Defense Notices


All students and faculty are welcome to attend the final defense of EECS graduate students completing their M.S. or Ph.D. degrees. Defense notices for M.S./Ph.D. presentations for this year and several previous years are listed below in reverse chronological order.

Students who are nearing the completion of their M.S./Ph.D. research should schedule their final defenses through the EECS graduate office at least THREE WEEKS PRIOR to their presentation date so that there is time to complete the degree requirements check, and post the presentation announcement online.

Upcoming Defense Notices

Xiangyu Chen

Toward Data Efficient Learning in Computer Vision

When & Where:


Nichols Hall, Room 246

Committee Members:

Cuncong Zhong, Chair
Prasad Kulkarni
Fengjun Li
Bo Luo
Guanghui Wang

Abstract

Deep learning leads the performance in many areas of computer vision. Deep neural networks usually require a large amount of data to train a good model with the growing number of parameters. However, collecting and labeling a large dataset is not always realistic, e.g. to recognize rare diseases in the medical field. In addition, both collecting and labeling data are labor-intensive and time-consuming. In contrast, studies show that humans can recognize new categories with even a single example, which is apparently in the opposite direction of current machine learning algorithms. Thus, data-efficient learning, where the labeled data scale is relatively small, has attracted increased attention recently. According to the key components of machine learning algorithms, data-efficient learning algorithms can also be divided into three folders, data-based, model-based, and optimization-based. In this study, we investigate two data-based models and one model-based approach.

First, to collect more data to increase data quantity. The most direct way for data-efficient learning is to generate more data to mimic data-rich scenarios. To achieve this, we propose to integrate both spatial and Discrete Cosine Transformation (DCT) based frequency representations to finetune the classifier. In addition to the quantity, another property of data is the quality to the model, different from the quality to human eyes. As language carries denser information than natural images. To mimic language, we propose to explicitly increase the input information density in the frequency domain. The goal of model-based methods in data-efficient learning is mainly to make models converge faster. After carefully examining the self-attention modules in Vision Transformers, we discover that trivial attention covers useful non-trivial attention due to its large amount. To solve this issue, we proposed to divide attention weights into trivial and non-trivial ones by thresholds and suppress the accumulated trivial attention weights. Extensive experiments have been performed to demonstrate the effectiveness of the proposed models.


Yousif Dafalla

Web-Armour: Mitigating Reconnaissance and Vulnerability Scanning with Injecting Scan-Impeding Delays in Web Deployments

When & Where:


Nichols Hall, Room 250 (Gemini Room)

Committee Members:

Alex Bardas, Chair
Drew Davidson
Fengjun Li
Bo Luo
ZJ Wang

Abstract

Scanning hosts on the internet for vulnerable devices and services is a key step in numerous cyberattacks. Previous work has shown that scanning is a widespread phenomenon on the internet and commonly targets web application/server deployments. Given that automated scanning is a crucial step in many cyberattacks, it would be beneficial to make it more difficult for adversaries to perform such activity.

In this work, we propose Web-Armour, a mitigation approach to adversarial reconnaissance and vulnerability scanning of web deployments. The proposed approach relies on injecting scanning impeding delays to infrequently or rarely used portions of a web deployment. Web-Armour has two goals: First, increase the cost for attackers to perform automated reconnaissance and vulnerability scanning; Second, introduce minimal to negligible performance overhead to benign users of the deployment. We evaluate Web-Armour on live environments, operated by real users, and on different controlled (offline) scenarios. We show that Web-Armour can effectively lead to thwarting reconnaissance and internet-wide scanning.


Sandhya Kandaswamy

An Empirical Evaluation of Multi-Resource Scheduling for Moldable Workflows

When & Where:


Eaton Hall, Room 2001B

Committee Members:

