Search

The development of self-interference (SI) cancelation technology makes full-duplex (FD) communication possible. Considering the quality of service (QoS) of flows in small cells densely deployed scenario with limited time slot (TS) resources, this paper introduces the FD communication into the concurrent scheduling problem of millimeter-wave (mmWave) wireless backhaul network. We propose a QoS-aware FD concurrent scheduling algorithm to maximize the number of flows with their QoS requirements satisfied. Based on the contention graph, the algorithm makes full use of the FD condition. Both residual self-interference (RSI) and multi-user interference (MUI) are considered. Besides, it also fully considers the QoS requirements of flows and ensures the flows can be transmitted at high rates. Extensive simulations at 60GHz demonstrate that with high SI cancelation level and appropriate contention threshold, the proposed FD algorithm can achieve superior performance in terms of the number of flows with their QoS requirements satisfied and the system throughput compared with other stateof-of-the-art schemes., Comment: 9 pages, 10 figures

In the scenario where small cells are densely deployed, the millimeter wave (mmWave) wireless backhaul network has been widely used. However, mmWave is easily blocked by obstacles, and how to forward the data of the blocked flows is still a significant challenge. To ensure backhauling capacity, the quality of service (QoS) requirements of flows should be satisfied. In this paper, we investigate the problem of optimal scheduling to maximize the number of flows satisfying their QoS requirements with relays exploited to overcome blockage. To achieve a practical solution, we propose a relay-assisted and QoS aware scheduling scheme for the backhaul networks, called RAQS. It consists of a relay selection algorithm and a transmission scheduling algorithm. The relay selection algorithm selects non-repeating relays with high link rates for the blocked flows, which helps to achieve the QoS requirements of flows as soon as possible. Then, according to the results of relay selection, the transmission scheduling algorithm exploits concurrent transmissions to satisfy the QoS requirements of flows as much as possible. Extensive simulations show RAQS can effectively overcome the blockage problem, and increase the number of completed flows and network throughput compared with other schemes. In particular, the impact of relay selection parameter is also investigated to further guide the relay selection., Comment: 11 pages, 10 figures

Alzheimer's Disease (AD) is a progressive neurodegenerative disease. Amnestic mild cognitive impairment (MCI) is a common first symptom before the conversion to clinical impairment where the individual becomes unable to perform activities of daily living independently. Although there is currently no treatment available, the earlier a conclusive diagnosis is made, the earlier the potential for interventions to delay or perhaps even prevent progression to full-blown AD. Neuroimaging scans acquired from MRI and metabolism images obtained by FDG-PET provide in-vivo view into the structure and function (glucose metabolism) of the living brain. It is hypothesized that combining different image modalities could better characterize the change of human brain and result in a more accuracy early diagnosis of AD. In this paper, we proposed a novel framework to discriminate normal control(NC) subjects from subjects with AD pathology (AD and NC, MCI subjects convert to AD in future). Our novel approach utilizing a multimodal and multiscale deep neural network was found to deliver a 85.68\% accuracy in the prediction of subjects within 3 years to conversion. Cross validation experiments proved that it has better discrimination ability compared with results in existing published literature., Comment: 12 pages, 4 figures, Alzheimer's disease, deep learning, multimodal, early diagnosis, multiscale

Monitoring the number of insect pests is a crucial component in pheromone-based pest management systems. In this paper, we propose an automatic detection pipeline based on deep learning for identifying and counting pests in images taken inside field traps. Applied to a commercial codling moth dataset, our method shows promising performance both qualitatively and quantitatively. Compared to previous attempts at pest detection, our approach uses no pest-specific engineering which enables it to adapt to other species and environments with minimal human effort. It is amenable to implementation on parallel hardware and therefore capable of deployment in settings where real-time performance is required., Comment: Preprints accepted by Computers and electronics in agriculture

Regularization is essential when training large neural networks. As deep neural networks can be mathematically interpreted as universal function approximators, they are effective at memorizing sampling noise in the training data. This results in poor generalization to unseen data. Therefore, it is no surprise that a new regularization technique, Dropout, was partially responsible for the now-ubiquitous winning entry to ImageNet 2012 by the University of Toronto. Currently, Dropout (and related methods such as DropConnect) are the most effective means of regularizing large neural networks. These amount to efficiently visiting a large number of related models at training time, while aggregating them to a single predictor at test time. The proposed FaMe model aims to apply a similar strategy, yet learns a factorization of each weight matrix such that the factors are robust to noise.

In this report, we describe a Theano-based AlexNet (Krizhevsky et al., 2012) implementation and its naive data parallelism on multiple GPUs. Our performance on 2 GPUs is comparable with the state-of-art Caffe library (Jia et al., 2014) run on 1 GPU. To the best of our knowledge, this is the first open-source Python-based AlexNet implementation to-date., Comment: ICLR 2015 workshop camera-ready version

The human visual system is able to recognize objects despite transformations that can drastically alter their appearance. To this end, much effort has been devoted to the invariance properties of recognition systems. Invariance can be engineered (e.g. convolutional nets), or learned from data explicitly (e.g. temporal coherence) or implicitly (e.g. by data augmentation). One idea that has not, to date, been explored is the integration of latent variables which permit a search over a learned space of transformations. Motivated by evidence that people mentally simulate transformations in space while comparing examples, so-called "mental rotation", we propose a transforming distance. Here, a trained relational model actively transforms pairs of examples so that they are maximally similar in some feature space yet respect the learned transformational constraints. We apply our method to nearest-neighbour problems on the Toronto Face Database and NORB.