arXiv subjects

Leilei Sun

Publications and source records attributed to Leilei Sun.

46 records · Page 3Linked to original sources

arXiv2022

WASP-35 and HAT-P-30/WASP-51: re-analysis using TESS and ground-based transit photometry

High-precision transit observations provide excellent opportunities for characterizing the physical properties of exoplanetary systems. These physical properties supply many pieces of information for unvealing the internal structure, external atmosphere, and dynamical history of the planets. We present revised properties of transiting systems WASP-35 and HAT-P-30/WASP-51 through analyzing newly available TESS photometry and ground-based observations obtained at 1m telescope of Yunnan Observatories as well as from the literature. The improved system parameters are consistent with the previous results. Furthermore, we find that HAT-P-30b/WASP-51b's transits show significant timing variation which cannot be explained by decaying orbit due to tidal dissipation and the Rømer effect, while both apsidal precession and an additional perturbing body could reproduce this signal through our comprehensive dynamical simulations. Because both of them are valuable targets which are suitable for transmission spectroscopy, we make some predictions for atmospheric properties of WASP-35b and HAT-P-30b/WASP-51b based on newly derived system parameters.

astro-ph.EP↗

arXiv2021

A highly mutually-inclined, compact warm-Jupiter system KOI-984 ?

The discovery of a population of close-orbiting giant planets ($\le$ 1 au) has raised a number of questions about their origins and dynamical histories. These issues have still not yet been fully resolved, despite over 20 years of exoplanet detections and a large number of discovered exoplanets. In particular, it is unclear whether warm Jupiters (WJs) form in situ, or whether they migrate from further outside and are even currently migrating to form hot Jupiters (HJs). Here, we report the possible discovery and characterization of the planets in a highly mutually-inclined ($I_{\rm mut}\simeq 45^\circ$), compact two-planet system (KOI-984), in which the newly discovered warm Jupiter KOI-984$c$ is on a 21.5-day, moderately eccentric ($e\simeq 0.4$) orbit, in addition to a previously known 4.3-day planet candidate KOI-984$b$. Meanwhile, the orbital configuration of a moderately inclined ($I_{\rm mut}\simeq 15^\circ$), low-mass ($m_{c}\simeq 24 M_{\oplus}$;$P_b\simeq 8.6$ days) perturbing planet near 1:2 mean motion resonace with KOI-984$b$ could also well reproduce observed transit timing variations and transit duration variations of KOI-984$b$. Such an eccentric WJ with a close-in sibling would pose a challenge to the proposed formation and migration mechanisms of WJs, if the first scenario is supported with more evidences in near future; this system with several other well-measured inclined WJ systems (e.g., Kepler-419 and Kepler-108) may provide additional clues for the origin and dynamical histories of WJs.

astro-ph.EP↗

arXiv2021

Deep Human-guided Conditional Variational Generative Modeling for Automated Urban Planning

Urban planning designs land-use configurations and can benefit building livable, sustainable, safe communities. Inspired by image generation, deep urban planning aims to leverage deep learning to generate land-use configurations. However, urban planning is a complex process. Existing studies usually ignore the need of personalized human guidance in planning, and spatial hierarchical structure in planning generation. Moreover, the lack of large-scale land-use configuration samples poses a data sparsity challenge. This paper studies a novel deep human guided urban planning method to jointly solve the above challenges. Specifically, we formulate the problem into a deep conditional variational autoencoder based framework. In this framework, we exploit the deep encoder-decoder design to generate land-use configurations. To capture the spatial hierarchy structure of land uses, we enforce the decoder to generate both the coarse-grained layer of functional zones, and the fine-grained layer of POI distributions. To integrate human guidance, we allow humans to describe what they need as texts and use these texts as a model condition input. To mitigate training data sparsity and improve model robustness, we introduce a variational Gaussian embedding mechanism. It not just allows us to better approximate the embedding space distribution of training data and sample a larger population to overcome sparsity, but also adds more probabilistic randomness into the urban planning generation to improve embedding diversity so as to improve robustness. Finally, we present extensive experiments to validate the enhanced performances of our method.

cs.CV↗

arXiv2021

Analysis for full face mechanical behaviors through spatial deduction model with real-time monitoring data

