arXiv subjects

Ming Ding

Publications and source records attributed to Ming Ding.

At least 199 records · Page 11Linked to original sources

arXiv2019

Towards Knowledge-Based Recommender Dialog System

In this paper, we propose a novel end-to-end framework called KBRD, which stands for Knowledge-Based Recommender Dialog System. It integrates the recommender system and the dialog generation system. The dialog system can enhance the performance of the recommendation system by introducing knowledge-grounded information about users' preferences, and the recommender system can improve that of the dialog generation system by providing recommendation-aware vocabulary bias. Experimental results demonstrate that our proposed model has significant advantages over the baselines in both the evaluation of dialog generation and recommendation. A series of analyses show that the two systems can bring mutual benefits to each other, and the introduced knowledge contributes to both their performances.

cs.CL↗

arXiv2019

Beam Allocation for Millimeter-Wave MIMO Tracking Systems

In this paper, we propose a new beam allocation strategy aiming to maximize the average successful tracking probability (ASTP) of time-varying millimeter-wave MIMO systems. In contrast to most existing works that employ one transmitting-receiving (Tx-Rx) beam pair once only in each training period, we investigate a more general framework, where the Tx-Rx beam pairs are allowed to be used repeatedly to improve the received signal powers in specific directions. In the case of orthogonal Tx-Rx beam pairs, a power-based estimator is employed to track the time-varying AoA and AoD of the channel, and the resulting training beam pair sequence design problem is formulated as an integer nonlinear programming (I-NLP) problem. By dividing the feasible region into a set of subregions, the formulated I-NLP is decomposed into a series of concave sub I-NLPs, which can be solved by recursively invoking a nonlinear branch-and-bound algorithm. To reduce the computational cost, we relax the integer constraints of each sub I-NLP and obtain a low-complexity solution via solving the Karush-Kuhn-Tucker conditions of their relaxed problems. For the case when the Tx-Rx beam pairs are overlapped in the angular space, we estimate the updated AoA and AoD via an orthogonal matching pursuit (OMP) algorithm. Moreover, since no explicit expression for the ASTP exists for the OMP-based estimator, we derive a closed-form lower bound of the ASTP, based on which a favorable beam pair allocation strategy can be obtained. Numerical results demonstrate the superiority of the proposed beam allocation strategy over existing benchmarks.

eess.SP↗

arXiv2019

Cognitive Knowledge Graph Reasoning for One-shot Relational Learning

Inferring new facts from existing knowledge graphs (KG) with explainable reasoning processes is a significant problem and has received much attention recently. However, few studies have focused on relation types unseen in the original KG, given only one or a few instances for training. To bridge this gap, we propose CogKR for one-shot KG reasoning. The one-shot relational learning problem is tackled through two modules: the summary module summarizes the underlying relationship of the given instances, based on which the reasoning module infers the correct answers. Motivated by the dual process theory in cognitive science, in the reasoning module, a cognitive graph is built by iteratively coordinating retrieval (System 1, collecting relevant evidence intuitively) and reasoning (System 2, conducting relational reasoning over collected information). The structural information offered by the cognitive graph enables our model to aggregate pieces of evidence from multiple reasoning paths and explain the reasoning process graphically. Experiments show that CogKR substantially outperforms previous state-of-the-art models on one-shot KG reasoning benchmarks, with relative improvements of 24.3%-29.7% on MRR. The source code is available at https://github.com/THUDM/CogKR.

cs.LG↗

arXiv2019

Cognitive Graph for Multi-Hop Reading Comprehension at Scale

We propose a new CogQA framework for multi-hop question answering in web-scale documents. Inspired by the dual process theory in cognitive science, the framework gradually builds a \textit{cognitive graph} in an iterative process by coordinating an implicit extraction module (System 1) and an explicit reasoning module (System 2). While giving accurate answers, our framework further provides explainable reasoning paths. Specifically, our implementation based on BERT and graph neural network efficiently handles millions of documents for multi-hop reasoning questions in the HotpotQA fullwiki dataset, achieving a winning joint $F_1$ score of 34.9 on the leaderboard, compared to 23.6 of the best competitor.

