arXiv subjects

Liang Wang

Publications and source records attributed to Liang Wang.

At least 595 records · Page 33Linked to original sources

arXiv2019

RL-Based User Association and Resource Allocation for Multi-UAV enabled MEC

In this paper, multi-unmanned aerial vehicle (UAV) enabled mobile edge computing (MEC), i.e., UAVE is studied, where several UAVs are deployed as flying MEC platform to provide computing resource to ground user equipments (UEs). Compared to the traditional fixed location MEC, UAV enabled MEC (i.e., UAVE) is particular useful in case of temporary events, emergency situations and on-demand services, due to its high flexibility, low cost and easy deployment features. However, operation of UAVE faces several challenges, two of which are how to achieve both 1) the association between multiple UEs and UAVs and 2) the resource allocation from UAVs to UEs, while minimizing the energy consumption for all the UEs. To address this, we formulate the above problem into a mixed integer nonlinear programming (MINLP), which is difficult to be solved in general, especially in the large-scale scenario. We then propose a Reinforcement Learning (RL)-based user Association and resource Allocation (RLAA) algorithm to tackle this problem efficiently and effectively. Numerical results show that the proposed RLAA can achieve the optimal performance with comparison to the exhaustive search in small scale, and have considerable performance gain over other typical algorithms in large-scale cases.

cs.NI↗

arXiv2019

Translational control of gene expression via interacting feedback loops

Translation is a key step in the synthesis of proteins. Accordingly, cells have evolved an intricate array of control mechanisms to regulate this process. By constructing a multi-component mathematical framework for translation we uncover how translation may be controlled via interacting feedback loops. Our results reveal that this interplay gives rise to a remarkable range of protein synthesis dynamics, including oscillations, step-change and bistability. This suggests that cells may have recourse to a much richer set of control mechanisms than was previously understood.

physics.bio-ph↗

arXiv2019

An Attention Enhanced Graph Convolutional LSTM Network for Skeleton-Based Action Recognition

Skeleton-based action recognition is an important task that requires the adequate understanding of movement characteristics of a human action from the given skeleton sequence. Recent studies have shown that exploring spatial and temporal features of the skeleton sequence is vital for this task. Nevertheless, how to effectively extract discriminative spatial and temporal features is still a challenging problem. In this paper, we propose a novel Attention Enhanced Graph Convolutional LSTM Network (AGC-LSTM) for human action recognition from skeleton data. The proposed AGC-LSTM can not only capture discriminative features in spatial configuration and temporal dynamics but also explore the co-occurrence relationship between spatial and temporal domains. We also present a temporal hierarchical architecture to increases temporal receptive fields of the top AGC-LSTM layer, which boosts the ability to learn the high-level semantic representation and significantly reduces the computation cost. Furthermore, to select discriminative spatial information, the attention mechanism is employed to enhance information of key joints in each AGC-LSTM layer. Experimental results on two datasets are provided: NTU RGB+D dataset and Northwestern-UCLA dataset. The comparison results demonstrate the effectiveness of our approach and show that our approach outperforms the state-of-the-art methods on both datasets.

cs.CV↗

arXiv2019

Forecasting model based on information-granulated GA-SVR and ARIMA for producer price index

The accuracy of predicting the Producer Price Index (PPI) plays an indispensable role in government economic work. However, it is difficult to forecast the PPI. In our research, we first propose an unprecedented hybrid model based on fuzzy information granulation that integrates the GA-SVR and ARIMA (Autoregressive Integrated Moving Average Model) models. The fuzzy-information-granulation-based GA-SVR-ARIMA hybrid model is intended to deal with the problem of imprecision in PPI estimation. The proposed model adopts the fuzzy information-granulation algorithm to pre-classification-process monthly training samples of the PPI, and produced three different sequences of fuzzy information granules, whose Support Vector Regression (SVR) machine forecast models were separately established for their Genetic Algorithm (GA) optimization parameters. Finally, the residual errors of the GA-SVR model were rectified through ARIMA modeling, and the PPI estimate was reached. Research shows that the PPI value predicted by this hybrid model is more accurate than that predicted by other models, including ARIMA, GRNN, and GA-SVR, following several comparative experiments. Research also indicates the precision and validation of the PPI prediction of the hybrid model and demonstrates that the model has consistent ability to leverage the forecasting advantage of GA-SVR in non-linear space and of ARIMA in linear space.