Hongyang Sun, Chair
Suzanne Shontz
Heechul Yun


Abstract

Resource scheduling plays a vital role in High-Performance Computing (HPC) systems. However, most scheduling research in HPC has focused on only a single type of resource (e.g., computing cores or I/O resources). With the advancement in hardware architectures and the increase in data-intensive HPC applications, there is a need to simultaneously embrace a diverse set of resources (e.g., computing cores, cache, memory, I/O, and network resources) in the design of runtime schedulers for improving the overall application performance. This thesis performs an empirical evaluation of a recently proposed multi-resource scheduling algorithm for minimizing the overall completion time (or makespan) of computational workflows comprised of moldable parallel jobs. Moldable parallel jobs allow the scheduler to select the resource allocations at launch time and thus can adapt to the available system resources (as compared to rigid jobs) while staying easy to design and implement (as compared to malleable jobs). The algorithm was proven to have a worst-case approximation ratio that grows linearly with the number of resource types for moldable workflows. In this thesis, a comprehensive set of simulations is conducted to empirically evaluate the performance of the algorithm using synthetic workflows generated by DAGGEN and moldable jobs that exhibit different speedup profiles. The results show that the algorithm fares better than the theoretical bound predicts, and it consistently outperforms two baseline heuristics under a variety of parameter settings, illustrating its robust practical performance.


Bernaldo Luc

FPGA Implementation of an FFT-Based Carrier Frequency Estimation Algorithm

When & Where:


Eaton Hall, Room 2001B

Committee Members:

Erik Perrins, Chair
Morteza Hashemi
Rongqing Hui


Abstract

Carrier synchronization is an essential part of digital communication systems. In essence, carrier synchronization is the process of estimating and correcting any carrier phase and frequency differences between the transmitted and received signals. Typically, carrier synchronization is achieved using a phase lock loop (PLL) system; however, this method is unreliable when experiencing frequency offsets larger than 30 kHz. This thesis evaluates the FPGA implementation of a combined FFT and PLL-based carrier phase synchronization system. The algorithm includes non-data-aided, FFT-based, frequency estimator used to initialize a data-aided, PLL-based phase estimator. The frequency estimator algorithm employs a resource-efficient strategy of averaging several small FFTs instead of using one large FFT, which results in a rough estimate of the frequency offset. Since it is initialized with a rough frequency estimate, this hybrid design allows the PLL to start in a state close to frequency lock and focus mainly on phase synchronization. The results show that the algorithm demonstrates comparable performance, based on performance metrics such as bit-error rate (BER) and estimator error variance, to alternative frequency estimation strategies and simulation models. Moreover, the FFT-initialized PLL approach improves the frequency acquisition range of the PLL while achieving similar BER performance as the PLL-only system.


Rakshitha Vidhyashankar

An empirical study of temporal knowledge graph and link prediction using longitudinal editorial data

When & Where:


Eaton Hall, Room 2001B

Committee Members:

Zijun Yao, Chair
Prasad Kulkarni
Hongyang Sun


Abstract

Natural Language Processing (NLP) is an application of Machine Learning (ML) which focuses on deriving useful and underlying facts through the semantics in articles to automatically extract insights about how information can be pictured, presented, and interpreted.  Knowledge graphs, as a promising medium for carrying the structured linguistical piece, can be a desired target for learning and visualization through artificial neural networks, in order to identify the absent information and understand the hidden transitive relationship among them. In this study, we aim to construct Temporal Knowledge Graphs of sematic information to facilitate better visualization of editorial data. Further, A neural network-based approach for link prediction is carried out on the constructed knowledge graphs. This study uses news articles in English language, from New York Times (NYT) collected over a period of time for experiments. The sentences in these articles can be decomposed into Part-Of-Speech (POS) Tags to give a triple t = {sub, pred, obj}. A directed Graph G (V, E) is constructed using POS tags, such that the Set of Vertices is the grammatical constructs that appear in the sentence and the Set of Edges is the directed relation between the constructs. The main challenge that arises with knowledge graphs is the storage constraints that arise in lieu of storing the graph information. The study proposes ways by which this can be handled. Once these graphs are constructed, a neural architecture is trained to learn the graph embeddings which can be utilized to predict the potentially missing links which are transitive in nature. The results are evaluated using learning-to-rank metrics such Mean Reciprocal Rank (MRR). 