Mechanical analysis for the full face of tunnel structure is crucial to maintain stability, which is a challenge in classical analytical solutions and data analysis. Along this line, this study aims to develop a spatial deduction model to obtain the full-faced mechanical behaviors through integrating mechanical properties into pure data-driven model. The spatial tunnel structure is divided into many parts and reconstructed in a form of matrix. Then, the external load applied on structure in the field was considered to study the mechanical behaviors of tunnel. Based on the limited observed monitoring data in matrix and mechanical analysis results, a double-driven model was developed to obtain the full-faced information, in which the data-driven model was the dominant one and the mechanical constraint was the secondary one. To verify the presented spatial deduction model, cross-test was conducted through assuming partial monitoring data are unknown and regarding them as testing points. The well agreement between deduction results with actual monitoring results means the proposed model is reasonable. Therefore, it was employed to deduct both the current and historical performance of tunnel full face, which is crucial to prevent structural disasters.

cs.LG↗

arXiv2021

LightXML: Transformer with Dynamic Negative Sampling for High-Performance Extreme Multi-label Text Classification

Extreme Multi-label text Classification (XMC) is a task of finding the most relevant labels from a large label set. Nowadays deep learning-based methods have shown significant success in XMC. However, the existing methods (e.g., AttentionXML and X-Transformer etc) still suffer from 1) combining several models to train and predict for one dataset, and 2) sampling negative labels statically during the process of training label ranking model, which reduces both the efficiency and accuracy of the model. To address the above problems, we proposed LightXML, which adopts end-to-end training and dynamic negative labels sampling. In LightXML, we use generative cooperative networks to recall and rank labels, in which label recalling part generates negative and positive labels, and label ranking part distinguishes positive labels from these labels. Through these networks, negative labels are sampled dynamically during label ranking part training by feeding with the same text representation. Extensive experiments show that LightXML outperforms state-of-the-art methods in five extreme multi-label datasets with much smaller model size and lower computational complexity. In particular, on the Amazon dataset with 670K labels, LightXML can reduce the model size up to 72% compared to AttentionXML.

cs.CL↗

arXiv2020

Hybrid Micro/Macro Level Convolution for Heterogeneous Graph Learning

Heterogeneous graphs are pervasive in practical scenarios, where each graph consists of multiple types of nodes and edges. Representation learning on heterogeneous graphs aims to obtain low-dimensional node representations that could preserve both node attributes and relation information. However, most of the existing graph convolution approaches were designed for homogeneous graphs, and therefore cannot handle heterogeneous graphs. Some recent methods designed for heterogeneous graphs are also faced with several issues, including the insufficient utilization of heterogeneous properties, structural information loss, and lack of interpretability. In this paper, we propose HGConv, a novel Heterogeneous Graph Convolution approach, to learn comprehensive node representations on heterogeneous graphs with a hybrid micro/macro level convolutional operation. Different from existing methods, HGConv could perform convolutions on the intrinsic structure of heterogeneous graphs directly at both micro and macro levels: A micro-level convolution to learn the importance of nodes within the same relation, and a macro-level convolution to distinguish the subtle difference across different relations. The hybrid strategy enables HGConv to fully leverage heterogeneous information with proper interpretability. Moreover, a weighted residual connection is designed to aggregate both inherent attributes and neighbor information of the focal node adaptively. Extensive experiments on various tasks demonstrate not only the superiority of HGConv over existing methods, but also the intuitive interpretability of our approach for graph analysis.

cs.LG↗

arXiv2020

Coupled Layer-wise Graph Convolution for Transportation Demand Prediction

Graph Convolutional Network (GCN) has been widely applied in transportation demand prediction due to its excellent ability to capture non-Euclidean spatial dependence among station-level or regional transportation demands. However, in most of the existing research, the graph convolution was implemented on a heuristically generated adjacency matrix, which could neither reflect the real spatial relationships of stations accurately, nor capture the multi-level spatial dependence of demands adaptively. To cope with the above problems, this paper provides a novel graph convolutional network for transportation demand prediction. Firstly, a novel graph convolution architecture is proposed, which has different adjacency matrices in different layers and all the adjacency matrices are self-learned during the training process. Secondly, a layer-wise coupling mechanism is provided, which associates the upper-level adjacency matrix with the lower-level one. It also reduces the scale of parameters in our model. Lastly, a unitary network is constructed to give the final prediction result by integrating the hidden spatial states with gated recurrent unit, which could capture the multi-level spatial dependence and temporal dynamics simultaneously. Experiments have been conducted on two real-world datasets, NYC Citi Bike and NYC Taxi, and the results demonstrate the superiority of our model over the state-of-the-art ones.

