arXiv subjects

Ning Hu

Publications and source records attributed to Ning Hu.

33 records · Page 2Linked to original sources

arXiv2023

Exploring the Spatiotemporal Features of Online Food Recommendation Service

Online Food Recommendation Service (OFRS) has remarkable spatiotemporal characteristics and the advantage of being able to conveniently satisfy users' needs in a timely manner. There have been a variety of studies that have begun to explore its spatiotemporal properties, but a comprehensive and in-depth analysis of the OFRS spatiotemporal features is yet to be conducted. Therefore, this paper studies the OFRS based on three questions: how spatiotemporal features play a role; why self-attention cannot be used to model the spatiotemporal sequences of OFRS; and how to combine spatiotemporal features to improve the efficiency of OFRS. Firstly, through experimental analysis, we systemically extracted the spatiotemporal features of OFRS, identified the most valuable features and designed an effective combination method. Secondly, we conducted a detailed analysis of the spatiotemporal sequences, which revealed the shortcomings of self-attention in OFRS, and proposed a more optimized spatiotemporal sequence method for replacing self-attention. In addition, we also designed a Dynamic Context Adaptation Model to further improve the efficiency and performance of OFRS. Through the offline experiments on two large datasets and online experiments for a week, the feasibility and superiority of our model were proven.

cs.IR↗

arXiv2023

ESMC: Entire Space Multi-Task Model for Post-Click Conversion Rate via Parameter Constraint

Large-scale online recommender system spreads all over the Internet being in charge of two basic tasks: Click-Through Rate (CTR) and Post-Click Conversion Rate (CVR) estimations. However, traditional CVR estimators suffer from well-known Sample Selection Bias and Data Sparsity issues. Entire space models were proposed to address the two issues via tracing the decision-making path of "exposure_click_purchase". Further, some researchers observed that there are purchase-related behaviors between click and purchase, which can better draw the user's decision-making intention and improve the recommendation performance. Thus, the decision-making path has been extended to "exposure_click_in-shop action_purchase" and can be modeled with conditional probability approach. Nevertheless, we observe that the chain rule of conditional probability does not always hold. We report Probability Space Confusion (PSC) issue and give a derivation of difference between ground-truth and estimation mathematically. We propose a novel Entire Space Multi-Task Model for Post-Click Conversion Rate via Parameter Constraint (ESMC) and two alternatives: Entire Space Multi-Task Model with Siamese Network (ESMS) and Entire Space Multi-Task Model in Global Domain (ESMG) to address the PSC issue. Specifically, we handle "exposure_click_in-shop action" and "in-shop action_purchase" separately in the light of characteristics of in-shop action. The first path is still treated with conditional probability while the second one is treated with parameter constraint strategy. Experiments on both offline and online environments in a large-scale recommendation system illustrate the superiority of our proposed methods over state-of-the-art models. The real-world datasets will be released.

cs.AI↗

arXiv2022

BASM: A Bottom-up Adaptive Spatiotemporal Model for Online Food Ordering Service

Online Food Ordering Service (OFOS) is a popular location-based service that helps people to order what you want. Compared with traditional e-commerce recommendation systems, users' interests may be diverse under different spatiotemporal contexts, leading to various spatiotemporal data distribution, which limits the fitting capacity of the model. However, numerous current works simply mix all samples to train a set of model parameters, which makes it difficult to capture the diversity in different spatiotemporal contexts. Therefore, we address this challenge by proposing a Bottom-up Adaptive Spatiotemporal Model(BASM) to adaptively fit the spatiotemporal data distribution, which further improve the fitting capability of the model. Specifically, a spatiotemporal-aware embedding layer performs weight adaptation on field granularity in feature embedding, to achieve the purpose of dynamically perceiving spatiotemporal contexts. Meanwhile, we propose a spatiotemporal semantic transformation layer to explicitly convert the concatenated input of the raw semantic to spatiotemporal semantic, which can further enhance the semantic representation under different spatiotemporal contexts. Furthermore, we introduce a novel spatiotemporal adaptive bias tower to capture diverse spatiotemporal bias, reducing the difficulty to model spatiotemporal distinction. To further verify the effectiveness of BASM, we also novelly propose two new metrics, Time-period-wise AUC (TAUC) and City-wise AUC (CAUC). Extensive offline evaluations on public and industrial datasets are conducted to demonstrate the effectiveness of our proposed modle. The online A/B experiment also further illustrates the practicability of the model online service. This proposed method has now been implemented on the Ele.me, a major online food ordering platform in China, serving more than 100 million online users.

