Cryptosystem designers frequently assume that secrets will be manipulated in closed, reliable computing environments. Unfortunately, actual computers and microchips leak information about the operations they process. This paper examines specific methods for analyzing power consumption measurements to find secret keys from tamper resistant devices. We also discuss approaches for building cryptosystems that can operate securely in existing hardware that leaks information.

/pdf/differential-power-analysis-4t7fjmuc7b.pdf

Differential Power Analysis

This monograph provides an overview of general deep learning methodology and its applications to a variety of signal and information processing tasks. The application areas are chosen with the following three criteria in mind: (1) expertise or knowledge of the authors; (2) the application areas that have already been transformed by the successful use of deep learning technology, such as speech recognition and computer vision; and (3) the application areas that have the potential to be impacted significantly by deep learning and that have been experiencing research growth, including natural language and text processing, information retrieval, and multimodal information processing empowered by multi-task deep learning.

/pdf/deep-learning-methods-and-applications-5cyi610sjz.pdf

Deep Learning: Methods and Applications

This ebook is the first authorized digital version of Kernighan and Ritchie's 1988 classic, The C Programming Language (2nd Ed.). One of the best-selling programming books published in the last fifty years, "K&R" has been called everything from the "bible" to "a landmark in computer science" and it has influenced generations of programmers. Available now for all leading ebook platforms, this concise and beautifully written text is a "must-have" reference for every serious programmers digital library.

As modestly described by the authors in the Preface to the First Edition, this "is not an introductory programming manual; it assumes some familiarity with basic programming concepts like variables, assignment statements, loops, and functions. Nonetheless, a novice programmer should be able to read along and pick up the language, although access to a more knowledgeable colleague will help."

The C programming language

Recently, the hybrid deep neural network (DNN)- hidden Markov model (HMM) has been shown to significantly improve speech recognition performance over the conventional Gaussian mixture model (GMM)-HMM. The performance improvement is partially attributed to the ability of the DNN to model complex correlations in speech features. In this paper, we show that further error rate reduction can be obtained by using convolutional neural networks (CNNs). We first present a concise description of the basic CNN and explain how it can be used for speech recognition. We further propose a limited-weight-sharing scheme that can better model speech features. The special structure such as local connectivity, weight sharing, and pooling in CNNs exhibits some degree of invariance to small shifts of speech features along the frequency axis, which is important to deal with speaker and environment variations. Experimental results show that CNNs reduce the error rate by 6%-10% compared with DNNs on the TIMIT phone recognition and the voice search large vocabulary speech recognition tasks.

/pdf/convolutional-neural-networks-for-speech-recognition-59lskxzlgf.pdf

Convolutional neural networks for speech recognition

From the Publisher:
Connectionist Speech Recognition: A Hybrid Approach describes the theory and implementation of a method to incorporate neural network approaches into state-of-the-art continuous speech recognition systems based on Hidden Markov Models (HMMs) to improve their performance. In this framework, neural networks (and in particular, multilayer perceptrons or MLPs) have been restricted to well-defined subtasks of the whole system, i.e., HMM emission probability estimation and feature extraction. The book describes a successful five year international collaboration between the authors. The lessons learned form a case study that demonstrates how hybrid systems can be developed to combine neural networks with more traditional statistical approaches. The book illustrates both the advantages and limitations of neural networks in the framework of a statistical system. Using standard databases and comparing with some conventional approaches, it is shown that MLP probability estimation can improve recognition performance. Other approaches are discussed, though there is no such unequivocal experimental result for these methods. Connectionist Speech Recognition: A Hybrid Approach is of use to anyone intending to use neural networks for speech recognition or within the framework provided by an existing successful statistical approach. This includes research and development groups working in the field of speech recognition, both with standard and neural network approaches, as well as other pattern recognition and/or neural network researchers. This book is also suitable as a text for advanced courses on neural networks or speech processing.

Connectionist Speech Recognition: A Hybrid Approach

The demand for multimedia mobile terminals has created a need for low power implementation of video compression algorithms. In this paper we consider different implementations for the discrete cosine transform. The effect of pipelining and parallelism on reducing the power dissipation is considered for fast discrete cosine transform algorithms, as well as ROM-based algorithms.

Low-power implementation of discrete cosine transform

Due to CMOS technology scaling, devices are getting smaller, faster, and operating at lower supply voltages. The reduced capacitances and power supply voltages and the increased chip density to perform more functionality result in increasing the soft errors and making them one of the essential design constraints at the same level as delay and power. Even though the impact of process variations on the performance and the power consumption has been investigated by many researchers, its impact on soft errors has not been paid enough attention. This impact is investigated in this paper for 65-nm CMOS technology. The soft error yield is defined in this paper similar to the timing yield and the power yield. This paper shows that the soft error yield of the sense-amplifier based flip flop (SA-FF) is very poor. Therefore, soft error mitigation techniques are required when using this flip-flop topology. The semi-dynamic flip-flop (SD-FF) exhibits the best soft error yield behavior with a very high performance at the expense of large power requirement. Finally, some design insights are proposed to guide flip-flops designers to select the best flip-flop topology that satisfies their specific circuit soft error rate constraints.

Comparative analysis of process variation impact on flip-flops soft error rate

A SRAM includes an ECL input buffer connected between an address bus and a W-OR predecoder array. The logic output of the W-OR predecoder array is applied to a level translator array and level shifted. The level shifted output of the level translator array is supplied to a plurality of self-resetting word-line decoder and driver (WLDD) circuits. The WLDD circuits supply activation pulses to selected blocks of memory in a memory cell array. Sense amplifiers sense and latch-in the data stored in the activated selected blocks of memory. The design of the W-OR predecoder array, level translator array, WLDD circuits and sense amplifiers is such to reduce the overall power consumption of the SRAM.

Low-power BiCMOS/ECL SRAM

As VLSI technologies scale into the deep submicron (DSM) realm, the minimum feature size continues to shrink. In contrast, the average die size is expected to remain constant or to slightly increase with each technology generation. This results in an average increase in the global interconnect lengths. In order to mitigate the impact of these global wires, buffer insertion is the most widely used technique. However, unconstrained buffering is bound to adversely affect the overall chip performance. In fact, the number of global interconnect buffers is expected to reach several hundreds of thousands to achieve an appropriate timing closure. This increase in buffers is destined to have a huge impact on the chip power consumption. In order to mitigate the impact of the power consumed by the interconnect buffers, a low power multi-pin routing methodology is proposed. The problem is tackled based on a graph theoretic representation of the interconnects that explores the various buffer sizing options in order to identify the lowest power path that can be assigned to the interconnect being routed. The formulation was found to be of a pseudo-polynomial complexity which fits well within the context of the increased number of buffers. The methodology is tested using the MCNC floorplan benchmarks to verify the correctness and the complexity. These tests showed that an average power saving as high as 45% with a 10% sacrifice in delay is observed

A Low-Power Multi-Pin Maze Routing Methodology

Current mode logic (CML) is used to obtain NMOS VLSI logic gates with nanosecond delays at a speed power product of 1.5 pJ. The gates use a reference voltage generated by the difference between an enhancement- and a depletion-type MOSFET. A depletion-type MOSFET is used for the current sources and the loads. The gates have been designed with a logic swing of 2 V in a 2.5-µm NMOS VLSI technology, simulated, and their circuit performance is studied.

Mohamed I. Elmasry

Papers

Low-power implementation of discrete cosine transform

Comparative analysis of process variation impact on flip-flops soft error rate

Low-power BiCMOS/ECL SRAM

A Low-Power Multi-Pin Maze Routing Methodology

Nanosecond NMOS VLSI current mode logic