Engineering multiversion neural-net systems
نویسندگان
چکیده
In this paper we address the problem of constructing reliable neural-net implementations, given the assumption that any particular implementation will not be totally correct. The approach taken in this paper is to organize the inevitable errors so as to minimize their impact in the context of a multiversion system, i.e., the system functionality is reproduced in multiple versions, which together will constitute the neural-net system. The unique characteristics of neural computing are exploited in order to engineer reliable systems in the form of diverse, multiversion systems that are used together with a "decision strategy" (such as majority vote). Theoretical notions of "methodological diversity" contributing to the improvement of system performance are implemented and tested. An important aspect of the engineering of an optimal system is to overproduce the components and then choose an optimal subset. Three general techniques for choosing final system components are implemented and evaluated. Several different approaches to the effective engineering of complex multiversion systems designs are realized and evaluated to determine overall reliability as well as reliability of the overall system in comparison to the lesser reliability of component substructures.
منابع مشابه
Software diversity: practical statistics for its measurement and exploitation
The topic of this paper is the exploitation of diversity to enhance computer system reliability. It is well-established that a diverse system composed of multiple alternative versions is more reliable than any single version alone, and this knowledge has occasionally been exploited in safety-critical applications. However, it is not clear what this property is, nor how the available diversity i...
متن کاملData-deened Problems and Multiversion Neural-net Systems
Data-de ned problems are not restricted to any particular problem domain, and are common. Data-de ned problems, as the name suggests, are de ned by a set of input-output mappings, and for the problems of particular interest the full details of how the inputs are related to the outputs are unknown. Such problems present the traditional programmer, whether using AI techniques or not, with a di cu...
متن کاملAn Experimental Evaluation of Methodological Diversity in Multiversion Software Reliability
N-version programming has long been mooted as a method of improving software reliability. Earlier studies, which generated apparently discouraging results, ooered pessimistic prognostications for this general strategy. However, further study of the problem has both reened the evaluation procedure and revealed new opportunities for improvement. In particular, exploitation of methodological diver...
متن کاملMulti-Version Systems of Neural Networks for Predicting the Risk of Osteoporosis
This paper describes the methodology of developing multiversion systems using neural networks in the hope of improving their performance and reliability. However, a system implemented by simply combining N neural nets may not necessarily deliver a better result than the individual versions alone. A critical factor to success is the diversity among these versions, which is high probability that ...
متن کاملA multiversion mechanism for intra-object concurrency
In this paper, we propose an extended concurrent object model for distributed multiuser systems called the multiversion concurrent object (MCO). The original concurrent object model is simple because it excludes internal concurrency, but this may cause long blocking on its execution during message sending and external device accesses. Thus, these blockings may decrease system performance. We in...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Neural computation
دوره 8 4 شماره
صفحات -
تاریخ انتشار 1996