cs.LG↗

arXiv2020

Defending Water Treatment Networks: Exploiting Spatio-temporal Effects for Cyber Attack Detection

While Water Treatment Networks (WTNs) are critical infrastructures for local communities and public health, WTNs are vulnerable to cyber attacks. Effective detection of attacks can defend WTNs against discharging contaminated water, denying access, destroying equipment, and causing public fear. While there are extensive studies in WTNs attack detection, they only exploit the data characteristics partially to detect cyber attacks. After preliminary exploring the sensing data of WTNs, we find that integrating spatio-temporal knowledge, representation learning, and detection algorithms can improve attack detection accuracy. To this end, we propose a structured anomaly detection framework to defend WTNs by modeling the spatio-temporal characteristics of cyber attacks in WTNs. In particular, we propose a spatio-temporal representation framework specially tailored to cyber attacks after separating the sensing data of WTNs into a sequence of time segments. This framework has two key components. The first component is a temporal embedding module to preserve temporal patterns within a time segment by projecting the time segment of a sensor into a temporal embedding vector. We then construct Spatio-Temporal Graphs (STGs), where a node is a sensor and an attribute is the temporal embedding vector of the sensor, to describe the state of the WTNs. The second component is a spatial embedding module, which learns the final fused embedding of the WTNs from STGs. In addition, we devise an improved one class-SVM model that utilizes a new designed pairwise kernel to detect cyber attacks. The devised pairwise kernel augments the distance between normal and attack patterns in the fused embedding space. Finally, we conducted extensive experimental evaluations with real-world data to demonstrate the effectiveness of our framework.

cs.CR↗

arXiv2020

Predicting Temporal Sets with Deep Neural Networks

Given a sequence of sets, where each set contains an arbitrary number of elements, the problem of temporal sets prediction aims to predict the elements in the subsequent set. In practice, temporal sets prediction is much more complex than predictive modelling of temporal events and time series, and is still an open problem. Many possible existing methods, if adapted for the problem of temporal sets prediction, usually follow a two-step strategy by first projecting temporal sets into latent representations and then learning a predictive model with the latent representations. The two-step approach often leads to information loss and unsatisfactory prediction performance. In this paper, we propose an integrated solution based on the deep neural networks for temporal sets prediction. A unique perspective of our approach is to learn element relationship by constructing set-level co-occurrence graph and then perform graph convolutions on the dynamic relationship graphs. Moreover, we design an attention-based module to adaptively learn the temporal dependency of elements and sets. Finally, we provide a gated updating mechanism to find the hidden shared patterns in different sequences and fuse both static and dynamic information to improve the prediction performance. Experiments on real-world data sets demonstrate that our approach can achieve competitive performances even with a portion of the training data and can outperform existing methods with a significant margin.

cs.LG↗

arXiv2019

Kepler-411: a four-planet system with an active host star

We present a detailed characterization of the Kepler-411 system (KOI 1781). This system was previously known to host two transiting planets: one with a period of 3 days ($R=2.4R_\oplus$; Kepler-411b) and one with a period of 7.8 days ($R=4.4R_\oplus$; Kepler-411c), as well as a transiting planetary candidate with a 58-day period ($R=3.3R_\oplus$; KOI 1781.03) from Kepler photometry. Here, we combine Kepler photometry data and new transit timing variation (TTV) measurements from all the Kepler quarters with previous adaptive-optics imaging results, and dynamical simulations, in order to constrain the properties of the Kepler-411 system. From our analysis, we obtain masses of 25.6$\pm$2.6$M_\oplus$ for Kepler-411b and 26.4$\pm$5.9 $M_\oplus$ for Kepler-411c, and we confirm the planetary nature of KOI 1781.03 with a mass of 15.2$\pm$5.1$M_\oplus$, hence the name Kepler-411d. Furthermore, by assuming near-coplanarity of the system (mutual inclination below $30^\circ$), we discover a nontransiting planet, Kepler-411e, with a mass of 10.8$\pm$1.1$M_\oplus$ on a 31.5-day orbit, which has a strong dynamical interaction with Kepler-411d. With densities of $1.71\pm0.39$~g\,cm$^{-3}$ and $2.32\pm0.83$~g\,cm$^{-3}$, both Kepler-411c and Kepler-411d belong to the group of planets with a massive core and a significant fraction of volatiles. Although Kepler-411b has a sub-Neptune size, it belongs to the group of rocky planets.

astro-ph.EP↗