Jace Kline

A Framework for Assessing Decompiler Inference Accuracy of Source-Level Program Constructs

When & Where:


Eaton Hall, Room 2001B

Committee Members:

Prasad Kulkarni, Chair
Perry Alexander
Bo Luo


Abstract

Decompilation is the process of reverse engineering a binary program into an equivalent source code representation with the objective to recover high-level program constructs such as functions, variables, data types, and control flow mechanisms. Decompilation is applicable in many contexts, particularly for security analysts attempting to decipher the construction and behavior of malware samples. However, due to the loss of information during compilation, this process is naturally speculative and thus is prone to inaccuracy. This inherent speculation motivates the idea of an evaluation framework for decompilers.

In this work, we present a novel framework to quantitatively evaluate the inference accuracy of decompilers, regarding functions, variables, and data types. Within our framework, we develop a domain-specific language (DSL) for representing such program information from any "ground truth" or decompiler source. Using our DSL, we implement a strategy for comparing ground truth and decompiler representations of the same program. Subsequently, we extract and present insightful metrics illustrating the accuracy of decompiler inference regarding functions, variables, and data types, over a given set of benchmark programs. We leverage our framework to assess the correctness of the Ghidra decompiler when compared to ground truth information scraped from DWARF debugging information. We perform this assessment over a subset of the GNU Core Utilities (Coreutils) programs and discuss our findings.


Jaypal Singh

EvalIt: Skill Evaluation using block chain

When & Where:


Eaton Hall, Room 2001B

Committee Members:

Drew Davidson, Chair
David Johnson
Hongyang Sun


Abstract

Skills validation is a key issue when hiring workers. Companies and universities often face difficulties in determining an applicant's skills because certification of the skills claimed by an applicant is usually not readily verifiable and verification is costly. Also, from applicant's perspective, skill evaluation from industry expert is valuable instead of learning a generalized course with certification. Most of the certification programs are easy and proved not so fruitful in learning the required work skills. Blockchain has been proposed in the literature for functional verification and tamper-proof information storage in a decentralized way. "EvalIt" is a blockchain-based Dapp that addresses the above issues and guarantees some desirable properties. The Dapp facilitates skill evaluation efforts through payments using tokens that it collects from payments made by users of the platform.


Soma Pal

Properties of Profile-guided Compiler Optimization with GCC and LLVM

When & Where:


Eaton Hall, Room 2001B

Committee Members:

Prasad Kulkarni, Chair
Mohammad Alian
Tamzidul Hoque


Abstract

Profile-guided optimizations (PGO) are a class of sophisticated compiler transformations that employ information regarding the profile or execution time behavior of a program to improve program performance, typically speed. PGOs for popular language platforms, like C, C++, and Java, are generally regarded as a mature and mainstream technology and are supported by most standard compilers. Consequently, properties and characteristics of PGOs are assumed to be established and known but have rarely been systematically studied with multiple mainstream compilers.

The goal of this work is to explore and report some important properties of PGOs in mainstream compilers, specifically GCC and LLVM in this work. We study the performance delivered by PGOs at the program and function-level, impact of different execution profiles on PGO performance, and compare relative PGO benefit delivered by different mainstream compilers. We also built the experimental framework to conduct this research. We expect that our work will help focus future research and assist in building frameworks to field PGOs in actual systems.


Past Defense Notices

Dates

Anushka Bhattacharya

Predicting In-Season Soil Mineral Nitrogen in Corn Production Using Deep Learning Model

When & Where:


Nichols Hall, Room 246

Committee Members:

Taejoon Kim, Chair
Morteza Hashemi
Dorivar Ruiz Diaz


Abstract

One of the biggest challenges in nutrient management in corn (Zea mays) production is determining the amount of plant-available nitrogen (N) that will be supplied to the crop by the soil. Measuring a soil’s N-supplying power is quite difficult and approximations are often used in-lieu of intensive soil testing. This can lead to under/over-fertilization of crops, and in turn increased risk of crop N-deficiencies or environmental degradation. In this paper, we propose a deep learning algorithm to predict the inorganic-N content of the soil on a given day of the growing season. Since the historic data for inorganic nitrogen (IN) is scarce, deep learning has not yet been implemented in predicting fertilizer content. To overcome this hurdle, Generative Adversarial Network (GAN) is used to produce synthetic IN data and is trained using offline simulation data from the Decision Support System for Agrotechnology Transfer (DSSAT). Additionally, the time-series prediction problem is solved using long-short term memory (LSTM) neural networks. This model proves to be economical as it gives an estimate without the need for comprehensive soil testing, overcomes the issue of limited available data, and the accuracy makes it reliable for use.