cs.LG↗

arXiv2022

Spatiotemporal-Enhanced Network for Click-Through Rate Prediction in Location-based Services

In Location-Based Services(LBS), user behavior naturally has a strong dependence on the spatiotemporal information, i.e., in different geographical locations and at different times, user click behavior will change significantly. Appropriate spatiotemporal enhancement modeling of user click behavior and large-scale sparse attributes is key to building an LBS model. Although most of existing methods have been proved to be effective, they are difficult to apply to takeaway scenarios due to insufficient modeling of spatiotemporal information. In this paper, we address this challenge by seeking to explicitly model the timing and locations of interactions and proposing a Spatiotemporal-Enhanced Network, namely StEN. In particular, StEN applies a Spatiotemporal Profile Activation module to capture common spatiotemporal preference through attribute features. A Spatiotemporal Preference Activation is further applied to model the personalized spatiotemporal preference embodied by behaviors in detail. Moreover, a Spatiotemporal-aware Target Attention mechanism is adopted to generate different parameters for target attention at different locations and times, thereby improving the personalized spatiotemporal awareness of the model.Comprehensive experiments are conducted on three large-scale industrial datasets, and the results demonstrate the state-of-the-art performance of our methods. In addition, we have also released an industrial dataset for takeaway industry to make up for the lack of public datasets in this community.

cs.IR↗

arXiv2021

Gradient zigzag metamaterial beams as broadband vibration isolators for beam-like structures

Phononic crystals (PCs) and metamaterials are artificially structured materials with the unprecedented property of the existence of complete stop-bands. However, the generally narrow width of the stop-bands in such materials severely limits their applicability. Here, a particular kind of gradient zigzag metamaterial beams exclusively with flat bands within a broad frequency range is proposed based upon our previously proposed uniform counterparts. The calculated band structures by the Finite Element Method (FEM) and the Spectral Element Method (SEM) show the distinctly different characteristics of the gradient zigzag metamaterial beams from the conventional PCs and metamaterials, where the band structures are totally filled with flat bands within a broad frequency range without the normal pass bands. These flat bands are caused by the local resonance of certain parts in the gradient zigzag metamaterial beams, and the elastic waves with the frequencies at these flat bands are trapped in the resonant parts and cannot propagate. As a result, our proposed gradient zigzag metamaterial beams can be used as broadband vibration isolators for beam-like structures, which is verified by both numerical simulations and experiments conducted on the 3D-printed samples. This design strategy opens a new avenue for designing and constructing novel broadband vibration isolators.

physics.app-ph↗

arXiv2020

Dust Attenuation Curve for Local Subgalactic Star-forming Regions

We compile a sample of about 157,000 spaxels from the Mapping Nearby Galaxies at the Apache Point Observatory survey to derive the average dust attenuation curve for subgalactic star-forming regions of local star-forming galaxies (SFGs) in the optical wavelength, following the method of \cite{Calzetti1994}. We obtain a $D_n(4000)$-independent average attenuation curve for spaxels with $1.1\leq D_n(4000)<1.3$, which is similar to the one derived from either local starbursts or normal SFGs. We examine whether and how the shape of the average attenuation curve changes with several local and global physical properties. For spaxels with $1.2\leq D_n(4000)<1.3$, we find no dependence on either local or global physical properties for the shape of the average attenuation curve. However, for spaxels with younger stellar population ($1.1\leq D_n(4000)<1.2$), shallower average attenuation curves are found for star-forming regions with smaller stellar mass surface density, smaller star formation rate surface density, or those residing in the outer region of galaxies. These results emphasize the risk of using one single attenuation curve to correct the dust reddening for all types of star-forming regions, especially for those with fairly young stellar population.