stat.AP↗

arXiv2019

Iteratively Learning Embeddings and Rules for Knowledge Graph Reasoning

Reasoning is essential for the development of large knowledge graphs, especially for completion, which aims to infer new triples based on existing ones. Both rules and embeddings can be used for knowledge graph reasoning and they have their own advantages and difficulties. Rule-based reasoning is accurate and explainable but rule learning with searching over the graph always suffers from efficiency due to huge search space. Embedding-based reasoning is more scalable and efficient as the reasoning is conducted via computation between embeddings, but it has difficulty learning good representations for sparse entities because a good embedding relies heavily on data richness. Based on this observation, in this paper we explore how embedding and rule learning can be combined together and complement each other's difficulties with their advantages. We propose a novel framework IterE iteratively learning embeddings and rules, in which rules are learned from embeddings with proper pruning strategy and embeddings are learned from existing triples and new triples inferred by rules. Evaluations on embedding qualities of IterE show that rules help improve the quality of sparse entity embeddings and their link prediction results. We also evaluate the efficiency of rule learning and quality of rules from IterE compared with AMIE+, showing that IterE is capable of generating high quality rules more efficiently. Experiments show that iteratively learning embeddings and rules benefit each other during learning and prediction.

cs.AI↗

arXiv2019

Angular momentum of $z\sim 1.5$ galaxies and their local analogues with adaptive optics

We present stellar specific angular momentum $j_*$ measurements of two $z\sim 1.5$ galaxies in the KGES sample and 12 DYNAMO $z\sim 0.1$ analogues of high-redshift galaxies. We combine natural seeing integral field spectroscopic data to trace line emission out to high multiples of effective radius $r_e$, with adaptive optics assisted Keck/OSIRIS observations to trace the rapid rise in rotation curve in the inner regions. Our spaxel-wise integration method gives results that are on average within measurement uncertainty of the traditional rotation curve model method. At $z\sim 0$, combining GMOS and OSIRIS datasets improves the measurement uncertainty in $j_*$ from 13\% (GMOS only) or 16\% (OSIRIS only) to 10\%. At $z\sim 1.5$, systematics allow for at best 20\% uncertainty on $j_*$. DYNAMO analogues of high-$z$ galaxies have low $j_*$ for their stellar mass $M_*$, and low bulge-to-total light ratio $β$ for their $j_*/M_*$. The high-$z$ galaxy COSMOS 127977 has $j_*/M_*$ consistent with normal local disk galaxies, while UDS 78317 is consistent with local analogues. However, our high-resolution OSIRIS data reveal that UDS 78317 may be a merging system. We report a relationship between distance to the $β-j_*/M_*$ plane and the ratio of velocity dispersion to rotational velocity $σ/v_{max}$, where galaxies that deviate more from the plane are more dispersion-dominated due to turbulence. Much of the scatter in $M_*-j_*$ that is not explained by variations in the bulge-to-total ratio or evolution with redshift may be driven by increased turbulence due to star formation, or by treating mergers as rotating disks.

astro-ph.GA↗

arXiv2019

Size-scaling of clump instabilities in turbulent, feedback regulated disks

We explore the scaling between the size of star-forming clumps and rotational support in massively star-forming galactic disks. The analysis relies on simulations of a clumpy galaxy at $z=2$ and the observed DYNAMO sample of rare clumpy analogs at $z\approx0.1$ to test a predictive clump size scaling proposed by \citet{Fisher2017ApJ...839L...5F} in the context of the Violent Disk Instability (VDI) theory. We here determine the clump sizes using a recently presented 2-point estimator, which is robust against resolution/noise effects, hierarchical clump substructure, clump-clump overlap and other galactic substructure. After verifying Fisher's clump scaling relation for the DYNAMO observations, we explore whether this relation remains characteristic of the VDI theory, even if realistic physical processes, such as local asymetries and stellar feedback, are included in the model. To this end, we rely on hydrodynamic zoom-simulations of a Milky Way-mass galaxy with four different feedback prescriptions. We find that, during its marginally stable epoch at $z=2$, this mock galaxy falls on the clump scaling relation, although its position on this relation depends on the feedback model. This finding implies that Toomre-like stability considerations approximately apply to large ($\sim\rm kpc$) instabilities in marginally stable turbulent disks, irrespective of the feedback model, but also emphasizes that the global clump distribution of a turbulent disk depends strongly on feedback.