cs.CL↗

arXiv2019

Survey on UAV Cellular Communications: Practical Aspects, Standardization Advancements, Regulation, and Security Challenges

The rapid growth of consumer Unmanned Aerial Vehicles (UAVs) is creating promising new business opportunities for cellular operators. On the one hand, UAVs can be connected to cellular networks as new types of user equipment, therefore generating significant revenues for the operators that can guarantee their stringent service requirements. On the other hand, UAVs offer the unprecedented opportunity to realize UAV-mounted flying base stations that can dynamically reposition themselves to boost coverage, spectral efficiency, and user quality of experience. Indeed, the standardization bodies are currently exploring possibilities for serving commercial UAVs with cellular networks. Industries are beginning to trial early prototypes of flying base stations or user equipments, while academia is in full swing researching mathematical and algorithmic solutions to address interesting new problems arising from flying nodes in cellular networks. In this article, we provide a comprehensive survey of all of these developments promoting smooth integration of UAVs into cellular networks. Specifically, we survey (i) the types of consumer UAVs currently available off-the-shelf, (ii) the interference issues and potential solutions addressed by standardization bodies for serving aerial users with the existing terrestrial base stations, (iii) the challenges and opportunities for assisting cellular communications with UAV-based flying relays and base stations, (iv) the ongoing prototyping and test bed activities, (v) the new regulations being developed to manage the commercial use of UAVs, and (vi) the cyber-physical security of UAV-assisted cellular communications.

cs.NI↗

arXiv2019

Sampling Free TDOA Localization in Millimeter Wave Networks

Time difference of arrival (TDOA) is a widely used technique for localizing a radio transmitter from the difference in signal arrival times at multiple receivers. For TDOA to work, the individual receivers must estimate the respective signal arrival times precisely, which requires sampling the signal at least double the rate of its highest frequency content, commonly known as the Nyquist rate. Such sampling is less practical for the millimeter wave band comprising of frequencies in 30-300 GHz range. In this paper, we propose an energy detection architecture for accurately estimating the time of arrival from a single picosecond Gaussian pulse, which enables TDOA localization without sampling at the receiver. We derive the closed form expression of the estimated time of arrival and validate it via simulation. We demonstrate that the proposed sampling-free TDOA can localize millimeter wave transmitters as accurately as the conventional sampling-based TDOA.

eess.SP↗

arXiv2019

Differentially Private Release of High-Dimensional Datasets using the Gaussian Copula

We propose a generic mechanism to efficiently release differentially private synthetic versions of high-dimensional datasets with high utility. The core technique in our mechanism is the use of copulas. Specifically, we use the Gaussian copula to define dependencies of attributes in the input dataset, whose rows are modelled as samples from an unknown multivariate distribution, and then sample synthetic records through this copula. Despite the inherently numerical nature of Gaussian correlations we construct a method that is applicable to both numerical and categorical attributes alike. Our mechanism is efficient in that it only takes time proportional to the square of the number of attributes in the dataset. We propose a differentially private way of constructing the Gaussian copula without compromising computational efficiency. Through experiments on three real-world datasets, we show that we can obtain highly accurate answers to the set of all one-way marginal, and two-and three-way positive conjunction queries, with 99\% of the query answers having absolute (fractional) error rates between 0.01 to 3\%. Furthermore, for a majority of two-way and three-way queries, we outperform independent noise addition through the well-known Laplace mechanism. In terms of computational time we demonstrate that our mechanism can output synthetic datasets in around 6 minutes 47 seconds on average with an input dataset of about 200 binary attributes and more than 32,000 rows, and about 2 hours 30 mins to execute a much larger dataset of about 700 binary attributes and more than 5 million rows. To further demonstrate scalability, we ran the mechanism on larger (artificial) datasets with 1,000 and 2,000 binary attributes (and 5 million rows) obtaining synthetic outputs in approximately 6 and 19 hours, respectively.

