<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet type="text/xsl" href="static/style.xsl"?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-11-21T01:19:20Z</responseDate><request verb="GetRecord" identifier="oai:macsphere.mcmaster.ca:11375/30409" metadataPrefix="dim">https://prod-ms-be.lib.mcmaster.ca/server/oai/request</request><GetRecord><record><header><identifier>oai:macsphere.mcmaster.ca:11375/30409</identifier><datestamp>2026-09-24T18:38:08Z</datestamp><setSpec>com_11375_271</setSpec><setSpec>col_11375_272</setSpec></header><metadata><dim:dim xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:doc="http://www.lyncode.com/xoai" xsi:schemaLocation="http://www.dspace.org/xmlns/dspace/dim http://www.dspace.org/schema/dim.xsd">
   <dim:field mdschema="dc" element="contributor" qualifier="advisor">Dumitrescu, Sorina</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="author">Zendehboodi, Sara</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="department" lang="en_US">Electrical and Computer Engineering</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="accessioned">2024-10-11T18:52:12Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="available">2024-10-11T18:52:12Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="issued">2024-11</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">http://hdl.handle.net/11375/30409</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">https://doi.org/10.71548/24948</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="abstract" lang="en_US">This thesis addresses the design of distributed scalar quantizers (DSQs) for two sensors,&#xd;
tailored to maximize the classification accuracy for a pre-trained binary linear classifier&#xd;
at the central node, diverging from traditional designs that prioritize data reconstruction&#xd;
quality.&#xd;
The first contribution of this thesis is the development of efficient globally optimal&#xd;
DSQ design algorithms for two correlated discrete sources when the quantizer cells are&#xd;
assumed to be convex. First, it is shown that the problem is equivalent to a minimum&#xd;
weight path problem (with certain constraints) in a weighted directed acyclic graph.&#xd;
The latter problem can be solved using dynamic programming with O(K_1K_2M^4) computational&#xd;
complexity, where Ki, is the number of cells for the quantizer of source i,&#xd;
i = 1, 2, and M is the size of the union of the sources’ alphabets. Additionally, it is&#xd;
proved that the dynamic programming algorithm can be expedited by a factor of M by&#xd;
exploiting the so called Monge property, for scenarios where the pre-trained classifier is&#xd;
the optimal classifier for the unquantized sources.&#xd;
Next, the design of so-called staggered DSQs (SDSQs) is addressed, i.e., DSQ’s with&#xd;
K_1 = K_2 = K and with the thresholds of the two quantizers being interleaved. First, a&#xd;
faster dynamic programming algorithm with only O(KM^2) time complexity is devised&#xd;
for the design of the SDSQ that minimizes an upperbound on the classification error.&#xd;
This sped up is obtained by simplifying the graph model for the problem. Moreover,&#xd;
it is shown that this algorithm can also be further accelerated by a factor of M when&#xd;
the pre-trained linear classifier is the optimal classifier. Furthermore, some theoretical&#xd;
results are derived that provide support to imposing the above constraints to the DSQ&#xd;
design problem in the case when the pre-trained classifier is optimal. First, it is shown&#xd;
that when the sources (discrete or continuous) satisfy a certain symmetry property, the&#xd;
SDSQ that minimizes the modified cost also minimizes the original cost within the class&#xd;
of DSQs without the staggerness constraint. For continuous sources, it is also shown&#xd;
that the SDSQ that minimizes the modified cost also minimizes the original cost and all&#xd;
quantizer thresholds are distinct, even if the sources do not satisfy the aforementioned&#xd;
symmetry condition. The latter result implies that DSQs with identical encoders are&#xd;
not optimal even when the sources has the same marginal distribution, a fact which is&#xd;
proved here for the first time, up to our knowledge.&#xd;
The last (but not least) contribution of this thesis resides in leveraging the aforementioned&#xd;
results to obtain efficient globally optimal solution algorithms for the problem&#xd;
of decentralized detection under the probability of error criterion of two discrete vector&#xd;
sources that are conditionally independent given any class label. The previously&#xd;
known globally optimal solution has O(N^(K_1+K_2+1)) time complexity, where N is the&#xd;
size of the union of the alphabets of the two sources. We show that by applying an&#xd;
appropriate transformation to each vector source, the problem reduces to the problem&#xd;
of designing the optimal DSQ with convex cells in the transformed scalar domain for&#xd;
a scenario where the pre-trained linear classifier is the optimal classifier. We conclude&#xd;
that the problem can be solved by a much faster algorithm with only O(K_1K_2N^3) time&#xd;
complexity. Similarly, for the case of equal quantizer rates, the problem can be solved&#xd;
in O(KN) operations if the sources satisfy an additional symmetry condition. Furthermore,&#xd;
our results prove the conjecture that for continuous sources, imposing the&#xd;
constraint that the encoders be identical precludes optimality, even when the marginal&#xd;
distributions of the sources are the same.</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="degreetype" lang="en_US">Thesis</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="degree" lang="en_US">Doctor of Philosophy (PhD)</dim:field>
   <dim:field mdschema="dc" element="language" qualifier="iso" lang="en_US">en</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Distributed Scalar Quantizer</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Binary Linear Classification</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Decentralized Detection</dim:field>
   <dim:field mdschema="dc" element="title" lang="en_US">Study and Design of Globally Optimal Distributed Scalar Quantizer for Binary Linear Classification</dim:field>
   <dim:field mdschema="dc" element="type" lang="en_US">Thesis</dim:field>
   <dim:field mdschema="others" element="access-status">open.access</dim:field>
</dim:dim></metadata></record></GetRecord></OAI-PMH>