arXiv subjects

Nikolaos Pappas

Publications and source records attributed to Nikolaos Pappas.

At least 199 records · Page 11Linked to original sources

arXiv2019

GILE: A Generalized Input-Label Embedding for Text Classification

Neural text classification models typically treat output labels as categorical variables which lack description and semantics. This forces their parametrization to be dependent on the label set size, and, hence, they are unable to scale to large label sets and generalize to unseen ones. Existing joint input-label text models overcome these issues by exploiting label descriptions, but they are unable to capture complex label relationships, have rigid parametrization, and their gains on unseen labels happen often at the expense of weak performance on the labels seen during training. In this paper, we propose a new input-label model which generalizes over previous such models, addresses their limitations and does not compromise performance on seen labels. The model consists of a joint non-linear input-label embedding with controllable capacity and a joint-space-dependent classification unit which is trained with cross-entropy loss to optimize classification performance. We evaluate models on full-resource and low- or zero-resource text classification of multilingual news and biomedical text with a large label set. Our model outperforms monolingual and multilingual models which do not leverage label semantics and previous joint input-label space models in both scenarios.

cs.CL↗

arXiv2018

Age of Information and Throughput in a Shared Access Network with Heterogeneous Traffic

We consider a cognitive shared access scheme consisting of a high priority primary node and a low priority network with $N$ secondary nodes accessing the spectrum. Assuming bursty traffic at the primary node, saturated queues at the secondary nodes, and multipacket reception capabilities at the receivers, we derive analytical expressions of the time average age of information of the primary node and the throughput of the secondary nodes. We formulate two optimization problems, the first aiming to minimize the time average age of information of the primary node subject to an aggregate secondary throughput requirement. The second problem aims to maximize the aggregate secondary throughput of the network subject to a maximum time average staleness constraint. Our results provide guidelines for the design of a multiple access system with multipacket reception capabilities that fulfils both timeliness and throughput requirements.

cs.NI↗

arXiv2018

Maximum Throughput Scheduling for Multi-connectivity in Millimeter-Wave Networks

Multi-connectivity is emerging as promising solution to provide reliable communications and seamless connectivity at the millimeter-wave frequency range. Due to the obstacles that cause frequent interruptions at such high frequency range, connectivity to multiple cells can drastically increase the network performance in terms of throughput and reliability by coordination among the network elements. In this paper, we propose an algorithm for the link scheduling optimization that maximizes the network throughput for multi-connectivity in millimeter-wave cellular networks. The considered approach exploits a centralized architecture, fast link switching, proactive context preparation and data forwarding between millimeter-wave access points and the users. The proposed algorithm is able to numerically approach the global optimum and to quantify the potential gain of multi-connectivity in millimeter-wave cellular networks.

cs.NI↗

arXiv2018

Anti-Gravitating Brane-World Solutions for a de Sitter Brane in Scalar-Tensor Gravity

In the context of a five-dimensional theory with a scalar field non-minimally-coupled to gravity, we look for solutions that describe novel black-string or maximally-symmetric solutions in the bulk. The brane line-element is found to describe a Schwarzschild-(Anti)-de Sitter spacetime, and, here, we choose to study solutions with a positive four-dimensional cosmological constant. We consider two different forms of the coupling function of the scalar field to the bulk scalar curvature, a linear and a quadratic one. In the linear case, we find solutions where the theory, close to our brane, mimics an ordinary gravitational theory with a minimally-coupled scalar field giving rise to an exponentially decreasing warp factor in the absence of a negative bulk cosmological constant. The solution is characterised by the presence of a normal gravity regime around our brane and an anti-gravitating regime away from it. In the quadratic case, there is no normal-gravity regime at all, however, scalar field and energy-momentum tensor components are well-defined and an exponentially decreasing warp factor emerges again. We demonstrate that, in the context of this theory, the emergence of a positive cosmological constant on our brane is always accompanied by an anti-gravitating regime in the five-dimensional bulk.

gr-qc↗

arXiv2018

Optimal Policies for Status Update Generation in a Wireless System with Heterogeneous Traffic

A large body of applications that involve monitoring, decision making, and forecasting require timely status updates for their efficient operation. Age of Information (AoI) is a newly proposed metric that effectively captures this requirement. Recent research on the subject has derived AoI optimal policies for the generation of status updates and AoI optimal packet queueing disciplines. Unlike previous research we focus on low-end devices that typically support monitoring applications in the context of the Internet of Things. We acknowledge that these devices host a diverse set of applications some of which are AoI sensitive while others are not. Furthermore, due to their limited computational resources they typically utilize a simple First-In First-Out (FIFO) queueing discipline. We consider the problem of optimally controlling the status update generation process for a system with a source-destination pair that communicates via a wireless link, whereby the source node is comprised of a FIFO queue and two applications, one that is AoI sensitive and one that is not. We formulate this problem as a dynamic programming problem and utilize the framework of Markov Decision Processes to derive optimal policies for the generation of status update packets. Due to the lack of comparable methods in the literature, we compare the derived optimal policies against baseline policies, such as the zero-wait policy, and investigate the performance of all policies for a variety of network configurations. Results indicate that existing status update policies fail to capture the trade-off between frequent generation of status updates and queueing delay and thus perform poorly.