cs.CR↗

arXiv2018

A Brief History on the Theoretical Analysis of Dense Small Cell Wireless Networks

This article provides dives into the fundamentals of dense and ultra-dense small cell wireless networks, discussing the reasons why dense and ultra-dense small cell networks are fundamentally different from sparse ones, and why the well-known linear scaling law of capacity with the base station (BS) density in the latter does not apply to the former. In more detail, we review the impact of the following factors on ultradense networks (UDNs), (i) closed-access operations and line-of-sight conditions, (ii) the near-field effect, (iii) the antenna height difference between small cell BSs and user equipments (UEs), and (iv) the surplus of idle-mode-enabled small cell BSs with respect to UEs. Combining all these network characteristics and features, we present a more realistic capacity scaling law for UDNs, which indicates (i) the existence of an optimum BS density to maximise the area spectral efficiency (ASE) for a given finite UE density, and (ii) the existence of an optimum density of UEs that can be simultaneously scheduled across the network to maximise the ASE for a given finite BS density.

cs.NI↗

arXiv2018

Semi-supervised Learning on Graphs with Generative Adversarial Nets

We investigate how generative adversarial nets (GANs) can help semi-supervised learning on graphs. We first provide insights on working principles of adversarial learning over graphs and then present GraphSGAN, a novel approach to semi-supervised learning on graphs. In GraphSGAN, generator and classifier networks play a novel competitive game. At equilibrium, generator generates fake samples in low-density areas between subgraphs. In order to discriminate fake samples from the real, classifier implicitly takes the density property of subgraph into consideration. An efficient adversarial learning algorithm has been developed to improve traditional normalized graph Laplacian regularization with a theoretical guarantee. Experimental results on several different genres of datasets show that the proposed GraphSGAN significantly outperforms several state-of-the-art methods. GraphSGAN can be also trained using mini-batch, thus enjoys the scalability advantage.

cs.SI↗

arXiv2018

Spectral Network Embedding: A Fast and Scalable Method via Sparsity

Network embedding aims to learn low-dimensional representations of nodes in a network, while the network structure and inherent properties are preserved. It has attracted tremendous attention recently due to significant progress in downstream network learning tasks, such as node classification, link prediction, and visualization. However, most existing network embedding methods suffer from the expensive computations due to the large volume of networks. In this paper, we propose a $10\times \sim 100\times$ faster network embedding method, called Progle, by elegantly utilizing the sparsity property of online networks and spectral analysis. In Progle, we first construct a \textit{sparse} proximity matrix and train the network embedding efficiently via sparse matrix decomposition. Then we introduce a network propagation pattern via spectral analysis to incorporate local and global structure information into the embedding. Besides, this model can be generalized to integrate network information into other insufficiently trained embeddings at speed. Benefiting from sparse spectral network embedding, our experiment on four different datasets shows that Progle outperforms or is comparable to state-of-the-art unsupervised comparison approaches---DeepWalk, LINE, node2vec, GraRep, and HOPE, regarding accuracy, while is $10\times$ faster than the fastest word2vec-based method. Finally, we validate the scalability of Progle both in real large-scale networks and multiple scales of synthetic networks.

