Nripsuta Saxena

USC Viterbi School of Engineering

Department of Computer Science

Fifth year Ph.D. student

Advisor: Prof. Cyrus Shahabi

Thesis: Towards a fairer society

3MT Presentation: Algorithms for a fairer world

Abstract: Algorithms are increasingly being used for decision-making in high-impact domains with an almost tangible impact on people: hiring, healthcare, and credit applications are just some examples. Trained on data collected from the real world, these algorithms also pick up on the biases that get encoded in data. While ideally algorithms should be completely fair, that is a complex task (currently there is no consensus on what it means to be "fair" in different situations). My research aims to design fairer algorithms by placing constraints so that the outcomes are more fair than they are currently.

Mustafa Can Gursoy

USC Viterbi School of Engineering

Department of Electrical and Computer Engineering

Fourth year Ph.D. student

Advisor: Prof. Urbashi Mitra

Thesis: Engineering Nature: Bio-Inspired Molecular Communications

3MT Presentation: Engineering Nature: Bio-Inspired Molecular Communications

Abstract: Every living thing interacts with its environment and other members of its community. For bacteria, these interactions occur through exchanging chemicals. Thus, considering the key role of bacteria in human biology (e.g., gut microbiome) and the ecosystem in general, engineering bacterial communication can enable a vast range of applications: from biomedical systems to new energy sources. However, chemical communication is a slow and unreliable process, which hinders these goals. My research finds unique solutions to these problems by understanding the fundamentals of chemical signaling in microbial communities, as well as using the insight to propose more efficient signaling strategies.

Heramb Nemlekar

USC Viterbi School of Engineering

Department of Computer Science

Fourth year Ph.D. student

Advisor:  Prof. Stefanos Nikolaidis

Thesis: Efficiently learning human preferences for proactive robot assistance in assembly tasks 

3MT Presentation: Efficiently learning human preferences for personalized robot assistance 

Abstract: Robots that support humans in assembly tasks need to adapt efficiently to the preferences of their users. While prior work has focused on learning user preferences from explicit demonstrations, obtaining this data can be costly in real-world assemblies. Thus, I propose leveraging prior knowledge of (i) similarities in the preferences of different users in a given task and (ii) similarities in the preferences of a given user in different tasks for efficient robot adaptation. Through user studies of an IKEA bookcase and model airplane assembly, I show how my proposed approach enables robots to proactively and accurately assist humans.

Mahmoud Abdelrahman

USC Viterbi School of Engineering

Department of Electrical Engineering

Third year Ph.D. student

Advisor: Prof. Demetrios Christodoulides

Thesis: A thermodynamic description of highly multimoded nonlinear optical systems

3MT Presentation: Thermodynamic theory of highly multimoded nonlinear optical systems

Abstract: Recently, there has been a resurgence of interest in optical wave propagation in multimoded nonlinear photonic configurations. What has incited a flurry of activities in this field is that the sheer complexity associated with such multimode systems, allows them to display a plethora of intriguing phenomena that have no counterparts in single-mode settings. Meanwhile, it has led to new theoretical challenges in understanding their collective dynamics. Here, we present a self-consistent optical thermodynamic theory to predict the convoluted behavior of these structures as well as unravel a host of novel applications such as all-optical refrigeration via Carnot cycles.

Ahmed A. Alzanki

USC Viterbi School of Engineering

Department of Industrial and Systems Engineering

Seventh year Ph.D. student

Advisor: Prof. Ali E. Abbas

Thesis: Examining Methods of Assessing Trade-offs in Life and Death Decisions

3MT Presentation: Examining Methods of Assessing Trade-offs in Life and Death Decisions

Abstract: Decisions involving trade-offs between life, death, and money are being faced every day by policymakers, economists, and even individuals. Examples include government imposing stay-at-home decisions during a pandemic. Because of the severity of the monetary and safety consequences in these decisions, a proper assessment of trade-offs is necessary to make better decisions. This dissertation investigates several methods for assessing trade-offs in life and death decisions theoretically and practically through a behavioral experiment to provide decision-makers guidance for structuring their trade-off assessments. The investigation is based on individuals’ preferences, complexity, ease of use, and consistency in the estimation of the trade-offs.