Krushi Patel

Image Classification & Segmentation based on Enhanced CNN and Transformer Networks

When & Where:


Nichols Hall, Room 250 - Gemini Room

Committee Members:

Fengjun Li, Chair
Prasad Kulkarni
Bo Luo
Cuncong Zhong
Guanghui Wang

Abstract

Convolutional Neural Networks (CNNs) have significantly improved the performance on various computer vision tasks such as image recognition and segmentation based on their rich representation power. To enhance the performance of CNN, a self-attention module is embedded after each layer in the network. Recently proposed Transformer-based models achieve outstanding performance by employing a multi-head self-attention module as the main building block. However, several challenges still need to be addressed, such as (1) focusing only on class-specified limited channels in CNN; (2) limited respective field in the local transformer; and (3) addition of redundant features and lack of multi-scale features in U-Net type segmentation architecture.

In our work, we propose new strategies to address these issues. First, we propose a novel channel-based self-attention module to diversify the focus more on the discriminative and significant channels, and the module can be embedded at the end of any backbone network for image classification. Second, to limit the noise added by the shallow layers of an encoder in U-Net type architecture, we replaced the skip connections with the Adaptive Global Context Module (AGCM). In addition, we introduced the Semantic Feature Enhancement Module (SFEM) for multi-scale feature enhancement in polyp segmentation. Third, we propose a Multi-scaled Overlapped Attention (MOA) mechanism in the local transformer-based network for image classification to establish the long-range dependencies and initiate the neighborhood window communication.


Justinas Lialys

Parametrically resonant surface plasmon polaritons

When & Where:


2001B Eaton Hall

Committee Members:

Alessandro Salandrino, Chair
Kenneth Demarest
Shima Fardad
Rongqing Hui
Xinmai Yang

Abstract

The surface electromagnetic waves that propagate along a metal-dielectric or a metal-air interface are called surface plasmon polaritons (SPPs). These SPPs are advantageous in a broad range of applications, including in optical waveguides to increase the transmission rates of carrier waves, in near field optics to enhance the resolution beyond the diffraction limit, and in Raman spectroscopy to amplify the Raman signal. However, they have an inherent limitation:  as the tangential wavevector component of propagation is larger than what is permitted for the homogenous plane wave in the dielectric medium, this poses a phase-matching issue. In other words, the available spatial vector in the dielectric at a given frequency is smaller than what is required by SPP to be excited. The most commonly known technique to bypass this problem is by using the Otto and Kretschmann configurations. A glass prism is used to increase the available spatial vector in dielectric/air. Other methods are the evanescent field directional coupling, optical grating, localized scatterers, and coupling via highly focused beams. However, even with all these methods at our disposal, it is still challenging to couple SPPs that have a large propagation constant. 

As SPPs apply to a wide range of purposes, it is vitally important to overcome the SPP excitation dilemma. Presented here is a novel way to efficiently inject power into SPPs via temporal modulation of the dielectric adhered to the metal. In this configuration, the dielectric constant is modulated in time using an incident pump field. As a result of the induced changes in the dielectric constant, we show that efficient phase-matched coupling can be achieved even by a perpendicularly incident uniform plane wave. This novel method of exciting SPPs paves the way for further understanding and implementation of SPPs in a plethora of applications. For example, optical waveguides can be investigated under such excitation. Hence, this technique opens new possibilities in conventional plasmonics, as well as in the emerging field of nonlinear plasmonics. 


Andrei Elliott

Promise Land: Proving Correctness with Strongly Typed Javascript-Style Promises

When & Where:


Nichols Hall, Room 250, Gemini Room

Committee Members:

Matt Moore, Chair
Perry Alexander
Drew Davidson


Abstract

Code that can run asynchronously is important in a wide variety of situations, from user interfaces to communication over networks, to the use of concurrency for performance gains. One widely used method of specifying asynchronous control flow is the Promise model as used in Javascript. Promises are powerful, but can be confusing and hard-to-debug. This problem is exacerbated by Javascript’s permissive type system, where erroneous code is likely to fail silently, with values being implicitly coerced into unexpected types at runtime.

