Search

Reduced voltage operation is an effective technique for substantial energy efficiency improvement in digital circuits. This brief introduces a simple approach for enabling reduced voltage operation of Deep Neural Network (DNN) accelerators by mere software modifications. Conventional approaches for enabling reduced voltage operation e.g., Timing Error Detection (TED) systems, incur significant development costs and overheads, while not being applicable to the off-the-shelf components. Contrary to those, the solution proposed in this paper relies on algorithm-based error detection, and hence, is implemented with low development costs, does not require any circuit modifications, and is even applicable to commodity devices. By showcasing the solution through experimenting on popular DNNs, i.e., LeNet and VGG16, on a GPU platform, we demonstrate 18% to 25% energy saving with no accuracy loss of the models and negligible throughput compromise (< 3.9%), considering the overheads from integration of the error detection schemes into the DNN. The integration of presented algorithmic solution into the design is simpler when compared conventional TED based techniques that require extensive circuit-level modifications, cell library characterizations or special support from the design tools.

Beamforming is a signal processing technique where an array of antenna elements can be steered to transmit and receive radio signals in a specific direction. The usage of millimeter wave (mmWave) frequencies and multiple input multiple output (MIMO) beamforming are considered as the key innovations of 5th Generation (5G) and beyond communication systems. The technique initially performs a beam alignment procedure, followed by data transfer in the aligned directions between the transmitter and the receiver. Traditionally, beam alignment involves periodical and exhaustive beam sweeping at both transmitter and the receiver, which is a slow process causing extra communication overhead with MIMO and massive MIMO radio units. In applications such as beam tracking, angular velocity, beam steering etc., the beam alignment procedure is optimized by estimating the beam directions using first order polynomial approximations. Recent learning-based SOTA strategies for fast mmWave beam alignment also require exploration over exhaustive beam pairs during the training procedure, causing overhead to learning strategies for higher antenna configurations. In this work, we first optimize the beam alignment cost functions e.g. the data rate, to reduce the beam sweeping overhead by applying polynomial approximations of its partial derivatives which can then be solved as a system of polynomial equations using well-known tools from algebraic geometry. At this point, a question arises: 'what is a good polynomial approximation?' In this work, we attempt to obtain a 'good polynomial approximation'. Preliminary experiments indicate that our estimated polynomial approximations attain a so-called sweet-spot in terms of the solver speed and accuracy, when evaluated on test beamforming problems., Comment: Accepted for publication in the SIGSAM's ACM Communications in Computer Algebra, as an extended abstract

Few-shot Class-Incremental Learning (FSCIL) presents a unique challenge in Machine Learning (ML), as it necessitates the Incremental Learning (IL) of new classes from sparsely labeled training samples without forgetting previous knowledge. While this field has seen recent progress, it remains an active exploration area. This paper aims to provide a comprehensive and systematic review of FSCIL. In our in-depth examination, we delve into various facets of FSCIL, encompassing the problem definition, the discussion of the primary challenges of unreliable empirical risk minimization and the stability-plasticity dilemma, general schemes, and relevant problems of IL and Few-shot Learning (FSL). Besides, we offer an overview of benchmark datasets and evaluation metrics. Furthermore, we introduce the Few-shot Class-incremental Classification (FSCIC) methods from data-based, structure-based, and optimization-based approaches and the Few-shot Class-incremental Object Detection (FSCIOD) methods from anchor-free and anchor-based approaches. Beyond these, we present several promising research directions within FSCIL that merit further investigation.

Remote photoplethysmography (rPPG) offers a state-of-the-art, non-contact methodology for estimating human pulse by analyzing facial videos. Despite its potential, rPPG methods can be susceptible to various artifacts, such as noise, occlusions, and other obstructions caused by sunglasses, masks, or even involuntary facial contact, such as individuals inadvertently touching their faces. In this study, we apply image processing transformations to intentionally degrade video quality, mimicking these challenging conditions, and subsequently evaluate the performance of both non-learning and learning-based rPPG methods on the deteriorated data. Our results reveal a significant decrease in accuracy in the presence of these artifacts, prompting us to propose the application of restoration techniques, such as denoising and inpainting, to improve heart-rate estimation outcomes. By addressing these challenging conditions and occlusion artifacts, our approach aims to make rPPG methods more robust and adaptable to real-world situations. To assess the effectiveness of our proposed methods, we undertake comprehensive experiments on three publicly available datasets, encompassing a wide range of scenarios and artifact types. Our findings underscore the potential to construct a robust rPPG system by employing an optimal combination of restoration algorithms and rPPG techniques. Moreover, our study contributes to the advancement of privacy-conscious rPPG methodologies, thereby bolstering the overall utility and impact of this innovative technology in the field of remote heart-rate estimation under realistic and diverse conditions., Comment: 8 pages, 7 figures, 4 tables

