Distributed deep learning (DL) plays a critical role in many wireless Internet of Things (IoT) applications including remote camera deployment. This work addresses three practical challenges in cyber-deployment of distributed DL over band-limited channels. Specifically, many IoT systems consist of sensor nodes for raw data collection and encoding, and servers for learning and inference tasks. Adaptation of DL over band-limited network data links has only been scantly addressed. A second challenge is the need for pre-deployed encoders being compatible with flexible decoders that can be upgraded or retrained. The third challenge is the robustness against erroneous training labels. Addressing these three challenges, we develop a hierarchical learning strategy to improve image classification accuracy over band-limited links between sensor nodes and servers. Experimental results show that our hierarchically-trained models can improve link spectrum efficiency without performance loss, reduce storage and computational complexity, and achieve robustness against training label corruption.

Vertical sectorization introduces considerable gain to particular scenarios where user equipment (UE) are distributed in 3D domain, i.e. high-rise buildings, in terms of coverage and network capacity thanks to availability of active antenna systems. However, due to the huge varieties in distribution of UEs and different physical environments, presenting a comprehensive analytical framework is quite challenging. From this aspect, most available studies on vertical sectorization are limited to present only empirical results. In this paper, we introduce a novel methodology to forecast the performance of NB-IoT systems over urban scenarios. In particular, a logistic distribution-based analytical framework is exploited in order to calculate the group probabilities for each available UE. Based on these probabilities, we propose a scheduling framework with beamforming which improves physical resource block (PRB) utilization by over 50% compared to the case with no scheduling.

The past decade has witnessed the rising dominance of deep learning and artificial intelligence in a wide range of applications. In particular, the ocean of wireless smartphones and IoT devices continue to fuel the tremendous growth of edge/cloud-based machine learning (ML) systems, including image/speech recognition and classification. To overcome the infrastructural barrier of limited network bandwidth in cloud ML, existing solutions have mainly relied on traditional compression codecs such as JPEG that were historically engineered for human-end users instead of ML algorithms. Traditional codecs do not necessarily preserve features important to ML algorithms under limited bandwidth, leading to potentially inferior performance. This work investigates application-driven optimization of programmable commercial codec settings for networked learning tasks such as image classification. Based on the foundation of variational autoencoders (VAEs), we develop an end-to-end networked learning framework by jointly optimizing the codec and classifier without reconstructing images for a given data rate (bandwidth). Compared with the standard JPEG codec, the proposed VAE joint compression and classification framework achieves classification accuracy improvement by over 10% and 4%, respectively, for CIFAR-10 and ImageNet-1k data sets at data rate of 0.8 bpp. Our proposed VAE-based models show 65%–99% reductions in encoder size,$\times 1.5$–$\times 13.1$improvements in inference speed, and 25%–99% savings in power compared to baseline models. We further show that a simple decoder can reconstruct images with sufficient quality without compromising classification accuracy.

An increasing share of image and video content is analyzed by machines rather than viewed by humans, and therefore it becomes relevant to optimize codecs for such applications where the analysis is performed remotely. Unfortunately, conventional coding tools are challenging to specialize for machine tasks as they were originally designed for human perception. However, neural network based codecs can be jointly trained end-to-end with any convolutional neural network (CNN)-based task model. In this paper, we propose to study an end-to-end framework enabling efficient image compression for remote machine task analysis, using a chain composed of a compression module and a task algorithm that can be optimized end-to-end. We show that it is possible to significantly improve the task accuracy when fine-tuning jointly the codec and the task networks, especially at low bitrates. Depending on training or deployment constraints, selective fine-tuning can be applied only on the encoder, decoder or task network and still achieve rate-accuracy improvements over an off-the-shelf codec and task network. Our results also demonstrate the flexibility of end-to-end pipelines for practical applications.

JPEG2000 (j2k) is a highly popular format for image and video compression. It plays a major role in the rapidly growing applications of cloud based image classification. Considering limited network bandwidth, we propose an end-to-end deep learning framework to achieve faster and more accurate classification by directly training a deep CNN image classifier using the CDF 9/7 Discrete Wavelet Transformed (DWT) coefficients from j2k-compressed images without image reconstruction. We demonstrate additional computation savings by utilizing shallower CNN to achieve classification of good accuracy. Furthermore, we present DWT-centric augmentation transformations to achieve more accurate classification without added cost. Achieving faster and more accurate classification for j2k encoded images, the proposed solution is well suited for joint compression and cloud-based image and video classification over limited channel bandwidth.

Signal detection in colored noise with an unknown covariance matrix has a myriad of applications in diverse scientific/engineering fields. The test statistic is the largest generalized eigenvalue (l.g.e.) of the whitened sample covariance matrix, which is constructed via m-dimensional p signal-plusnoise samples and m-dimensional n noise-only samples. A finite dimensional characterization of this statistic under the alternative hypothesis has hitherto been an open problem. We answer this problem by deriving cumulative distribution function (c.d.f.) of this l.g.e. via the powerful orthogonal polynomial approach, exploiting the deformed Jacobi unitary ensemble (JUE). Two special cases and an asymptotic version of the c.d.f. are also derived. With this new c.d.f., we comprehensively analyze the receiver operating characteristics (ROC) of the detector. Importantly, when the noise-only covariance matrix is nearly rank deficient (i.e., m = n), we show that (a) when m and p increase such that m/p is fixed, at each fixed signal-to-noise ratio (SNR), there exists an optimal ROC profile. We also establish a tight approximation of it; and (b) asymptotically, reliable signal detection is always possible if SNR scales with m.

This paper investigates the classical statistical signal processing problem of detecting a signal in the presence of colored noise with an unknown covariance matrix. In particular, we consider a scenario wherem-dimensionalppossible signal-plus-noise samples andm-dimensionalnnoise-only samples are available at the detector. Then the presence of a signal can be detected using the largest generalized eigenvalue (l.g.e.) of the so called whitened sample covariance matrix. This amounts to statistically characterizing the maximum eigenvalue of the deformed Jacobi unitary ensemble (JUE). To do this, we employ the powerful orthogonal polynomial approach to determine a new finite dimensional expression for the cumulative distribution function (c.d.f.) of the l.g.e. of the deformed JUE. This new c.d.f. expression facilitates the further analysis of the receiver operating characteristics (ROC) of the detector. It turns out that, form=n, when m andpincrease such thatm/pis fixed, there exists an optimal ROC profile for each fixed signal-to-noise ratio (SNR). In this respect, we have established a tight approximation for the corresponding optimal ROC profile.

The performance of cloud based image classification depends critically on its allocated bandwidth. Traditional data compression methods can negatively impact classification accuracy under limited bandwidth. We investigate the design of bandwidth efficient quantization for image encoding and compression with minimum classification accuracy loss. This work develops a simple neural network framework for joint quantization and classification. The proposed 'QuanNet' can optimize the quantization intervals of JPEG2000 encoder to minimize the classification loss. We show that our quantizer optimization can achieve significant accuracy improvement for a given channel bandwidth. Similarly, significant bandwidth can be saved to achieve a desired accuracy for cloud based image classification.