The present work implements Javascript-style Promises in Haskell, translating the model to a strongly typed framework where we can use the type system to rule out some classes of bugs.

Common errors – such as failure to call one of the callbacks of an executor, which would, in Javascript, leave the Promise in an eternally-pending deadlock state – can be detected for free by the type system at compile time and corrected without even needing to run the code.

We also demonstrate that Promises form a monad, providing a monad instance that allows code using Promises to be written using Haskell’s do notation.


Hoang Trong Mai

Design and Development of Multi-band and Ultra-wideband Antennas and Circuits for Ice and Snow Radar Measurements

When & Where:


Nichols Hall, Room 317

Committee Members:

Carl Leuschen, Chair
Fernando Rodriguez-Morales, Co-Chair
Christopher Allen


Abstract

Remote sensing based on radar technology has been successfully used for several decades as an effective tool of scientific discovery. A particular application of radar remote sensing instruments is the systematic monitoring of ice and snow masses in both hemispheres of the Earth. The operating requirements of these instruments are driven by factors such as science requirements and platform constraints, often necessitating the development of custom electronic components to enable the desired radar functionality.

This work focuses on component development and trade studies for two multichannel radar systems. First, this thesis presents the design and implementation of two dual-polarized ultra-wideband antennas for a ground-based dual-band ice penetrating radar. The first antenna operates at UHF (600–900 MHz) while the second antenna operates at VHF (140–215 MHz). Each antenna element is composed of two orthogonal octagon-shaped dipoles, two inter-locked printed circuit baluns and an impedance matching network for each polarization. Prototype of each band shows a VSWR of less than 2:1 at both polarizations over a fractional bandwidth exceeding 40%. The antennas developed offer cross-polarization isolation larger than 30 dB, an E-plane 3-dB beamwidth of 69 degrees, and a gain of at least 4 dBi with a variation of ± 1 dB across the bandwidth. This design with high power handling in mind also allows for straightforward adjustment of the antenna dimensions to meet other bandwidth constrains. It is being used as the basis for an airborne system.

Next, this work documents design details and measured performance of an improved and integrated x16 frequency multiplier system for an airborne snow-probing radar. This sub-system produces a 40 – 56 GHz linear frequency sweep from a 2.5 – 3.5 GHz chirp and mixes it down to the 2 – 18 GHz range.  The resulting chirp is used for transmission and analog de-chirping of the receive signal. The initial prototype developed through this work provided a higher level of integration and wider fractional bandwidth (>135%) compared to earlier versions implemented with the same frequency plan and a path to guide future realizations.

Lastly, this work documents a series of trade studies on antenna array configurations for both radar systems using electromagnetic simulation tools and measurements.


Xi Mo

Convolutional Neural Network in Pattern Recognition

When & Where:


Zoom Meeting, please contact jgrisafe@ku.edu for link.

Committee Members:

Cuncong Zhong, Chair
Taejoon Kim
Fengjun Li
Bo Luo
Hauzhen Fang

Abstract

Since convolutional neural network (CNN) was first implemented by Yann LeCun et al. in 1989, CNN and its variants have been widely implemented to numerous topics of pattern recognition, and have been considered as the most crucial techniques in the field of artificial intelligence and computer vision. This dissertation not only demonstrates the implementation aspect of CNN, but also lays emphasis on the methodology of neural network (NN) based classifier.

As known to many, one general pipeline of NN-based classifier can be recognized as three stages: pre-processing, inference by models, and post-processing. To demonstrate the importance of pre-processing techniques, this dissertation presents how to model actual problems in medical pattern recognition and image processing by introducing conceptual abstraction and fuzzification. In particular, a transformer on the basis of self-attention mechanism, namely beat-rhythm transformer, greatly benefits from correct R-peak detection results and conceptual fuzzification.

Recently proposed self-attention mechanism has been proven to be the top performer in the fields of computer vision and natural language processing. In spite of the pleasant accuracy and precision it has gained, it usually consumes huge computational resources to perform self-attention. Therefore, realtime global attention network is proposed to make a better trade-off between efficiency and performance for the task of image segmentation. To illustrate more on the stage of inference, we also propose models to detect polyps via Faster R-CNN - one of the most popular CNN-based 2D detectors, as well as a 3D object detection pipeline for regressing 3D bounding boxes from LiDAR points and stereo image pairs powered by CNN.