astro-ph.GA↗

arXiv2019

Dressing as a Whole: Outfit Compatibility Learning Based on Node-wise Graph Neural Networks

With the rapid development of fashion market, the customers' demands of customers for fashion recommendation are rising. In this paper, we aim to investigate a practical problem of fashion recommendation by answering the question "which item should we select to match with the given fashion items and form a compatible outfit". The key to this problem is to estimate the outfit compatibility. Previous works which focus on the compatibility of two items or represent an outfit as a sequence fail to make full use of the complex relations among items in an outfit. To remedy this, we propose to represent an outfit as a graph. In particular, we construct a Fashion Graph, where each node represents a category and each edge represents interaction between two categories. Accordingly, each outfit can be represented as a subgraph by putting items into their corresponding category nodes. To infer the outfit compatibility from such a graph, we propose Node-wise Graph Neural Networks (NGNN) which can better model node interactions and learn better node representations. In NGNN, the node interaction on each edge is different, which is determined by parameters correlated to the two connected nodes. An attention mechanism is utilized to calculate the outfit compatibility score with learned node representations. NGNN can not only be used to model outfit compatibility from visual or textual modality but also from multiple modalities. We conduct experiments on two tasks: (1) Fill-in-the-blank: suggesting an item that matches with existing components of outfit; (2) Compatibility prediction: predicting the compatibility scores of given outfits. Experimental results demonstrate the great superiority of our proposed method over others.

cs.IR↗

arXiv2019

Session-based Recommendation with Graph Neural Networks

The problem of session-based recommendation aims to predict user actions based on anonymous sessions. Previous methods model a session as a sequence and estimate user representations besides item representations to make recommendations. Though achieved promising results, they are insufficient to obtain accurate user vectors in sessions and neglect complex transitions of items. To obtain accurate item embedding and take complex transitions of items into account, we propose a novel method, i.e. Session-based Recommendation with Graph Neural Networks, SR-GNN for brevity. In the proposed method, session sequences are modeled as graph-structured data. Based on the session graph, GNN can capture complex transitions of items, which are difficult to be revealed by previous conventional sequential methods. Each session is then represented as the composition of the global preference and the current interest of that session using an attention network. Extensive experiments conducted on two real datasets show that SR-GNN evidently outperforms the state-of-the-art session-based recommendation methods consistently.

cs.IR↗

arXiv2019

Simultaneous Subspace Clustering and Cluster Number Estimating based on Triplet Relationship

In this paper we propose a unified framework to simultaneously discover the number of clusters and group the data points into them using subspace clustering. Real data distributed in a high-dimensional space can be disentangled into a union of low-dimensional subspaces, which can benefit various applications. To explore such intrinsic structure, state-of-the-art subspace clustering approaches often optimize a self-representation problem among all samples, to construct a pairwise affinity graph for spectral clustering. However, a graph with pairwise similarities lacks robustness for segmentation, especially for samples which lie on the intersection of two subspaces. To address this problem, we design a hyper-correlation based data structure termed as the \textit{triplet relationship}, which reveals high relevance and local compactness among three samples. The triplet relationship can be derived from the self-representation matrix, and be utilized to iteratively assign the data points to clusters. Three samples in each triplet are encouraged to be highly correlated and are considered as a meta-element during clustering, which show more robustness than pairwise relationships when segmenting two densely distributed subspaces. Based on the triplet relationship, we propose a unified optimizing scheme to automatically calculate clustering assignments. Specifically, we optimize a model selection reward and a fusion reward by simultaneously maximizing the similarity of triplets from different clusters while minimizing the correlation of triplets from same cluster. The proposed algorithm also automatically reveals the number of clusters and fuses groups to avoid over-segmentation. Extensive experimental results on both synthetic and real-world datasets validate the effectiveness and robustness of the proposed method.

