands are 23.4 and 22.9 mag, and the full depths of three epochs are about 24.1 and 23.5 mag, respectively.

astro-ph.GA↗

arXiv2018

The Second Data Release of the Beijing-Arizona Sky Survey

This paper presents the second data release (DR2) of the Beijing-Arizona Sky Survey (BASS). BASS is an imaging survey of about 5400 deg$^2$ in $g$ and $r$ bands using the 2.3 m Bok telescope. DR2 includes the observations as of July 2017 obtained by BASS and Mayall $z$-band Legacy Survey (MzLS). This is our first time to include the MzLS data covering the same area as BASS. BASS and MzLS have respectively completed about 72% and 76% of their observations. The two surveys will be served for the spectroscopic targeting of the upcoming Dark Energy Spectroscopic Instrument. Both BASS and MzLS data are reduced by the same pipeline. We have updated the basic data reduction and photometric methods in DR2. In particular, source detections are performed on stacked images, and photometric measurements are co-added from single-epoch images based on these sources. The median 5$σ$ depths with corrections of the Galactic extinction are 24.05, 23.61, and 23.10 mag for $g$, $r$, and $z$ bands, respectively. The DR2 data products include stacked images, co-added catalogs, and single-epoch images and catalogs. The BASS website (http://batc.bao.ac.cn/BASS/) provides detailed information and links to download the data.

astro-ph.IM↗

arXiv2018

The relation between HI Gas and Star Formation Properties in Nearby Galaxies

In this paper, we present some correlations of neutral hydrogen HI gas and physical properties of galaxies to investigate the role of atomic gas in governing galaxy evolution. We build a HI-detected sample including 70 galaxies that are detected by ALFALFA in a 15 deg$^2$ region, and derive their star formation distribution based on the images of $\rm Hα$ narrow-band observed here. In general, HI-detected galaxies have low surface density of stellar mass and active star formation. Additionally, most of the galaxies are in good agreement with the star-forming main sequence, consistent with the previous findings. We confirm the dependence of star formation (SF) in galaxies on HI gas at least on global scale, i.e., star formation rate (SFR) generally increases with HI mass, specific star formation rate (SSFR$\equiv$SFR/M$_*$) increases with HI fraction ($f_{HI}$) even for a given stellar mass, and \hi-based star formation efficiency (SFE) mildly increases with the stellar mass and SFR surface density. Based on the distribution of stellar mass and star formation, we calculate the morphology indices of the sample, and analyze the dependence of $f_{HI}$ and SFE on them. The weak correlations between SFE and morphological indexes imply a weak physical link between HI and star formation in small scale. We find that $f_{HI}$ mildly increases with the asymmetry and decreases with the concentration of galaxies, suggesting that the HI gas supply and its effect are likely correlated with external processes in the extended disks of galaxies.

astro-ph.GA↗

arXiv2018

LinkNBed: Multi-Graph Representation Learning with Entity Linkage

Knowledge graphs have emerged as an important model for studying complex multi-relational data. This has given rise to the construction of numerous large scale but incomplete knowledge graphs encoding information extracted from various resources. An effective and scalable approach to jointly learn over multiple graphs and eventually construct a unified graph is a crucial next step for the success of knowledge-based inference for many downstream applications. To this end, we propose LinkNBed, a deep relational learning framework that learns entity and relationship representations across multiple graphs. We identify entity linkage across graphs as a vital component to achieve our goal. We design a novel objective that leverage entity linkage and build an efficient multi-task training procedure. Experiments on link prediction and entity linkage demonstrate substantial improvements over the state-of-the-art relational learning approaches.

cs.LG↗

arXiv2018

Multi-color optical monitoring of ten blazars from 2005 to 2011

We carried out multi-color optical monitoring of a sample of ten blazars from 2005 to 2011. The sample contains 3 LBLs, 2 IBLs, 4 HBLs, and 1 FSRQ. Our monitoring focused on the long-term variability and the sample included nine BL Lac objects and one flat-spectrum radio quasar. A total number of 14799 data points were collected. This is one of the largest optical database for a sample of ten blazars. All objects showed significant variability except OT 546. Because of the low sampling on each single night, only BL Lacertae was observed to have intra-day variability on 2006 November 6. Most BL Lac objects showed a bluer-when-brighter chromatism, while the flat-spectrum radio quasar, 3C 454.3, displayed a redder-when-brighter trend. The BWB color behaviors of most BL Lacs can be at least partly interpreted by the fact of increasing variation amplitude with increasing frequency observed in these objects. The average spectral index of LBLs is around 1.5, as expected from the model dominated by Synchrotron Self-Compton (SSC) loss. The optical emission of HBL is probably contaminated by the thermal emission from the host galaxies. Correlation analysis did not reveal any time delay between variations at different wavelengths.

