1,721,214 research outputs found
A Probably Approximately Correct Framework to estimate Performance Degradation in Embedded Systems
No full textIntelligence for embedded systems: a methodological approach
Full text linkAddressing current issues of which any engineer or computer scientist should be aware, this monograph is a response to the need to adopt a new computational paradigm as the methodological basis for designing pervasive embedded systems with sensor capabilities. The requirements of this paradigm are to control complexity, to limit cost and energy consumption, and to provide adaptation and cognition abilities allowing the embedded system to interact proactively with the real world. The quest for such intelligence requires the formalization of a new generation of intelligent systems able to exploit advances in digital architectures and in sensing technologies. The book sheds light on the theory behind intelligence for embedded systems with specific focus on: · robustness (the robustness of a computational flow and its evaluation); · intelligence (how to mimic the adaptation and cognition abilities of the human brain), · the capacity to learn in non-stationary and evolving environments by detecting changes and reacting accordingly; and · a new paradigm that, by accepting results that are correct in probability, allows the complexity of the embedded application the be kept under control. Theories, concepts and methods are provided to motivate researchers in this exciting and timely interdisciplinary area. Applications such as porting a neural network from a high-precision platform to a digital embedded system and evaluating its robustness level are described. Examples show how the methodology introduced can be adopted in the case of cyber-physical systems to manage the interaction between embedded devices and physical world.. Researchers and graduate students in computer science and various engineering-related disciplines will find the methods and approaches propounded in Intelligence for Embedded Systems of great interest. The book will also be an important resource for practitioners working on embedded systems and applications
Randomised Algorithms: A System-Level, Poly-time Analysis of Robust Computation
No full textProvides a methodology for analyzing the performance degradation of a computation once it has been affected by perturbations. The suggested methodology, by relaxing all assumptions made in the related literature, provides design guidelines for the subsequent implementation of complex computations in physical devices. Implementation issues, such as finite precision representation, fluctuations of the production parameters and aging effects, can be studied directly at the system level, independent of any technological aspect and quantization technique. Only the behavioral description of the computational flow, which is assumed to be Lebesgue-measurable, and the architecture to be investigated are needed. The suggested analysis is based on the theory of randomized algorithms, which transforms the computationally intractable problem of robustness investigation into a polynomial-time algorithm by resorting to probability
Just-in-time Adaptive Classifiers. Part II. Designing the classifier
No full textAging effects, environmental changes, thermal drifts, and soft and hard faults affect physical systems by changing their nature and behavior over time. To cope with a process evolution adaptive solutions must be envisaged to track its dynamics; in this direction, adaptive classifiers are generally designed by assuming the stationary hypothesis for the process generating the data with very few results addressing nonstationary environments. This paper proposes a methodology based on -nearest neighbor (NN) classifiers for designing adaptive classification systems able to react to changing conditions just-in-time (JIT), i.e., exactly when it is needed. -NN classifiers have been selected for their computational-free training phase, the possibility to easily estimate the model complexity and keep under control the computational complexity of the classifier through suitable data reduction mechanisms. A JIT classifier requires a temporal detection of a (possible) process deviation (aspect tackled in a companion paper) followed by an adaptive management of the knowledge base (KB) of the classifier to cope with the process change. The novelty of the proposed approach resides in the general framework supporting the real-time update of the KB of the classification system in response to novel information coming from the process both in stationary conditions (accuracy improvement) and in nonstationary ones (process tracking) and in providing a suitable estimate of . It is shown that the classification system grants consistency once the change targets the process generating the data in a new stationary state, as it is the case in many real applications
A Genetic-Based Application Oriented Approach to Optimize RFID-like Passive Sensor Devices for Homeland Security
No full textA Perturbation Size-independent Analysis of Robustness in Feedforward Neural Networks by Randomized Algorithms
No full text
- …