astro-ph.GA↗

arXiv2020

Physical Properties of H II Regions in M51 from Spectroscopic Observations

M51 and NGC 5195 is an interacting system that can be explored in great details with ground-based telescopes. The H II regions in M51 were observed using the 2.16 m telescope of the National Astronomical Observatories of the Chinese Academy of Sciences and the 6.5 m Multiple Mirror Telescope with spatial resolution of less than $\sim100$ pc. We obtain a total of 113 spectra across the galaxy and combine the literature data of Croxall et al. to derive a series of physical properties, including the gas-phase extinction, stellar population age, star formation rate (SFR) surface density, and oxygen abundance. The spatial distributions and radial profiles of these properties are investigated in order to study the characteristics of M51 and the clues to the formation and evolution of this galaxy. M51 presents a mild radial extinction gradient. The lower gas-phase extinction in the north spiral arms compared to the south arms are possibly caused by the past encounters with the companion galaxy of NGC 5195. A number of H II regions have the stellar age between 50 and 500 Myr, consistent with the recent interaction history by simulations in the literatures. The SFR surface density presents a mild radial gradient, which is ubiquitous in spiral galaxies. There is a negative metallicity gradient of $-0.08$ dex $R_{e}^{-1}$ in the disk region, which is also commonly found in many spiral galaxies. It is supported by the "inside-out" scenario of galaxy formation. We find a positive abundance gradient of 0.26 dex $R_{e}^{-1}$ in the inner region. There are possible reasons causing the positive gradient, including the freezing of the chemical enrichment due to the star-forming quenching in the bulge and the gas infall and dilution due to the pseudobulge growth and/or galactic interaction.

astro-ph.GA↗

arXiv2019

The role of second-order radial density gradient for helicon power absorption

To reveal the mysterious and still controversial mechanism of high ionization efficiency during helicon discharges, this work focuses particularly on the role of second-order derivative in radial density profile, both analytical and numerically. It is found that: (i) the radially localized potential well that plays a critical role in resonance power absorption from antenna to plasma is localized where the second-order derivative vanishes, (ii) the power absorption increases for positive second-order derivative, decreases for negative second-order derivative, and maximizes where second-order derivative becomes zero, (iii) the power absorption decreases near plasma core and increases near plasma edge when the radial location of vanishing second-order derivative moves outwards, which is also a process of Trivelpiece-Gould mode overwhelming helicon mode. These findings can be very interesting for helicon plasma applications that require certain power distribution or heat flux configuration, e. g. material processing, which can be controlled by adjusting the profile and zero-crossing location of second-order radial density gradient.

physics.plasm-ph↗

arXiv2018

Coupling of RF Antennas to Large Volume Helicon Plasma

Large volume helicon plasma sources are of particular interest for large scale semiconductor processing, high power plasma propulsion and recently plasma-material interaction under fusion conditions. This work is devoted to studying the coupling of four typical RF antennas to helicon plasma with infinite length and diameter of $0.5$~m, and exploring its frequency dependence in the range of $13.56-70$~MHz for coupling optimization. It is found that loop antenna is more efficient than half helix, Boswell and Nagoya III antennas for power absorption; radially parabolic density profile overwhelms Gaussian density profile in terms of antenna coupling for low-density plasma, but the superiority reverses for high-density plasma. Increasing the driving frequency results in power absorption more near plasma edge, but the overall power absorption increases with frequency. Perpendicular stream plots of wave magnetic field, wave electric field and perturbed current are also presented. This work can serve as an important reference for the experimental design of large volume helicon plasma source with high RF power.

physics.plasm-ph↗

arXiv2018

What determines the local metallicity of galaxies: global stellar mass, local stellar mass surface density or star formation rate?