Artificial intelligence (AI) has become a part of everyday conversation and our lives. It is considered as the new electricity that is revolutionizing the world. AI is heavily invested in both industry and academy. However, there is also a lot of hype in the current AI debate. AI based on so-called deep learning has achieved impressive results in many problems, but its limits are already visible. AI has been under research since the 1940s, and the industry has seen many ups and downs due to over-expectations and related disappointments that have followed. The purpose of this book is to give a realistic picture of AI, its history, its potential and limitations. We believe that AI is a helper, not a ruler of humans. We begin by describing what AI is and how it has evolved over the decades. After fundamentals, we explain the importance of massive data for the current mainstream of artificial intelligence. The most common representations for AI, methods, and machine learning are covered. In addition, the main application areas are introduced. Computer vision has been central to the development of AI. The book provides a general introduction to computer vision, and includes an exposure to the results and applications of our own research. Emotions are central to human intelligence, but little use has been made in AI. We present the basics of emotional intelligence and our own research on the topic. We discuss super-intelligence that transcends human understanding, explaining why such achievement seems impossible on the basis of present knowledge,and how AI could be improved. Finally, a summary is made of the current state of AI and what to do in the future. In the appendix, we look at the development of AI education, especially from the perspective of contents at our own university., Comment: 234 pages. Published as an electronic publication at the University of Oulu, Finland, in December 2021, ISBN: 978-952-62-3199-0 link http://jultika.oulu.fi/Record/isbn978-952-62-3199-0

Inspired by the success of classical neural networks, there has been tremendous effort to develop classical effective neural networks into quantum concept. In this paper, a novel hybrid quantum-classical neural network with deep residual learning (Res-HQCNN) is proposed. We firstly analysis how to connect residual block structure with a quantum neural network, and give the corresponding training algorithm. At the same time, the advantages and disadvantages of transforming deep residual learning into quantum concept are provided. As a result, the model can be trained in an end-to-end fashion, analogue to the backpropagation in classical neural networks. To explore the effectiveness of Res-HQCNN , we perform extensive experiments for quantum data with or without noisy on classical computer. The experimental results show the Res-HQCNN performs better to learn an unknown unitary transformation and has stronger robustness for noisy data, when compared to state of the arts. Moreover, the possible methods of combining residual learning with quantum neural networks are also discussed., Comment: 37 pages, 13 figures

We propose a decision triggered data transmission and collection (DTDTC) protocol for condition monitoring and anomaly detection in the industrial Internet of things (IIoT). In the IIoT, the collection, processing, encoding, and transmission of the sensor readings are usually not for the reconstruction of the original data but for decision making at the fusion center. By moving the decision making process to the local end devices, the amount of data transmission can be significantly reduced, especially when normal signals with positive decisions dominate in the whole life cycle and the fusion center is only interested in collecting the abnormal data. The proposed concept combines compressive sensing, machine learning, data transmission, and joint decision making. The sensor readings are encoded and transmitted to the fusion center only when abnormal signals with negative decisions are detected. All the abnormal signals from the end devices are gathered at the fusion center for a joint decision with feedback messages forwarded to the local actuators. The advantage of such an approach lies in that it can significantly reduce the volume of data to be transmitted through wireless links. Moreover, the introduction of compressive sensing can further reduce the dimension of data tremendously. An exemplary case, i.e., diesel engine condition monitoring, is provided to validate the effectiveness and efficiency of the proposed scheme compared to the conventional ones., Comment: 5 pages, 4 figures, accepted by IEEE WCNC, 2020

