Kerns H. Powers. / "February 1, 1956." "This report is indentical with a thesis submitted to the Department of Electrical Engineering, M.I.T., 1956, in partial fulfillment of the requirements for the degree of Doctor of Science." / Bibliography: p. 104-105. / Army Signal Corps Contract DA36-039-sc-42607 Project 102B Dept. of the Army Project 3-99-10-022

Richard A. Silverman. / "May 25, 1955"--Cover. "Reprinted from the IRE transactions of the professional group on information theory, volume IT-1, number 3, December, 1955"--Cover. / Includes bibliographical references. / Army Signal Corps Contract DA36-039 sc-42607 Project 102B Dept. of the Army Project 3-99-10-022

Amiel Feinstein. / "June 1, 1954." "This report is identical with a thesis submitted to the Department of Physics, M.I.T., ... for the degree of Doctor of Philosophy." / Bibliography: p. 27-28. / Army Signal Corps Contract DA36-039 sc-100 Project 8-102-B-0 Dept. of the Army Project 3-99-10-022

We formulate and interpret several multi-modal registration methods in the context of a unified statistical and information theoretic framework. A unified interpretation clarifies the implicit assumptions of each method yielding a better understanding of their relative strengths and weaknesses. Additionally, we discuss a generative statistical model from which we derive a novel analysis tool, the "auto-information function", as a means of assessing and exploiting the common spatial dependencies inherent in multi-modal imagery. We analytically derive useful properties of the "auto-information" as well as verify them empirically on multi-modal imagery. Among the useful aspects of the "auto-information function" is that it can be computed from imaging modalities independently and it allows one to decompose the search space of registration problems.

In this dissertation, we introduce a relay protocol for multiple-relay networks called parity forwarding. The relay channel is a classic multiuser information theory problem introduced in 1971, modeling a network of three nodes: a source, a destination, and a relay node. The relay has no message of its own and assists the source to communicate to the destination. Of the two main coding techniques for the relay channel, decode-and-forward (DF) and compress-and-forward (CF), our focus is to understand the DF scheme for single- and multiple-relay channels. For the single-relay channel, we present an interpretation of the classic DF strategy from a linear-coding perspective. Identifying binning in the classic DF strategy as parity bit generation, we devise ensembles of low-density parity-check (LDPC) codes, called bilayer LDPC codes, to incorporate the parity bits generated by the relay in the decoding process at the destination. We develop code design techniques and optimize the parameters of the bilayer LDPC code structure to show that bilayer LDPC codes can approach the theoretical DF rate. Inspired by the relation between binning and parity bits, we introduce the parity forwarding protocol to improve on the achievable DF rate in a multiple-relay network. For a two-relay network, we show that the previous multihop DF protocol can be improved if the relays forward parity messages and receivers (relays or the destination)use a joint decoding scheme. Depending on the relation between relay messages and decoded messages at each relay, different parity forwarding protocols are possible. In this thesis, we present a structured characterization of a class of parity forwarding protocols for multiple-relay networks. We propose a tree structure to describe the relations between messages. Using this tree structure, we derive a closed-form expression for the parity forwarding rate in a relay network with an arbitrary number of relays. Finally, examples of new types of degraded multiple-relay networks are presented for which the parity forwarding protocol achieves capacity.