cs.CV↗

arXiv2018

Hierarchical Macro Strategy Model for MOBA Game AI

The next challenge of game AI lies in Real Time Strategy (RTS) games. RTS games provide partially observable gaming environments, where agents interact with one another in an action space much larger than that of GO. Mastering RTS games requires both strong macro strategies and delicate micro level execution. Recently, great progress has been made in micro level execution, while complete solutions for macro strategies are still lacking. In this paper, we propose a novel learning-based Hierarchical Macro Strategy model for mastering MOBA games, a sub-genre of RTS games. Trained by the Hierarchical Macro Strategy model, agents explicitly make macro strategy decisions and further guide their micro level execution. Moreover, each of the agents makes independent strategy decisions, while simultaneously communicating with the allies through leveraging a novel imitated cross-agent communication mechanism. We perform comprehensive evaluations on a popular 5v5 Multiplayer Online Battle Arena (MOBA) game. Our 5-AI team achieves a 48% winning rate against human player teams which are ranked top 1% in the player ranking system.

cs.MA↗

arXiv2018

Skeleton-Based Action Recognition with Spatial Reasoning and Temporal Stack Learning

Skeleton-based action recognition has made great progress recently, but many problems still remain unsolved. For example, most of the previous methods model the representations of skeleton sequences without abundant spatial structure information and detailed temporal dynamics features. In this paper, we propose a novel model with spatial reasoning and temporal stack learning (SR-TSL) for skeleton based action recognition, which consists of a spatial reasoning network (SRN) and a temporal stack learning network (TSLN). The SRN can capture the high-level spatial structural information within each frame by a residual graph neural network, while the TSLN can model the detailed temporal dynamics of skeleton sequences by a composition of multiple skip-clip LSTMs. During training, we propose a clip-based incremental loss to optimize the model. We perform extensive experiments on the SYSU 3D Human-Object Interaction dataset and NTU RGB+D dataset and verify the effectiveness of each network of our model. The comparison results illustrate that our approach achieves much better results than state-of-the-art methods.

cs.CV↗

arXiv2018

MV-RNN: A Multi-View Recurrent Neural Network for Sequential Recommendation

Sequential recommendation is a fundamental task for network applications, and it usually suffers from the item cold start problem due to the insufficiency of user feedbacks. There are currently three kinds of popular approaches which are respectively based on matrix factorization (MF) of collaborative filtering, Markov chain (MC), and recurrent neural network (RNN). Although widely used, they have some limitations. MF based methods could not capture dynamic user's interest. The strong Markov assumption greatly limits the performance of MC based methods. RNN based methods are still in the early stage of incorporating additional information. Based on these basic models, many methods with additional information only validate incorporating one modality in a separate way. In this work, to make the sequential recommendation and deal with the item cold start problem, we propose a Multi-View Recurrent Neural Network (MV-RNN}) model. Given the latent feature, MV-RNN can alleviate the item cold start problem by incorporating visual and textual information. First, At the input of MV-RNN, three different combinations of multi-view features are studied, like concatenation, fusion by addition and fusion by reconstructing the original multi-modal data. MV-RNN applies the recurrent structure to dynamically capture the user's interest. Second, we design a separate structure and a united structure on the hidden state of MV-RNN to explore a more effective way to handle multi-view features. Experiments on two real-world datasets show that MV-RNN can effectively generate the personalized ranking list, tackle the missing modalities problem and significantly alleviate the item cold start problem.

cs.IR↗

arXiv2018

Angular momentum evolution of bulge stars in disc galaxies in NIHAO

We study the origin of bulge stars and their angular momentum (AM) evolution in 10 spiral galaxies with baryonic masses above $10^{10}$M$_\odot$ in the NIHAO galaxy formation simulations. The simulated galaxies are in good agreement with observations of the relation between specific AM and mass of the baryonic component and the stellar bulge-to-total ratio ($B/T$). We divide the star particles at $z=0$ into disc and bulge components using a hybrid photometric/kinematic decomposition method that identifies all central mass above an exponential disc profile as the `bulge'. By tracking the bulge star particles back in time, we find that on average 95\% of the bulge stars formed {\it in situ}, 3\% formed {\it ex situ} in satellites of the same halo, and only 2\% formed {\it ex situ} in external galaxies. The evolution of the AM distribution of the bulge stars paints an interesting picture: the higher the final $B/T$ ratio, the more the specific AM remains preserved during the bulge formation. In all cases, bulge stars migrate significantly towards the central region, reducing their average galactocentric radius by roughly a factor 2, independently of the final $B/T$ value. However, in the higher $B/T$ ($\gtrsim0.2$) objects, the velocity of the bulge stars increases and the AM of the bulge is almost conserved, whereas at lower $B/T$ values, the velocity of the bulge stars decreases and the AM of bulge reduces. The correlation between the evolution of the AM and $B/T$ suggests that bulge and disc formation are closely linked and cannot be treated as independent processes.