cs.IT↗

arXiv2018

Throughput Analysis for Relay-Assisted Millimeter-Wave Wireless Networks

In this work, we analyze the throughput of random access multi-user relay-assisted millimeter-wave wireless networks, in which both the destination and the relay have multipacket reception capability. We consider a full-duplex network cooperative relay that stores the successfully received packets in a queue, for which we analyze the performance. Moreover, we study the effects on the network throughput of two different strategies, by which the source nodes transmit either a packet to both the destination and the relay in the same timeslot by using wider beams (broadcast approach) or to only one of these two by using narrower beams (fully directional approach). We consider the inter-beam interference at the receiver and show the optimal strategy with respect to several system parameters, e.g., positions and number of the nodes.

cs.NI↗

arXiv2018

Integrating Weakly Supervised Word Sense Disambiguation into Neural Machine Translation

This paper demonstrates that word sense disambiguation (WSD) can improve neural machine translation (NMT) by widening the source context considered when modeling the senses of potentially ambiguous words. We first introduce three adaptive clustering algorithms for WSD, based on k-means, Chinese restaurant processes, and random walks, which are then applied to large word contexts represented in a low-rank space and evaluated on SemEval shared-task data. We then learn word vectors jointly with sense vectors defined by our best WSD method, within a state-of-the-art NMT system. We show that the concatenation of these vectors, and the use of a sense selection mechanism based on the weighted average of sense vectors, outperforms several baselines including sense-aware ones. This is demonstrated by translation on five language pairs. The improvements are above one BLEU point over strong NMT baselines, +4% accuracy over all ambiguous nouns and verbs, or +20% when scored manually over several challenging words.

cs.CL↗

arXiv2018

Stable Throughput Region of the Two-User Interference Channel

We consider the two-user interference channel where two independent pairs communicate concurrently and investigate its stable throughput region. First, the stability region is characterized for the general case, i.e., without any specific consideration on the transmission and reception structures. Second, we explore two different interference harnessing strategies at the receiver: treating interference as noise and successive interference cancellation. Furthermore, we provide conditions for the convexity of the stability region and for which a certain receiver strategy leads to broader stability region. The impact of multiple transmit antennas on the stability region is briefly discussed. Finally, we study the effect of random access on the stability region of the two-user interference channel.

cs.IT↗

arXiv2018

Document-Level Neural Machine Translation with Hierarchical Attention Networks

Neural Machine Translation (NMT) can be improved by including document-level contextual information. For this purpose, we propose a hierarchical attention model to capture the context in a structured and dynamic manner. The model is integrated in the original NMT architecture as another level of abstraction, conditioning on the NMT model's own previous hidden states. Experiments show that hierarchical attention significantly improves the BLEU score over a strong NMT baseline with the state-of-the-art in context-aware methods, and that both the encoder and decoder benefit from context in complementary ways.

cs.CL↗

arXiv2018

Self-Attentive Residual Decoder for Neural Machine Translation

Neural sequence-to-sequence networks with attention have achieved remarkable performance for machine translation. One of the reasons for their effectiveness is their ability to capture relevant source-side contextual information at each time-step prediction through an attention mechanism. However, the target-side context is solely based on the sequence model which, in practice, is prone to a recency bias and lacks the ability to capture effectively non-sequential dependencies among words. To address this limitation, we propose a target-side-attentive residual recurrent network for decoding, where attention over previous words contributes directly to the prediction of the next word. The residual learning facilitates the flow of information from the distant past and is able to emphasize any of the previously translated words, hence it gains access to a wider context. The proposed model outperforms a neural MT baseline as well as a memory and self-attention network on three language pairs. The analysis of the attention learned by the decoder confirms that it emphasizes a wider context, and that it captures syntactic-like structures.

cs.CL↗

arXiv2018

Dynamic Power Control for Packets with Deadlines

Wireless devices need to adapt their transmission power according to the fluctuating wireless channel in order to meet constraints of delay sensitive applications. In this paper, we consider delay sensitivity in the form of strict packet deadlines arriving in a transmission queue. Packets missing the deadline while in the queue are dropped from the system. We aim at minimizing the packet drop rate under average power constraints. We utilize tools from Lyapunov optimization to find an approximate solution by selecting power allocation. We evaluate the performance of the proposed algorithm and show that it achieves the same performance in terms of packet drop rate with that of the Earliest Deadline First (EDF) when the available power is sufficient. However, our algorithm outperforms EDF regarding the trade-off between packet drop rate and average power consumption.

cs.IT↗

arXiv2018

Performance Analysis of a System with Bursty Traffic and Adjustable Transmission Times

In this work, we consider the case where a source with bursty traffic can adjust the transmission duration in order to increase the reliability. The source is equipped with a queue in order to store the arriving packets. We model the system with a discrete time Markov Chain, and we characterize the performance in terms of service probability and average delay per packet. The accuracy of the theoretical results is validated through simulations. This work serves as an initial step in order to provide a framework for systems with arbitrary arrivals and variable transmission durations and it can be utilized for the derivation of the delay distribution and the delay violation probability.