cs.SI↗

arXiv2018

Privacy Preservation in Location-Based Services: A Novel Metric and Attack Model

Recent years have seen rising needs for location-based services in our everyday life. Aside from the many advantages provided by these services, they have caused serious concerns regarding the location privacy of users. An adversary such as an untrusted location-based server can monitor the queried locations by a user to infer critical information such as the user's home address, health conditions, shopping habits, etc. To address this issue, dummy-based algorithms have been developed to increase the anonymity of users, and thus, protecting their privacy. Unfortunately, the existing algorithms only consider a limited amount of side information known by an adversary which may face more serious challenges in practice. In this paper, we incorporate a new type of side information based on consecutive location changes of users and propose a new metric called transition-entropy to investigate the location privacy preservation, followed by two algorithms to improve the transition-entropy for a given dummy generation algorithm. Then, we develop an attack model based on the Viterbi algorithm which can significantly threaten the location privacy of the users. Next, in order to protect the users from Viterbi attack, we propose an algorithm called robust dummy generation (RDG) which can resist against the Viterbi attack while maintaining a high performance in terms of the privacy metrics introduced in the paper. All the algorithms are applied and analyzed on a real-life dataset.

cs.CR↗

arXiv2018

The Essential Guide to Realizing 5G-Connected UAVs with Massive MIMO

What will it take for drones -- and the whole associated ecosystem -- to take off? Arguably, infallible command and control (C&C) channels for safe and autonomous flying, and high-throughput links for multi-purpose live video streaming. And indeed, meeting these aspirations may entail a full cellular support, provided through 5G-and-beyond hardware and software upgrades by both mobile operators and manufacturers of these unmanned aerial vehicles (UAVs). In this article, we vouch for massive MIMO as the key building block to realize 5G-connected UAVs. Through the sheer evidence of 3GPP-compliant simulations, we demonstrate how massive MIMO can be enhanced by complementary network-based and UAV-based solutions, resulting in consistent UAV C&C support, large UAV uplink data rates, and harmonious coexistence with legacy ground users.

cs.IT↗

arXiv2018

Cluster multiplication theorem in the quantum cluster algebra of type $A_{2}^{(2)}$

The objective of the present paper is to prove cluster multiplication theorem in the quantum cluster algebra of type $A_{2}^{(2)}$. As corollaries, we obtain bar-invariant $\mathbb{Z}[q^{\pm\frac{1}{2}}]$-bases established in [6], and naturally deduce the positivity of the elements in these bases. One bar-invariant basis as the triangular basis of this quantum cluster algebra is also explicitly described.

math.QA↗

arXiv2018

Performance Analysis for Practical Unmanned Aerial Vehicle Networks with LoS/NLoS Transmissions

In this paper, we provide a performance analysis for practical unmanned aerial vehicle (UAV)-enabled networks. By considering both line-of-sight (LoS) and non-line-of-sight (NLoS) transmissions between aerial base stations (BSs) and ground users, the coverage probability and the area spectral efficiency (ASE) are derived. Considering that there is no consensus on the path loss model for studying UAVs in the literature, in this paper, three path loss models, i.e., high-altitude model, low-altitude model and ultra-low-altitude model, are investigated and compared. Moreover, the lower bound of the network performance is obtained assuming that UAVs are hovering randomly according to homogeneous Poisson point process (HPPP), while the upper bound is derived assuming that UAVs can instantaneously move to the positions directly overhead ground users. From our analytical and simulation results for a practical UAV height of 50 meters, we find that the network performance of the high-altitude model and the low-altitude model exhibit similar trends, while that of the ultra-low-altitude model deviates significantly from the above two models. In addition, the optimal density of UAVs to maximize the coverage probability performance has also been investigated.

cs.NI↗

arXiv2018

New Experimental Limits on Exotic Spin-Spin-Velocity-Dependent Interactions By Using SmCo$_5$ Spin Sources

We report the latest results of searching for possible new macro-scale spin-spin-velocity-dependent forces (SSVDFs) based on specially designed iron-shielded SmCo$_5$ (ISSC) spin sources and a spin exchange relaxation free (SERF) co-magnetometer. The ISSCs have high net electron spin densities of about $1.7\times 10^{21}$ cm$^{-3}$, which mean high detecting sensitivity; and low magnetic field leakage of about $\sim$mG level due to iron shielding, which means low detecting noise. With help from the ISSCs, the high sensitivity SERF co-magnetometer, and the similarity analysis method, new constraints on SSVDFs with forms of $V_{6+7}$, $V_8$, $V_{15}$, and $V_{16}$ have been obtained, which represent the tightest limits in force range of 5 cm -- 1 km to the best of our knowledge.