In this dissertation, we introduce a relay protocol for multiple-relay networks called parity forwarding. The relay channel is a classic multiuser information theory problem introduced in 1971, modeling a network of three nodes: a source, a destination, and a relay node. The relay has no message of its own and assists the source to communicate to the destination. Of the two main coding techniques for the relay channel, decode-and-forward (DF) and compress-and-forward (CF), our focus is to understand the DF scheme for single- and multiple-relay channels. For the single-relay channel, we present an interpretation of the classic DF strategy from a linear-coding perspective. Identifying binning in the classic DF strategy as parity bit generation, we devise ensembles of low-density parity-check (LDPC) codes, called bilayer LDPC codes, to incorporate the parity bits generated by the relay in the decoding process at the destination. We develop code design techniques and optimize the parameters of the bilayer LDPC code structure to show that bilayer LDPC codes can approach the theoretical DF rate. Inspired by the relation between binning and parity bits, we introduce the parity forwarding protocol to improve on the achievable DF rate in a multiple-relay network. For a two-relay network, we show that the previous multihop DF protocol can be improved if the relays forward parity messages and receivers (relays or the destination)use a joint decoding scheme. Depending on the relation between relay messages and decoded messages at each relay, different parity forwarding protocols are possible. In this thesis, we present a structured characterization of a class of parity forwarding protocols for multiple-relay networks. We propose a tree structure to describe the relations between messages. Using this tree structure, we derive a closed-form expression for the parity forwarding rate in a relay network with an arbitrary number of relays. Finally, examples of new types of degraded multiple-relay networks are presented for which the parity forwarding protocol achieves capacity.

In this dissertation, we study the state-dependent two-user interference channel, where the state information is non-causally known at both transmitters but unknown to either of the receivers. We first propose two coding schemes for the discrete memoryless case: simultaneous encoding for the sub-messages in the first one and super-position encoding in the second one, both with rate splitting and Gel'fand-Pinsker coding. The corresponding achievable rate regions are established. Moreover, for the Gaussian case, we focus on the simultaneous encoding scheme and propose an active interference cancellation mechanism, which is a generalized dirty-paper coding technique, to partially eliminate the state effect at the receivers. The corresponding achievable rate region is then derived. We also propose several heuristic schemes for some special cases: the strong interference case, the mixed interference case, and the weak interference case. For the strong and mixed interference case, numerical results are provided to show that active interference cancellation significantly enlarges the achievable rate region. For the weak interference case, flexible power splitting instead of active interference cancellation improves the performance significantly. Moreover, we focus on the simplest symmetric case, where both direct link gains are the same with each other, and both interfering link gains are the same with each other. We apply the above coding scheme with different dirty paper coding parameters. When the state is additive and symmetric at both receivers, we study both strong and weak interference scenarios and characterize the theoretical gap between the achievable symmetric rate and the upper bound, which is shown to be less than 1/4 bit for the strong interference case and less than 3/4 bit for the weak interference case. Then we provide numerical evaluations of the achievable rates against the upper bound, which validates the theoretical analysis for both strong and weak interference scenarios. Finally, we define the generalized degrees of freedom for the symmetric Gaussian case, and compare the lower bounds against the upper bounds for both strong and weak interference cases. We also show that our achievable schemes can obtain the exact optimal values of the generalized degrees of freedom, i.e., the lower bounds meet the upper bounds for both strong and weak interference cases.

This thesis deals with one important aspect of Artificial Intelligence, knowledge representation and reasoning. Problems like knowledge representation hypothesis, production systems, and the proceduralist and declarativist controversy are discussed. The conventional concept of Al system is based on a model where the reasoning and knowledge base are separate. A different model is proposed in this thesis. In this model, the sensory knowledge is considered the basic elements of the knowledge base, and the higher level knowledge form a hierarchical structure. Reasoning power, in this model, is only part of the, knowledge structure. A frame-based knowledge representation system, WenLy, is presented. It is designed to construct a knowledge base from the bottom up, which is conformed to the proposed model in this thesis. The unique characteristic of WenLy is that its reasoning is accomplished by the code imbedded in the knowledge base instead of a separate inference engine. This is the correct approach according to the knowledge structure model herein proposed. An example of the application is shown. WenLy can be used by a robot to represent the robot's understanding of the outside world. The understanding, however, is limited by the bandwidth of the perception capabilities. It can be shown that the conflict between the declarative and procedural aspects of knowledge can be resolved by following the proposed model.

Thesis (Ph. D.)--University of Kentucky, 2001. / Title from document title page. Document formatted into pages; contains vii, 121 p. : ill. Includes abstract. Includes bibliographical references (p. 116-119).

Thesis (M.S.) in Spatial Information Science and Engineering--University of Maine, 2002. / Includes vita. Includes bibliographical references (leaves 101-106).