astro-ph.GA↗

arXiv2018

Metallicity and Kinematics of the Galactic halo from the LAMOST sample stars

We study the metallicity distribution and kinematic properties of 4,680 A/F/G/K-type giant stars with $|z|>$ 5 kpc selected from the LAMOST spectroscopic survey. The metallicity distribution of giant stars with 5 $<|z|\leqslant$ 15 kpc can be described by a three-peak Gaussian model with peaks at [Fe/H] $\sim-0.6\pm0.1$, $-1.2\pm0.3$ and $-2.0\pm0.2$, corresponding to the ratio of 19$\%$, 74$\%$ and 7$\%$, respectively. The $\rm{[α/Fe]}$ is used to associate the three peaks with the thick disk, inner-halo and outer-halo components of the Galaxy. The metallicity distribution of these giant stars, which is fit with Gaussians corresponding to the three components, show a growing fraction of inner-halo component and declining fraction of the thick-disk component with increasing distance from the Galactic plane. Adopting a galaxy potential model, we also derive the orbital parameters of the sample stars, such as orbit eccentricity and rotation velocity. The peak values of derived orbital eccentricity for stars covering different metallicity regions maintain $e\sim$ 0.75, independent of height above the plane, within the range 5$<|z|<$ 15 kpc. By comparing the MDFs of stars in different rotation velocity intervals, we find that the majority of the retrograde stars are more metal-poor than the prograde stars.

astro-ph.GA↗

arXiv2018

Photometric calibration for the Beijing-Arizona Sky Survey and Mayall z-band Legacy Survey

We present the photometric calibration of the Beijing-Arizona Sky Survey (BASS) and Mayall z-band Legacy Survey (MzLS), which are two of the three wide-field optical legacy imaging surveys to provide the baseline targeting data for the Dark Energy Spectroscopic Instrument (DESI) project. The method of our photometric calibration is subdivided into the external and internal processes. The former utilizes the point-source objects of Pan-STARRS1 survey (PS1) as the reference standards to achieve the zero points of the absolute flux for individual exposures. And then the latter revise the zero points to make them consistent across the survey based on multiple tilings and large offset overlaps. Our process achieves a homogeneous photometric calibration over most of the sky with precision better than 10 mmag for g and r bands, 15 mmag for z band. The accuracy of the calibration is better than 1% at the bright end (16-18 mag) over most of the survey area.

astro-ph.IM↗

arXiv2018

On certain combinatorial expansions of the Legendre-Stirling numbers

The Legendre-Stirling numbers of the second kind were introduced by Everitt et al. in the spectral theory of powers of the Legendre differential expressions. In this paper, we provide a combinatorial code for Legendre-Stirling set partitions. As an application, we obtain combinatorial expansions of the Legendre-Stirling numbers of both kinds. Moreover, we present grammatical descriptions of the Jacobi-Stirling numbers of both kinds.

math.CO↗

arXiv2018

Chemi-net: a graph convolutional network for accurate drug property prediction

Absorption, distribution, metabolism, and excretion (ADME) studies are critical for drug discovery. Conventionally, these tasks, together with other chemical property predictions, rely on domain-specific feature descriptors, or fingerprints. Following the recent success of neural networks, we developed Chemi-Net, a completely data-driven, domain knowledge-free, deep learning method for ADME property prediction. To compare the relative performance of Chemi-Net with Cubist, one of the popular machine learning programs used by Amgen, a large-scale ADME property prediction study was performed on-site at Amgen. The results showed that our deep neural network method improved current methods by a large margin. We foresee that the significantly increased accuracy of ADME prediction seen with Chemi-Net over Cubist will greatly accelerate drug discovery.

cs.LG↗

arXiv2018

The ascent-plateau statistics on Stirling permutations

In this paper, several variants of the ascent-plateau statistic are introduced, including flag ascent-plateau, double ascent and descent-plateau. We first study the flag ascent-plateau statistic on Stirling permutations by using context-free grammars. We then present a unified refinement of the ascent polynomials and the ascent-plateau polynomials. In particular, by using Foata and Strehl's group action, we prove two bistatistics over the set of Stirling permutations of order n are equidistributed.

math.CO↗

arXiv2017

Recurrence relations for binomial-Eulerian polynomials

Binomial-Eulerian polynomials were introduced by Postnikov, Reiner and Williams. In this paper, properties of the binomial-Eulerian polynomials, including recurrence relations and generating functions are studied. We present three constructive proofs of the recurrence relations for binomial-Eulerian polynomials. Moreover, we give a combinatorial interpretation of the Betti number of the complement of the k-equal real hyperplane arrangement.