The metallicity and its relationship with other galactic properties is a fundamental probe of the evolution of galaxies. In this work, we select about 750,000 star-forming spatial pixels from 1122 blue galaxies in the MaNGA survey to investigate the global stellar mass - local stellar mass surface density - gas-phase metallicity ($M_*$ - $Σ_*$ - $Z$ ) relation. At a fixed $M_*$, the metallicity increases steeply with increasing $Σ_*$. Similarly, at a fixed $Σ_*$, the metallicity increases strongly with increasing $M_*$ at low mass end, while this trend becomes less obvious at high mass end. We find the metallicity to be more strongly correlated to $Σ_*$ than to $M_*$. Furthermore, we construct a tight (0.07 dex scatter) $M_*$ - $Σ_*$ - $Z$ relation, which reduces the scatter in the $Σ_*$ - $Z$ relation by about 30$\%$ for galaxies with $7.8 < {\rm log}(M_*/M_\odot) < 11.0$, while the reduction of scatter is much weaker for high-mass galaxies. This result suggests that, especially for low-mass galaxies, the $M_*$ - $Σ_*$ - $Z$ relation is largely more fundamental than the $M_*$ - $Z$ and $Σ_*$ - $Z$ relations, meaning that both $M_*$ and $Σ_*$ play important roles in shaping the local metallicity. We also find that the local metallicity is probably independent on the local star formation rate surface density at a fixed $M_*$ and $Σ_*$. Our results are consistent with the scenario that the local metallicities in galaxies are shaped by the combination of the local stars formed in the history and the metal loss caused by galactic winds.

astro-ph.GA↗

arXiv2018

M101: Spectral Observations of HII Regions and Their Physical Properties

By using the Hectospec 6.5 m Multiple Mirror Telescope (MMT) and the 2.16 m telescope of National Astronomical Observatories, Chinese Academy of Sciences (NAOC), we obtained 188 high signal-to-noise ratio (S/N) spectra of HII regions in the nearby galaxy M101, which are the largest spectroscopic sample of HII regions for this galaxy so far. These spectra cover a wide range of regions on M101, which enables us to analyze two dimensional distributions of its physical properties. The physical parameters are derived from emission lines or stellar continuum, including stellar population age, electron temperature, oxygen abundance and etc. The oxygen abundances are derived using two empirical methods based on O3N2 and R$_{23}$ indicators, as well as the direct Te method when OIII$\lambda4363$ is available. By applying the harmonic decomposition analysis to the velocity field, we obtained line-of-sight rotation velocity of 71 km s$^{-1}$ and a position angle of 36 degree. The stellar age profile shows an old stellar population in galaxy center and a relative young stellar population in outer regions, suggesting an old bulge and a young disk. Oxygen abundance profile exhibits a clear break at $\sim$18 kpc, with a gradient of $-$0.0364 dex kpc$^{-1}$ in the inner region and $-$0.00686 dex kpc$^{-1}$ in the outer region. Our results agree with the "inside-out" disk growth scenario of M101.

astro-ph.GA↗

arXiv2017

Spectroscopic Observation and Analysis of HII regions in M33 with MMT: Temperatures and Oxygen Abundances

The spectra of 413 star-forming (or HII) regions in M33 (NGC 598) were observed by using the multifiber spectrograph of Hectospec at the 6.5-m Multiple Mirror Telescope (MMT). By using this homogeneous spectra sample, we measured the intensities of emission lines and some physical parameters, such as electron temperatures, electron densities, and metallicities. Oxygen abundances were derived via the direct method (when available) and two empirical strong-line methods, namely, O3N2 and N2. In the high-metallicity end, oxygen abundances derived from O3N2 calibration were higher than those derived from N2 index, indicating an inconsistency between O3N2 and N2 calibrations. We presented a detailed analysis of the spatial distribution of gas-phase oxygen abundances in M33 and confirmed the existence of the axisymmetric global metallicity distribution widely assumed in literature. Local variations were also observed and subsequently associated with spiral structures to provide evidence of radial migration driven by arms. Our O/H gradient fitted out to 1.1 $R_{25}$ resulted in slopes of $-0.17\pm0.03$, $-0.19\pm0.01$, and $-0.16\pm0.17$ dex $R_{25}^{-1}$ utilizing abundances from O3N2, N2 diagnostics, and direct method, respectively.

