etd AT Indian Institute of Science: Recent submissions
Now showing items 561-580 of 7574
-
Fault tolerance in feedforward neural networks using minimax optimization
Neural computing is currently being proposed as a viable solution to several problems. The approach of neural computing is to capture the guiding principles that underlie the functioning of the human brain and apply them ... -
Frames as abstractions for efficient multimedia object retrieval
Multimedia Information Systems (MMISs) integrate various media types - text, audio, video, graphics, and animation - for machine-processable storage, retrieval, and presentation. These systems are characterized by complex ... -
Application of trust region on methods to learning in feed forward neural networks
This thesis presents a novel algorithm for training feed-forward neural networks, addressing the limitations of the widely used back-propagation and Quickprop algorithms. While Quickprop is significantly faster than standard ... -
Algorithmic and architectural solutions for scheduling task graphs on multiprocessors
Scheduling is a critical challenge in the design of multicomputer systems, especially for minimizing execution time of parallel programs represented as task graphs. This thesis presents both algorithmic and architectural ... -
Reason maintenance and logic
This thesis studies the theory and applications of recison maintenance from a logical perspective. First a series of exam ples introduces the concepts, structures and techniques of reason m aintenance. T he exam ples ... -
Specification environment for distributed real time systems
Real-time computer systems usually find applications in safety-critical areas. Thus, real-time systems need to be extremely reliable. However, most of the real-time systems are inherently "distributed" in nature. The ... -
Algorithms for some problems on trees and graphs
Design and analysis of graph algorithms is an active area of research in Computer Science. Trees form an important subclass of graphs and have vast applications. In this thesis, we present some new algorithmic results on ... -
A Cognitive science approach for modelling representational organization and learning in humans
Artificial intelligence (AI) is concerned with discovering intelligence in human mind and its application to the problems in real world. As a consequence it is associated with understanding the nature of human mind, that ... -
knowledge teaching : an alternative strategy for knowkedge-base development
The development of knowledge-based systems relies heavily on the transfer of human expertise into a structured knowledge base - a process known as knowledge acquisition or knowledge elicitation. This process is widely ... -
Systolic architectures for realistic 3-D graphics.
Computer Graphics has come a long way from its humble beginning of lin£ drawing displays in the early sixties to the present day extremely realistic, synthetic, and photography quality pictures. As a result, interactive ... -
Algorithmic knowledge for a knowledge-based clustering environment.
Clustering is a process of partitioning a given set of objects into meaningful groups. It has potential applications in many areas, including: 1.Image Segmentation, 2. Speech/speaker Classification, 2.ECG Classificati ... -
Artificial intelligence research RR-76 : Machine Visualization of three dimensional objects
Computer Simulation of One Aspect of Human Visual Perceptual Capabilities This simulation focuses on the human ability to visualize object views resulting from the rotation of known objects. The objects considered are ... -
Decentralized optimization with appliation to flow control
Decentralized computation is considered an important design goal of large systems. This may be to increase the reliability of the system or to reduce the computation time, and in some situations, it is a necessity. ... -
Fault tolerant trasputer based firmware for parallel solution of partial differntial equations by the boundry element method
This work presents a fault-tolerant multi-transputer architecture capable of handling single-node failures without interrupting program execution. The system dynamically reconfigures the interconnections among transputers ... -
Diagnostic Reasoning with Surface and Deep Level Knowledge in Medical Domain
The aim of this work is to structure the domain knowledge based upon some conceptual models so that the knowledge can be efficiently used for diagnostic reasoning in medical domain. Medical domain is complex and incompletely ... -
Solution of algebraic riccati equations on wavefront array processors
The need for high performance computers has been felt for some time in application areas such as on-line control, system identification and estimation in real-time, complex control system design, etc. Parallel computers ... -
Analogies and their use in problem-solving
The thesis deals with the meaning of "analogy" and how it can be extracted and exploited for the following two classes of problems: CLASS 1: How can one transfer the theorems of one domain (in mathematics) to analogous ... -
Hierarchical Data Structures data structures based on quadtrees
In the present thesis, a size-based hierarchical ordering of the codes of the linear quadtree, called the Linear Hierarchical Quadtree (LHQT), is shown to reduce the time spent while exploring the adjacencies, without any ... -
Linear block source coding for binary memoryless sources
Despite the existence of well-defined rate-distortion functions for general binary memoryless sources (BMS), block source coding theory has traditionally been restricted to sources with equiprobable symbols, in which case ... -
S-Nets : Tool for the performance evaluation of hard real-time scheduling algorithms
A new Petri net-based tool is developed for an integrated analysis of schedulability and performance of Hard Real-Time Systems (HRTSs). The existing Deterministic Timed Petri Nets (DTPNs) do not facilitate modeling of ...