astro-ph.GA↗

arXiv2018

Dynamic equilibrium sets atomic content of galaxies across cosmic time

We analyze 88 independent high-resolution cosmological zoom-in simulations of disk galaxies in the NIHAO simulations suite to explore the connection between the atomic gas fraction and angular momentum of baryons throughout cosmic time. The study is motivated by the analytic model of \citet{obreschkow16}, which predicts a relation between the atomic gas fraction $f_{\rm atm}$ and the global atomic stability parameter $q \equiv jσ/ (GM)$, where $M$ and $j$ are the mass and specific angular momentum of the galaxy (stars+cold gas) and $σ$ is the velocity dispersion of the atomic gas. We show that the simulated galaxies follow this relation from their formation ($z\simeq4$) to present within $\sim 0.5$ dex. To explain this behavior, we explore the evolution of the local Toomre stability and find that $90\%$--$100\%$ of the atomic gas in all simulated galaxies is stable at any time. In other words, throughout the entire epoch of peak star formation until today, the timescale for accretion is longer than the timescale to reach equilibrium, thus resulting in a quasi-static equilibrium of atomic gas at any time. Hence, the evolution of $f_{\rm atm}$ depends on the complex hierarchical growth history primarily via the evolution of $q$. An exception are galaxies subject to strong environmental effects.

astro-ph.GA↗

arXiv2018

Role of the Plasmoid Instability in Magnetohydrodynamic Turbulence

The plasmoid instability in evolving current sheets has been widely studied due to its effects on the disruption of current sheets, the formation of plasmoids, and the resultant fast magnetic reconnection. In this Letter, we study the role of the plasmoid instability in two-dimensional magnetohydrodynamic (MHD) turbulence by means of high-resolution direct numerical simulations. At sufficiently large magnetic Reynolds number ($R_m=10^6$), the combined effects of dynamic alignment and turbulent intermittency lead to a copious formation of plasmoids in a multitude of intense current sheets. The disruption of current sheet structures facilitates the energy cascade towards small scales, leading to the breaking and steepening of the energy spectrum. In the plasmoid-mediated regime, the energy spectrum displays a scaling that is close to the spectral index $-2.2$ as proposed by recent analytic theories. We also demonstrate that the scale-dependent dynamic alignment exists in 2D MHD turbulence and the corresponding slope of the alignment angle is close to 0.25.

physics.plasm-ph↗

arXiv2018

Owl: A General-Purpose Numerical Library in OCaml

Owl is a new numerical library developed in the OCaml language. It focuses on providing a comprehensive set of high-level numerical functions so that developers can quickly build up data analytical applications. In this abstract, we will present Owl's design, core components, and its key functionality.

cs.MS↗

arXiv2018

Multi-Perspective Context Aggregation for Semi-supervised Cloze-style Reading Comprehension

Cloze-style reading comprehension has been a popular task for measuring the progress of natural language understanding in recent years. In this paper, we design a novel multi-perspective framework, which can be seen as the joint training of heterogeneous experts and aggregate context information from different perspectives. Each perspective is modeled by a simple aggregation module. The outputs of multiple aggregation modules are fed into a one-timestep pointer network to get the final answer. At the same time, to tackle the problem of insufficient labeled data, we propose an efficient sampling mechanism to automatically generate more training examples by matching the distribution of candidates between labeled and unlabeled data. We conduct our experiments on a recently released cloze-test dataset CLOTH (Xie et al., 2017), which consists of nearly 100k questions designed by professional teachers. Results show that our method achieves new state-of-the-art performance over previous strong baselines.

cs.CL↗