scispace - formally typeset
Search or ask a question
Topic

Latency (engineering)

About: Latency (engineering) is a research topic. Over the lifetime, 3729 publications have been published within this topic receiving 39210 citations. The topic is also known as: lag.


Papers
More filters
Proceedings ArticleDOI
Jinyu Li1, Rui Zhao1, Eric Sun1, Jeremy H. M. Wong1, Amit Das1, Zhong Meng1, Yifan Gong1 
17 Mar 2020
TL;DR: In this paper, a two-head trajectory LSTM (cltLSTM) was proposed to achieve high accuracy and low latency for automatic speech recognition, where one head has zero latency and the other head has a small latency.
Abstract: While the community keeps promoting end-to-end models over conventional hybrid models, which usually are long short-term memory (LSTM) models trained with a cross entropy criterion followed by a sequence discriminative training criterion, we argue that such conventional hybrid models can still be significantly improved. In this paper, we detail our recent efforts to improve conventional hybrid LSTM acoustic models for high-accuracy and low-latency automatic speech recognition. To achieve high accuracy, we use a contextual layer trajectory LSTM (cltLSTM), which decouples the temporal modeling and target classification tasks, and incorporates future context frames to get more information for accurate acoustic modeling. We further improve the training strategy with sequence-level teacher-student learning. To obtain low latency, we design a two-head cltLSTM, in which one head has zero latency and the other head has a small latency, compared to an LSTM. When trained with Microsoft’s 65 thousand hours of anonymized training data and evaluated with test sets with 1.8 million words, the proposed two-head cltLSTM model with the proposed training strategy yields a 28.2% relative WER reduction over the conventional LSTM acoustic model, with a similar perceived latency.

19 citations

Journal ArticleDOI
TL;DR: This paper characterize the reliability of the transmission phase in the finite blocklength regime and investigate the extreme event of queue length violation in the computation phase by applying extreme value theory.
Abstract: In this paper, we consider a multi-access edge computing (MEC) network with multiple servers. Due to the low latency constraints, the wireless data transmission/offloading is carried by finite blocklength codes. We characterize the reliability of the transmission phase in the finite blocklength regime and investigate the extreme event of queue length violation in the computation phase by applying extreme value theory. Under the assumption of perfect channel state information (CSI), we follow the obtained characterizations and provide an optimal framework design including server selection and time allocation aiming to minimize the overall error probability. Moreover, when only the outdated CSI is available, a deep reinforcement learning based design is proposed applying the deep deterministic policy gradient method. Via simulations, we validate the convexity proven in our analytical model and show the performance advantage of proposed analytical solution and learning-based solution comparing to the benchmark for perfect CSI and outdated CSI, respectively.

19 citations

Journal ArticleDOI
01 Mar 2005
TL;DR: Preliminary benchmark results showing exciting performances similar or better than those found in high-end commercial network systems are discussed.
Abstract: Developed by the APE group, APENet is a new high speed, low latency, 3-dimensional interconnect architecture optimized for PC clusters running LQCD-like numerical applications. The hardware implementation is based on a single PCI-X 133MHz network interface card hosting six independent bi-directional channels with a peak bandwidth of 676 MB/s each direction. We discuss preliminary benchmark results showing exciting performances similar or better than those found in high-end commercial network systems.

19 citations

Report SeriesDOI
Hailliang Zhao, Wei Du1, Wei Liu1, Tao Lei1, Qiwang Lei1 
01 Oct 2018
TL;DR: Simulation results illustrate that the algorithms could improve the ratio of offloading computation tasks by more than 10% while the QoE is guaranteed, and an online algorithm, namely, the LODCO-Based Genetic Algorithm with Greedy Policy, will be proposed.
Abstract: The increasing complexity of intelligent services requires new paradigm to overcome the problems caused by resource-limited mobile devices. Mobile edge computing systemwith energy harvesting devices is such a promising technology. By offloading the computation tasks to the MEC servers, users could experience services with low latency. In addition, energy harvesting technology releases the tension between high energy consumption of intelligent services and capacity-constrained mobile device batteries. However, in multi-user and multi-server scenarios where mobile devices can move arbitrarily, computation offloading strategies are faced with new challenges because of resource competition and server selection. In this paper, we develop an intelligent computation offloadingstrategy. An online algorithm, which is based on Lyapunov Optimization-based Dynamic Computation Offloading algorithm, is proposed. By choosing the execution mode among local execution, offloading execution and task dropping for each mobile device, our algorithm can asymptotically obtain the optimal results for the whole system. The algorithm not only inherits every advantage from the LODCO Algorithm but also adaptsperfectly to themore complex environment.Simulation results illustrate that the algorithmcanimprove the ratio of offloading computation tasks by more than 10% while the QoE is guaranteed.

19 citations

Journal ArticleDOI
TL;DR: A novel non-orthogonal multiple access (NOMA) scheme with WPT under FBL is first proposed to improve reliability and reduce latency, and the proposed scheme outperforms traditional OMA methods.
Abstract: This letter considers an uplink short-packet ultra-reliable low latency communications (URLLC) system with finite blocklength (FBL). A quasi-static Rayleigh fading channel is adopted in this letter, and we employ wireless power transfer (WPT) to guarantee the data packets transmission of user equipment (UE). A novel non-orthogonal multiple access (NOMA) scheme with WPT under FBL is first proposed to improve reliability and reduce latency. Furthermore, we adopt outage probability (OP) as an important URLLC metric, and deduce two closed-form approximate bounds of system OP for finite and infinite battery capacities to characterize the relationship between reliability and latency. Simulation results verify that the derived bounds are tight with theoretical values, and the proposed scheme outperforms traditional OMA methods.

19 citations


Network Information
Related Topics (5)
Network packet
159.7K papers, 2.2M citations
92% related
Server
79.5K papers, 1.4M citations
91% related
Wireless
133.4K papers, 1.9M citations
90% related
Wireless sensor network
142K papers, 2.4M citations
90% related
Wireless network
122.5K papers, 2.1M citations
90% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
202210
2021692
2020481
2019389
2018366
2017227