math.CO↗

arXiv2017

Message-Passing Receiver Design for Joint Channel Estimation and Data Decoding in Uplink Grant-Free SCMA Systems

The conventional grant-based network relies on the handshaking between base station and active users to achieve dynamic multi-user scheduling, which may cost large signaling overheads as well as system latency. To address those problems, the grant-free receiver design is considered in this paper based on sparse code multiple access (SCMA), one of the promising air interface technologies for 5G wireless networks. With the presence of unknown multipath fading, the proposed receiver blindly performs joint channel estimation and data decoding without knowing the user activity in the network. Based on the framework of belief propagation (BP), we formulate a messagepassing receiver for uplink SCMA that performs joint estimation iteratively. However, the direct application of BP for the multivariable detection problem is complex. Motivated by the idea of approximate inference, we use expectation propagation to project the intractable distributions into Gaussian families such that a linear complexity decoder is obtained.Simulation results show that the proposed receiver can detect active users in the network with a high accuracy and can achieve an improved bit-error-rate performance compared with existing methods.

eess.SP↗

arXiv2017

Neural Attentive Session-based Recommendation

Given e-commerce scenarios that user profiles are invisible, session-based recommendation is proposed to generate recommendation results from short sessions. Previous work only considers the user's sequential behavior in the current session, whereas the user's main purpose in the current session is not emphasized. In this paper, we propose a novel neural networks framework, i.e., Neural Attentive Recommendation Machine (NARM), to tackle this problem. Specifically, we explore a hybrid encoder with an attention mechanism to model the user's sequential behavior and capture the user's main purpose in the current session, which are combined as a unified session representation later. We then compute the recommendation scores for each candidate item with a bi-linear matching scheme based on this unified session representation. We train NARM by jointly learning the item and session representations as well as their matchings. We carried out extensive experiments on two benchmark datasets. Our experimental results show that NARM outperforms state-of-the-art baselines on both datasets. Furthermore, we also find that NARM achieves a significant improvement on long sessions, which demonstrates its advantages in modeling the user's sequential behavior and main purpose simultaneously.

cs.IR↗

arXiv2017

Mapping electronic states of dual-parallel and symmetric zigzag grain boundaries of graphene on highly oriented pyrolytic graphite

The grain boundaries (GBs) of a graphene surface were extensively studied because GBs with specific defect configurations result in the formation of new curved structures, which can be treated as new carbon allotropes. We studied the structures and electronic spectra of two periodic GBs in graphene on highly oriented pyrolytic graphite (HOPG) surfaces using scanning tunneling microscopy and spectroscopy (STM/S). Our results demonstrated that a GB consisting of dual parallel periodic dislocation cores of pentagonal-heptagonal (5-7) carbon rings gives rise to an enhanced localized state at 0.45 eV above the Dirac point in graphene surfaces, which is attributed to van Hove singularities (VHSs). Moreover, the energy positions of the localized states are varied between 0.40 and 0.47 eV depending on the site and eventually decayed to 0.36 eV. The variation of the energy positions is induced by two parallel GBs because of the higher electron density at the GB as a defect center and the lower electron away from the GB. Another periodic GB with a symmetric zigzag structural feature induced VHSs at -0.038 and 0.12 eV near the Fermi level. Moreover, intervalley scattering was observed on both these GBs. This means that carrier concentration and thus, the conductance around the periodic GBs can be significantly enhanced by localized density of states. This finding suggests that graphene with a properly embedded ordered GB is promising for improving the performance of graphene-based electronic devices.

cond-mat.mtrl-sci↗

arXiv2017

Word-representability of split graphs

Letters $x$ and $y$ alternate in a word $w$ if after deleting in $w$ all letters but the copies of $x$ and $y$ we either obtain a word $xyxy\cdots$ (of even or odd length) or a word $yxyx\cdots$ (of even or odd length). A graph $G=(V,E)$ is word-representable if and only if there exists a word $w$ over the alphabet $V$ such that letters $x$ and $y$ alternate in $w$ if and only if $xy\in E$. It is known that a graph is word-representable if and only if it admits a certain orientation called semi-transitive orientation. Word-representable graphs generalize several important classes of graphs such as $3$-colorable graphs, circle graphs, and comparability graphs. There is a long line of research in the literature dedicated to word-representable graphs. However, almost nothing is known on word-representability of split graphs, that is, graphs in which the vertices can be partitioned into a clique and an independent set. In this paper, we shed a light to this direction. In particular, we characterize in terms of forbidden subgraphs word-representable split graphs in which vertices in the independent set are of degree at most 2, or the size of the clique is 4. Moreover, we give necessary and sufficient conditions for an orientation of a split graph to be semi-transitive.

math.CO↗