The concept of reconfigurable intelligent surface (RIS) has been proposed to change the propagation of electromagnetic waves, e.g., reflection, diffraction, and refraction. To accomplish this goal, the phase values of the discrete RIS units need to be optimized. In this paper, we consider RIS-aided millimeter-wave (mmWave) multiple-input multiple-output (MIMO) systems for both accurate positioning and high data-rate transmission. We propose an adaptive phase shifter design based on hierarchical codebooks and feedback from the mobile station (MS). The benefit of the scheme lies in that the RIS does not require deployment of any active sensors and baseband processing units. During the update process of phase shifters, the combining vector at the MS is also sequentially refined. Simulation results show the performance improvement of the proposed algorithm over the random design scheme, in terms of both positioning accuracy and data rate. Moreover, the performance converges to exhaustive search scheme even in the low signal-to-noise ratio regime., Comment: 6 pages, 6 figures, submitted to IEEE WCNC

Millimeter-wave (mmWave) multiple-input multiple-output (MIMO) system for the fifth generation (5G) cellular communications can also enable single-anchor positioning and object tracking due to its large bandwidth and inherently high angular resolution. In this paper, we introduce the newly invented concept, large intelligent surface (LIS), to mmWave positioning systems, study the theoretical performance bounds (i.e., Cram\'er-Rao lower bounds) for positioning, and evaluate the impact of the number of LIS elements and the value of phase shifters on the position estimation accuracy compared to the conventional scheme with one direct link and one non-line-of-sight path. It is verified that better performance can be achieved with a LIS from the theoretical analyses and numerical study., Comment: 5 pages, 5 figures, conference

Low-level sensory data processing in many Internet-of-Things (IoT) devices pursue energy efficiency by utilizing sleep modes or slowing the clocking to the minimum. To curb the share of stand-by power dissipation in those designs, near-threshold/sub-threshold operational points or ultra-low-leakage processes in fabrication are employed. Those limit the clocking rates significantly, reducing the computing throughputs of individual processing cores. In this contribution we explore compensating for the performance loss of operating in near-threshold region (Vdd =0.6V) through massive parallelization. Benefits of near-threshold operation and massive parallelism are optimum energy consumption per instruction operation and minimized memory roundtrips, respectively. The Processing Elements (PE) of the design are based on Transport Triggered Architecture. The fine grained programmable parallel solution allows for fast and efficient computation of learnable low-level features (e.g. local binary descriptors and convolutions). Other operations, including Max-pooling have also been implemented. The programmable design achieves excellent energy efficiency for Local Binary Patterns computations.

In this contribution, it is proposes to limit the quantization search space of a successive approximation analog-to-digital converter through an analytic derivation of maximum possible sample-to-sample variation. The presented example design of the proposed ADC is an 8-bit 1MS/s ADC with SAR logic customized to incorporate this priori information while no modification has been required to the analog circuitry. In comparison to conventional SAR conversion, the proposed tracking approach yields significant reduction in total power consumption in oversampling mode. The power savings are due to the reduced number of SAR cycles, and voltage variation minimization across DAC capacitors. The design is reconfigurable both to conventional SAR sampling and the proposed tracking scheme. The approach is attractive for SAR ADCs embedded in very low power micro-controllers.

This paper summarizes the design of a programmable processor with transport triggered architecture (TTA) for decoding LDPC and turbo codes. The processor architecture is designed in such a manner that it can be programmed for LDPC or turbo decoding for the purpose of internetworking and roaming between different networks. The standard trellis based maximum a posteriori (MAP) algorithm is used for turbo decoding. Unlike most other implementations, a supercode based sum-product algorithm is used for the check node message computation for LDPC decoding. This approach ensures the highest hardware utilization of the processor architecture for the two different algorithms. Up to our knowledge, this is the first attempt to design a TTA processor for the LDPC decoder. The processor is programmed with a high level language to meet the time-to-market requirement. The optimization techniques and the usage of the function units for both algorithms are explained in detail. The processor achieves 22.64 Mbps throughput for turbo decoding with a single iteration and 10.12 Mbps throughput for LDPC decoding with five iterations for a clock frequency of 200 MHz., Comment: 8 pages, 7 figures, conference