Kazem Bakian Dogaheh

USC Viterbi School of Engineering

Department of Electrical and Computer Engineering

Sixth year Ph.D. student

Advisor: Prof. Mahta Moghaddam

Thesis: Microwave Instrumentation and Models for Studying the Subsurface Processes in Arctic Permafrost

3MT Presentation: Microwave Instrumentation and Models for Studying the Subsurface Processes in Arctic Permafrost

Abstract: Rising temperatures in the Arctic have raised several questions about the vulnerability of the Arctic permafrost ecosystem under a changing climate. To this end, NASA runs a decadal field campaign that integrates various in-situ and airborne science to achieve a better insight into Arctic ecosystem dynamics. The water and carbon cycles are the key aspects in this study and are not well understood in an era of rapid Arctic change. My dissertation is focused on developing microwave instrumentation and modeling approaches to help unravel the carbon-water cycle dynamics of permafrost soils and their influence on climate-driven vulnerabilities of the Arctic.

Arturo Cajal

USC Viterbi School of Engineering

Department of Aerospace and Mechanical Engineering

Sixth year Ph.D. student

Advisor: Prof. Alejandra Uranga

Thesis: Modeling Laminar to Turbulent Transition in RANS Simulations

3MT Presentation: Reduction of fuel burn and aircraft emission

Abstract: There is a need to develop environmentally sustainable aviation solutions, and improving the efficiency of aircraft systems is an effective way towards a cleaner future. Reducing drag on the aerodynamic surfaces of an aircraft is an effective way to reduce fuel burn and emissions. This work focuses on assessing and developing computational models to delay the onset of turbulent flow, having a higher drag when compared to a laminar flow.

Patrick Gelbach

USC Viterbi School of Engineering

Department of Biomedical Engineering

Fifth year Ph.D. student

Advisor: Prof. Stacey Finley

Thesis: Genome-scale Metabolic Modeling of Macrophages in Colorectal Cancer

3MT Presentation: Mathematical Modeling of Immune Cells in Colon Cancer

Abstract: Part of the difficulty in treating colon cancer is the presence of many different types of cells surrounding the tumor. Macrophages are one of those types of cells, and can take on a range of behaviors, either supporting or fighting the cancer. Macrophages have a highly specialized way of making and using energy (metabolism), which we have studied with mathematical approaches. We have developed and analyzed computational models characterizing metabolic differences between macrophage subgroups, have analyzed the range of behaviors that they can show, and have modeled interactions between different cell types, aimed at finding novel ways to combat cancer.

Haotian Hang

USC Viterbi School of Engineering

Department of Aerospace and Mechanical Engineering

Third year Ph.D. student

Advisor: Prof. Eva Kanso

3MT® Presentation: Control with Flow

Thesis: Control with Flow: Flow Sensing and Swimming Models Inspired by Fish

Abstract: Underwater navigation is challenging because visual and radio are disabled at deep ocean. However, structures in flow field are informative for efficient navigation. Fishes have fascinating ability in sensing and responding flow field around them. We are considering hydrodynamic trail following inspired by predation and schooling. We employed reinforcement learning to discover a sensory control strategy based on only local and instantaneous sensory cue. Moreover, we interpreted this strategy, proved its stability, and generalized it to different scenarios which are unseen during training. These works enrich our understanding biological behaviors and provide intuition for designing underwater robots.

Yuanzhong Pan

USC Viterbi School of Engineering

Department of Biomedical Engineering

Fifth year Ph.D. student

Advisor:  Prof. Steve Kay

Thesis: Pharmacological targeting of circadian rhythm genes in cancer

3MT Presentation: Fixing the twisted clock in cancer

Abstract: Normal cells in the body have intrinsic twenty-four-hour rhythm that is called the circadian clock. Some molecules called transcription factors in the cell nucleus maintain this rhythms. Cancer cells can hijack these molecules and twist their function to fuel their growth. We designed drugs that can stop the twisted working pattern of the hijacked transcription factors and therefore kill the cancer cells.

Powen Yao

USC Viterbi School of Engineering

Department of Computer Science

Eighth year Ph.D. student

Advisor: Prof. Michael Zyda

Thesis: Virtual Equipment System - Designing with Hyperphysical User Interface and Whole-Body Interaction in Extended Reality

3MT Presentation: Designing a Virtual Equipment System for Virtual Reality

Abstract: This thesis is about the design of a virtual equipment system in extended reality.

Two concepts are used as lenses for design. Hyper-physicality user interfaces and Whole-Body Interaction. Given that anything can be possible in extended reality, the user interface can follow rules of physics different from the real world. At the same time, the user is still very much limited by their body and thus we must consider what the user can or cannot do in the real world.

The thesis explores having objects in multiple dimensions, using gestures to interact with the equipment, and redefining what constitutes equipment.