Institution
Huawei
Company•Shenzhen, China•
About: Huawei is a company organization based out in Shenzhen, China. It is known for research contribution in the topics: Terminal (electronics) & Signal. The organization has 41417 authors who have published 44698 publications receiving 343496 citations. The organization is also known as: Huawei Technologies & Huawei Technologies Co., Ltd..
Papers published on a yearly basis
Papers
More filters
•
18 Sep 2002TL;DR: In this paper, an internet router is implemented as a network fabric of fabric routers and links, where the router receives data packets from trunk lines or other internet links and analyzes header information in the data packets to route the data packet to output internet links.
Abstract: An internet router is implemented as a network fabric of fabric routers and links. The internet router receives data packets from trunk lines or other internet links and analyzes header information in the data packets to route the data packets to output internet links. The line interface also analyzes the header to define a fabric path through the router fabric. The internet packets are broken into flits which are transferred through the router according to a wormhole routing protocol. Flits are stored in fabric routers at storage locations assigned to virtual channels corresponding to destination internet links. The virtual channels and links within the fabric define virtual networks in which congestion in one virtual network is substantially nonblocking to data flow through other virtual networks. Arbitration is performed at each fabric router to assign packets to virtual channels and to assign virtual channels to output fabric links. A virtual channel is enabled for possible assignment to an output fabric link upon receipt of an indication that an input buffer is available at the next fabric router of the path.
127 citations
••
13 May 2019TL;DR: A novel Feature Generation by Convolutional Neural Network (FGCNN) model with two components: Feature Generation and Deep Classifier, which significantly outperforms nine state-of-the-art models on three large-scale datasets.
Abstract: Click-Through Rate prediction is an important task in recommender systems, which aims to estimate the probability of a user to click on a given item. Recently, many deep models have been proposed to learn low-order and high-order feature interactions from original features. However, since useful interactions are always sparse, it is difficult for DNN to learn them effectively under a large number of parameters. In real scenarios, artificial features are able to improve the performance of deep models (such as Wide & Deep Learning), but feature engineering is expensive and requires domain knowledge, making it impractical in different scenarios. Therefore, it is necessary to augment feature space automatically. In this paper, We propose a novel Feature Generation by Convolutional Neural Network (FGCNN) model with two components: Feature Generation and Deep Classifier. Feature Generation leverages the strength of CNN to generate local patterns and recombine them to generate new features. Deep Classifier adopts the structure of IPNN to learn interactions from the augmented feature space. Experimental results on three large-scale datasets show that FGCNN significantly outperforms nine state-of-the-art models. Moreover, when applying some state-of-the-art models as Deep Classifier, better performance is always achieved, showing the great compatibility of our FGCNN model. This work explores a novel direction for CTR predictions: it is quite useful to reduce the learning difficulties of DNN by automatically identifying important features.
127 citations
••
01 Nov 2020TL;DR: This work proposes TernaryBERT, which ternarizes the weights in a fine-tuned BERT model, and leverages the knowledge distillation technique in the training process to reduce the accuracy degradation caused by the lower capacity of low bits.
Abstract: Transformer-based pre-training models like BERT have achieved remarkable performance in many natural language processing tasks. However, these models are both computation and memory expensive, hindering their deployment to resource-constrained devices. In this work, we propose TernaryBERT, which ternarizes the weights in a fine-tuned BERT model. Specifically, we use both approximation-based and loss-aware ternarization methods and empirically investigate the ternarization granularity of different parts of BERT. Moreover, to reduce the accuracy degradation caused by lower capacity of low bits, we leverage the knowledge distillation technique in the training process. Experiments on the GLUE benchmark and SQuAD show that our proposed TernaryBERT outperforms the other BERT quantization methods, and even achieves comparable performance as the full-precision model while being 14.9x smaller.
127 citations
••
01 Jan 2015TL;DR: An SDN‐based plastic architecture for 5G networks, designed to fulfill functional and performance requirements of new generation services and devices, and backward compatibility with legacy systems is guaranteed, and Control Plane and Data Plane are fully decoupled.
Abstract: In this paper, we describe an SDN-based plastic architecture for 5G networks, designed to fulfill functional and performance requirements of new generation services and devices. The 5G logical architecture is presented in detail, and key procedures for dynamic control plane instantiation, device attachment, and service request and mobility management are specified. Key feature of the proposed architecture is flexibility, needed to support efficiently a heterogeneous set of services, including Machine Type Communication, Vehicle to X and Internet of Things traffic. These applications are imposing challenging targets, in terms of end-to-end latency, dependability, reliability and scalability. Additionally, backward compatibility with legacy systems is guaranteed by the proposed solution, and Control Plane and Data Plane are fully decoupled. The three levels of unified signaling unify Access, Non-access and Management strata, and a clean-slate forwarding layer, designed according to the software defined networking principle, replaces tunneling protocols for carrier grade mobility. Copyright © 2014 John Wiley & Sons, Ltd.
127 citations
••
14 Jun 2020TL;DR: In this paper, Xu et al. used Gradually Vanishing Bridge (GVB) mechanism on both generator and discriminator to reduce the influence of residual domain-specific characteristics in domain-invariant representations.
Abstract: In unsupervised domain adaptation, rich domain-specific characteristics bring great challenge to learn domain-invariant representations. However, domain discrepancy is considered to be directly minimized in existing solutions, which is difficult to achieve in practice. Some methods alleviate the difficulty by explicitly modeling domain-invariant and domain-specific parts in the representations, but the adverse influence of the explicit construction lies in the residual domain-specific characteristics in the constructed domain-invariant representations. In this paper, we equip adversarial domain adaptation with Gradually Vanishing Bridge (GVB) mechanism on both generator and discriminator. On the generator, GVB could not only reduce the overall transfer difficulty, but also reduce the influence of the residual domain-specific characteristics in domain-invariant representations. On the discriminator, GVB contributes to enhance the discriminating ability, and balance the adversarial training process. Experiments on three challenging datasets show that our GVB methods outperform strong competitors, and cooperate well with other adversarial methods. The code is available at https://github.com/cuishuhao/GVB.
127 citations
Authors
Showing all 41483 results
Name | H-index | Papers | Citations |
---|---|---|---|
Yu Huang | 136 | 1492 | 89209 |
Xiaoou Tang | 132 | 553 | 94555 |
Xiaogang Wang | 128 | 452 | 73740 |
Shaobin Wang | 126 | 872 | 52463 |
Qiang Yang | 112 | 1117 | 71540 |
Wei Lu | 111 | 1973 | 61911 |
Xuemin Shen | 106 | 1221 | 44959 |
Li Chen | 105 | 1732 | 55996 |
Lajos Hanzo | 101 | 2040 | 54380 |
Luca Benini | 101 | 1453 | 47862 |
Lei Liu | 98 | 2041 | 51163 |
Tao Wang | 97 | 2720 | 55280 |
Mohamed-Slim Alouini | 96 | 1788 | 62290 |
Qi Tian | 96 | 1030 | 41010 |
Merouane Debbah | 96 | 652 | 41140 |