astro-ph.GA↗

arXiv2017

A sample of metal-poor galaxies identified from the LAMOST spectral survey

We present a sample of 48 metal-poor galaxies at $ z < 0.14$ selected from 92,510 galaxies in the LAMOST survey. These galaxies are identified for their detection of the auroral emission line \oiii$λ$4363 above $3σ$ level, which allows a direct measurement of the electron temperature and the oxygen abundance. The emission line fluxes are corrected for internal dust extinction using Balmer decrement method. With electron temperature derived from \oiii$λ\lambda4959,5007/\oiii\lambda4363$ and electron density from $\sii\lambda6731/\sii\lambda6717$, we obtain the oxygen abundances in our sample which range from $\zoh= 7.63$ (0.09 $\Zsun$) to $8.46$ (0.6 $\Zsun$). We find an extremely metal-poor galaxy with $\zoh=7.63 \pm 0.01$. With multiband photometric data from FUV to NIR and $\ha$ measurements, we also determine the stellar masses and star formation rates, based on the spectral energy distribution fitting and $\ha$ luminosity, respectively. We find that our galaxies have low and intermediate stellar masses with $\rm 6.39 \le log(M/M_{\sun})\le 9.27$, and high star formation rates (SFRs) with $\rm -2.18 \le log(SFR/M_{\sun} yr^{-1}) \le 1.95$. We also find that the metallicities of our galaxies are consistent with the local $T_e$-based mass-metallicity relation, while the scatter is about 0.28 dex. Additionally, assuming the coefficient of $\rm α=0.66$, we find most of our galaxies follow the local mass-metallicity-SFR relation, while a scatter about 0.24 dex exists, suggesting the mass-metallicity relation is weakly dependent on SFR for those metal-poor galaxies.

astro-ph.GA↗

arXiv2016

The metallicity evolution of blue compact dwarf galaxies from the intermediate redshift to the local Universe

We present oxygen abundance measurements for 74 blue compact dwarf (BCD) galaxies in the redshift range in [0.2, 0.5] using the strong-line method. The spectra of these objects are taken using Hectospec on the Multiple Mirror Telescope (MMT). More than half of these BCDs had dust attenuation corrected using the Balmer decrement method. For comparison, we also selected a sample of 2023 local BCDs from the Sloan Digital Sky Survey (SDSS) database. Based on the local and intermediate-z BCD samples, we investigated the cosmic evo- lution of the metallicity, star-formation rate (SFR), and Dn (4000) index. Com- pared with local BCDs, the intermediate-z BCDs had a systematically higher R23 ratio but similar O32 ratio. Interestingly, no significant deviation in the mass-metallicity (MZ) relation was found between the intermediate-z and lo- cal BCDs. Besides the metallicity, the intermediate-z BCDs also exhibited an SFR distribution that was consistent with local BCDs, suggesting a weak de- pendence on redshift. The intermediate-z BCDs seemed to be younger than the local BCDs with lower D4000 index values. The insignificant deviation in the mass-metallicity and mass-SFR relations between intermediate-z and local BCDs indicates that the relations between the global parameters of low-mass compact galaxies may be universal. These results from low mass compact galaxies could be used to place important observational constraints on galaxy formation and evolution models.

astro-ph.GA↗

arXiv2015

Spectral Synthesis via Mean Field approach Independent Component Analysis

In this paper, we apply a new statistical analysis technique, Mean Field approach to Bayesian Independent Component Analysis (MF-ICA), on galaxy spectral analysis. This algorithm can compress the stellar spectral library into a few Independent Components (ICs), and galaxy spectrum can be reconstructed by these ICs. Comparing to other algorithms which decompose a galaxy spectrum into a combination of several simple stellar populations, MF-ICA approach offers a large improvement in the efficiency. To check the reliability of this spectral analysis method, three different methods are used: (1) parameter-recover for simulated galaxies, (2) comparison with parameters estimated by other methods, and (3) consistency test of parameters from the Sloan Digital Sky Survey galaxies. We find that our MF-ICA method not only can fit the observed galaxy spectra efficiently, but also can recover the physical parameters of galaxies accurately. We also apply our spectral analysis method to the DEEP2 spectroscopic data, and find it can provide excellent fitting for those low signal-to-noise spectra.

astro-ph.GA↗