cs.IT↗

arXiv2018

Beyond Weight Tying: Learning Joint Input-Output Embeddings for Neural Machine Translation

Tying the weights of the target word embeddings with the target word classifiers of neural machine translation models leads to faster training and often to better translation quality. Given the success of this parameter sharing, we investigate other forms of sharing in between no sharing and hard equality of parameters. In particular, we propose a structure-aware output layer which captures the semantic structure of the output space of words within a joint input-output embedding. The model is a generalized form of weight tying which shares parameters but allows learning a more flexible relationship with input word embeddings and allows the effective capacity of the output layer to be controlled. In addition, the model shares weights across output classifiers and translation contexts which allows it to better leverage prior knowledge about them. Our evaluation on English-to-Finnish and English-to-German datasets shows the effectiveness of the method against strong encoder-decoder baselines trained with or without weight tying.

cs.CL↗

arXiv2018

Resource Optimization with Flexible Numerology and Frame Structure for Heterogeneous Services

We explore the potential of optimizing resource allocation with flexible numerology in frequency domain and variable frame structure in time domain, in presence of services with different types of requirements. We analyze the computational complexity and propose a scalable optimization algorithm based on searching in both the primal space and dual space that are complementary to each other. Numerical results show significant advantages of adopting flexibility in both time and frequency domains for capacity enhancement and meeting the requirements of mission critical services.

cs.IT↗

arXiv2018

Toward a Tactile Internet Reference Architecture: Vision and Progress of the IEEE P1918.1 Standard

The term Tactile Internet broadly refers to a communication network which is capable of delivering control, touch, and sensing/actuation information in real-time. The Tactile Internet is currently a topic of interest for various standardization bodies. The emerging IEEE P1918.1 standards working group is focusing on defining a framework for the Tactile Internet. The main objective of this article is to present a reference architecture for the Tactile Internet based on the latest developments within the IEEE P1918.1 standard. The article provides an in-depth treatment of various architectural aspects including the key entities, the interfaces, the functional capabilities and the protocol stack. A case study has been presented as a manifestation of the architecture. Performance evaluation demonstrates the impact of functional capabilities and the underlying enablers on user-level utility pertaining to a generic Tactile Internet application.

cs.NI↗

arXiv2018

Modeling and Analysis of MPTCP Proxy-based LTE-WLAN Path Aggregation

Long Term Evolution (LTE)-Wireless Local Area Network (WLAN) Path Aggregation (LWPA) based on Multi-path Transmission Control Protocol (MPTCP) has been under standardization procedure as a promising and cost-efficient solution to boost Downlink (DL) data rate and handle the rapidly increasing data traffic. This paper aims at providing tractable analysis for the DL performance evaluation of large-scale LWPA networks with the help of tools from stochastic geometry. We consider a simple yet practical model to determine under which conditions a native WLAN Access Point (AP) will work under LWPA mode to help increasing the received data rate. Using stochastic spatial models for the distribution of WLAN APs and LTE Base Stations (BSs), we analyze the density of active LWPA-mode WiFi APs in the considered network model, which further leads to closed-form expressions on the DL data rate and area spectral efficiency (ASE) improvement. Our numerical results illustrate the impact of different network parameters on the performance of LWPA networks, which can be useful for further performance optimization.

cs.NI↗

arXiv2018

QoS Provisioning in Large Wireless Networks

Quality of service (QoS) provisioning in next-generation mobile communications systems entails a deep understanding of the delay performance. The delay in wireless networks is strongly affected by the traffic arrival process and the service process, which in turn depends on the medium access protocol and the signal-to-interference-plus-noise ratio (SINR) distribution. In this work, we characterize the conditional distribution of the service process given the point process in Poisson bipolar networks. We then provide an upper bound on the delay violation probability combining tools from stochastic network calculus and stochastic geometry. Furthermore, we analyze the delay performance under statistical queueing constraints using the effective capacity formulation. The impact of QoS requirements, network geometry and link distance on the delay performance is identified. Our results provide useful insights for guaranteeing stringent delay requirements in large wireless networks.

cs.IT↗

arXiv2017

Multilingual Hierarchical Attention Networks for Document Classification

Hierarchical attention networks have recently achieved remarkable performance for document classification in a given language. However, when multilingual document collections are considered, training such models separately for each language entails linear parameter growth and lack of cross-language transfer. Learning a single multilingual model with fewer parameters is therefore a challenging but potentially beneficial objective. To this end, we propose multilingual hierarchical attention networks for learning document structures, with shared encoders and/or shared attention mechanisms across languages, using multi-task learning and an aligned semantic space as input. We evaluate the proposed models on multilingual document classification with disjoint label sets, on a large dataset which we provide, with 600k news documents in 8 languages, and 5k labels. The multilingual models outperform monolingual ones in low-resource as well as full-resource settings, and use fewer parameters, thus confirming their computational efficiency and the utility of cross-language transfer.

cs.CL↗