hep-ex↗

arXiv2018

Optimal Base Station Antenna Downtilt in Downlink Cellular Networks

From very recent studies, the area spectral efficiency (ASE) performance of downlink (DL) cellular networks will continuously decrease and finally to zero with the network densification in a fully loaded ultra-dense network (UDN) when the absolute height difference between a base station (BS) antenna and a user equipment (UE) antenna is larger than zero, which is referred as the ASE Crash. We revisit this issue by considering the impact of the BS antenna downtilt on the downlink network capacity. In general, there exists a height difference between a BS and a UE in practical networks. It is common to utilize antenna downtilt to adjust the direction of the vertical antenna pattern, and thus increase received signal power or reduce inter-cell interference power to improve network performance. This paper focuses on investigating the relationship between the base station antenna downtilt and the downlink network capacity in terms of the coverage probability and the ASE. The analytical results of the coverage probability and the ASE are derived, and we find that there exists an optimal antenna downtilt to achieve the maximal coverage probability for each base station density. Moreover, we derive numerically solvable expressions for the optimal antenna downtilt, which is a function of the base station density. Our theoretical and numerical results show that after applying the optimal antenna downtilt, the network performance can be improved significantly. Specifically, with the optimal antenna downtilt, the ASE crash can be delayed by nearly one order of magnitude in terms of the base station density.

cs.NI↗

arXiv2018

Promises and Caveats of Uplink IoT Ultra-Dense Networks

In this paper, by means of simulations, we evaluate the uplink (UL) performance of an Internet of Things (IoT) capable ultra-dense network (UDN) in terms of the coverage probability and the density of reliably working user equipments (UEs). From our study, we show the benefits and challenges that UL IoT UDNs will bring about in the future. In more detail, for a low-reliability criterion, such as achieving a UL signal-to-interference-plus-noise ratio (SINR) above 0 dB, the density of reliably working UEs grows quickly with the network densification, showing the potential of UL IoT UDNs. In contrast, for a high-reliability criterion, such as achieving a UL SINR above 10 dB, the density of reliably working UEs remains to be low in UDNs due to excessive inter-cell interference, which should be considered when operating UL IoT UDNs. Moreover, considering the existence of a non-zero antenna height difference between base stations (BSs) and UEs, the density of reliably working UEs could even decrease as we deploy more BSs. This calls for the usage of sophisticated interference management schemes and/or beam steering/shaping technologies in UL IoT UDNs.

cs.NI↗

arXiv2018

Dense Small Cell Networks: From Noise-Limited to Dense Interference-Limited

Considering both non-line-of-sight (NLoS) and line-of-sight (LoS) transmissions, the transitional behaviors from noise-limited regime to dense interference-limited regime have been investigated for the fifth generation (5G) small cell networks (SCNs). Besides, we identify four performance regimes based on base station (BS) density, i.e., (i) the noise-limited regime, (ii) the signal-dominated regime, (iii) the interference-dominated regime, and (iv) the interference-limited regime. To characterize the performance regime, we propose a unified framework analyzing the future 5G wireless networks over generalized shadowing/fading channels, in which the user association schemes based on the strongest instantaneous received power (SIRP) and the strongest average received power (SARP) can be studied, while NLoS/LoS transmissions and multi-slop path loss model are considered. Simulation results indicate that different factors, i.e., noise, desired signal, and interference, successively and separately dominate the network performance with the increase of BS density. Hence, our results shed new light on the design and management of SCNs in urban and rural areas with different BS deployment densities.

cs.NI↗