The goal for post-processing stage is to refine artifacts inferred by models. For the semantic segmentation task, the dilated continuous random field is proposed to be better fitted to CNN-based models than the widely implemented fully-connected continuous random field. Proposed approaches can be further integrated into a reinforcement learning architecture for robotics.


Sirisha Thippabhotla

An Integrated Approach for de novo Gene Prediction, Assembly and Biosynthetic Gene Cluster Discovery of Metagenomic Sequencing Data

When & Where:


Eaton Hall, Room 1

Committee Members:

Cuncong Zhong, Chair
Prasad Kulkarni
Fengjun Li
Zijun Yao
Liang Xu

Abstract

Metagenomics is the study of genomic content present in given microbial communities. Metagenomic functional analysis aims to quantify protein families and reconstruct metabolic pathways from the metagenome. It plays a central role in understanding the interaction between the microbial community and its host or environment. De novo functional analysis, which allows the discovery of novel protein families, remains challenging for high-complexity communities. There are currently three main approaches for recovering novel genes or proteins: de novo nucleotide assembly, gene calling, and peptide assembly. Unfortunately, their informational dependencies have been overlooked, and have been formulated as independent problems. 

In this work, we propose a novel de novo analysis pipeline that leverages these informational dependencies, to improve functional analysis of metagenomics data. Specifically, the pipeline will contain four novel modules: an assembly graph module, a graph-based gene calling module, a peptide assembly module, and a biosynthetic gene cluster (BGC) discovery module. The assembly graph module will be computational and memory efficient. It will be based on a combination of de Bruijn and string graphs. The assembly graphs contain important sequencing information, which can be further exploited to improve functional annotation. De novo gene-calling enables us to predict novel genes and protein sequences, that have not been previously characterized. We hypothesize that de novo gene calling can benefit from assembly graph structures, as they contain important start/stop codon information that provide stronger ORF signals. The assembly graph framework will be designed for both nucleotide and protein sequences. The resulting protein sequences from gene calling can be further assembled into longer protein contigs using our assembly framework. For the novel BGC module, the gene members of a BGC will be marked in the assembly graph. Finding a BGC can be achieved by identifying a path connecting its gene members in the assembly graph. Experimental results have shown that our proposed pipeline improved existing gene calling sensitivity on unassembled reads, achieving a 10-15% improvement in sensitivity over the state-of-the-art methods, at a high specificity (>90%). Our pipeline further allowed for more sensitive and accurate peptide assembly, recovering more reference proteins, delivering more hypothetical protein sequences.


Naveed Mahmud

Towards Complete Emulation of Quantum Algorithms using High-Performance Reconfigurable Computing

When & Where:


Zoom Meeting, please contact jgrisafe@ku.edu for link.

Committee Members:

Esam El-Araby, Chair
Perry Alexander
Prasad Kulkarni
Heechul Yun
Tyrone Duncan

Abstract

Quantum computing is a promising technology that can potentially demonstrate supremacy over classical computing in solving specific problems. At present, two critical challenges for quantum computing are quantum state decoherence, and low scalability of current quantum devices. Decoherence places constraints on realistic applicability of quantum algorithms as real-life applications usually require complex equivalent quantum circuits to be realized. For example, encoding classical data on quantum computers for solving I/O and data-intensive applications generally requires quantum circuits that violate decoherence constraints. In addition, current quantum devices are of small-scale having low quantum bit(qubit) counts, and often producing inaccurate or noisy measurements, which also impacts the realistic applicability of real-world quantum algorithms. Consequently, benchmarking of existing quantum algorithms and investigation of new applications are heavily dependent on classical simulations that use costly, resource-intensive computing platforms. Hardware-based emulation has been alternatively proposed as a more cost-effective and power-efficient approach. This work proposes a hardware-based emulation methodology for quantum algorithms, using cost-effective Field-Programmable Gate-Array(FPGA) technology. The proposed methodology consists of three components that are required for complete emulation of quantum algorithms; the first component models classical-to-quantum(C2Q) data encoding, the second emulates the behavior of quantum algorithms, and the third models the process of measuring the quantum state and extracting classical information, i.e., quantum-to-classical(Q2C) data decoding. The proposed emulation methodology is used to investigate and optimize methods for C2Q/Q2C data encoding/decoding, as well as several important quantum algorithms such as Quantum Fourier Transform(QFT), Quantum Haar Transform(QHT), and Quantum Grover’s Search(QGS). This work delivers contributions in terms of reducing complexities of quantum circuits, extending and optimizing quantum algorithms, and developing new quantum applications. For higher emulation performance and scalability of the framework, hardware design techniques and hardware architectural optimizations are investigated and proposed. The emulation architectures are designed and implemented on a high-performance-reconfigurable-computer(HPRC), and proposed quantum circuits are implemented on a state-of-the-art quantum processor. Experimental results show that the proposed hardware architectures enable emulation of quantum algorithms with higher scalability, higher accuracy, and higher throughput, compared to existing hardware-based emulators. As a case study, quantum image processing using multi-spectral images is considered for the experimental evaluations. 


Cecelia Horan

Open-Source Intelligence Investigations: Development and Application of Efficient Tools

When & Where:


2001B Eaton Hall

Committee Members:

Hossein Saiedian, Chair
Drew Davidson
Fengjun Li


Abstract

Open-source intelligence is a branch within cybercrime investigation that focuses on information collection and aggregation. Through this aggregation, investigators and analysts can analyze the data for connections relevant to the investigation. There are many tools that assist with information collection and aggregation. However, these often require enterprise licensing. A solution to enterprise licensed tools is using open-source tools to collect information, often by scraping websites. These tools provide useful information, but they provide a large number of disjointed reports. The framework we developed automates information collection, aggregates these reports, and generates one single graphical report. By using a graphical report, the time required for analysis is also reduced. This framework can be used for different investigations. We performed a case study regarding the performance of the framework with missing person case information. It showed a significant improvement in the time required for information collection and report analysis. 


Ishrak Haye

Invernet: An Adversarial Attack Framework to Infer Downstream Context Distribution Through Word Embedding Inversion

When & Where:


Nichols Hall, Room 246

Committee Members:

Bo Luo, Chair
Zijun Yao, Co-Chair
Alex Bardas
Fengjun Li

Abstract

Word embedding has become a popular form of data representation that is used to train deep neural networks in many natural
language processing tasks, such as Machine Translation, Question Answer Generation, Named Entity Recognition, Next
Word/Sentence Prediction etc. With embedding, each word is represented as a dense vector which captures its semantic relationship
with other words and can better empower Machine Learning models to achieve state-of-the-art performance.
However, due to the memory and time intensive nature of learning such word embeddings, transfer learning has emerged as a
common practice to warm start the training process. As a result, an efficient way is to initialize with pretrained word vectors and then
fine-tune those on downstream domain specific smaller datasets. This study aims to find whether we can infer the contextual
distribution (i.e., how words cooccur in a sentence driven by syntactic regularities) of the downstream datasets given that we have
access to the embeddings from both pre-training and fine-tuning processes.
In this work, we propose a focused sampling method along with a novel model inversion architecture “Invernet” to invert word
embeddings into the word-to-word context information of the fine-tuned dataset. We consider the popular word2Vec models
including CBOW, SkipGram, and GloVe based algorithms with various unsupervised settings. We conduct extensive experimental
study on two real-world news datasets: Antonio Gulli’s News Dataset from Hugging Face repository and a New York Times dataset
from both quantitative and qualitative perspectives. Results show that “Invernet” has been able to achieve an average F1 score of 0.75
and an average AUC score of 0.85 in an attack scenario.
A concerning pattern from our experiments reveal that embedding models that are generally considered superior in different tasks
tend to be more vulnerable to model inversion. Our results suggest that a significant amount of context distribution information from
the downstream dataset can potentially leak if an attacker gets access to the pretrained and fine-tuned word embeddings. As a result,
attacks using “Invernet” can jeopardize the privacy of the users whose data might have been used to fine-tune the word embedding
model.