STEM Fellows

Research Presentation: Two-Step Verification System Using Face Recognition

Presented by: Mukilan Narayanamoorthy — M.S. in Cybersecurity

Abstract: As technology use increases rapidly across the globe, the number of active social media users has reached 3.7 billion within just the past three years. This increasing demand urges social network companies to introduce more and better security mechanisms to protect the authenticity of the user. Several companies have introduced two-step verification mechanisms, like verification codes and graphical passwords, but these can easily be compromised with simple techniques, like SIM hijacking. To identify a more secure environment, we researched three approaches—aHash, pHash and dHash—to determine which offers the greatest protection of authenticity. Our findings suggest that dHash is the best approach, and one can use the combination of facial recognition and cryptography to safely ensure the authenticity of the user.

Research Presentation: Speech Emotion Recognition With Generative Data

Presented by: Benjamin Cohen — M.S. in Data Analytics and Visualization

Abstract: Traditionally, audio classification has been studied by manually selecting features from spectrograms. In the past few years, however, a lot of interest has been raised about how deep learning can make this process easier. This project applies deep-learning methods to speech-emotion recognition (the ability to discern a person’s emotions solely from their tone of voice). To do this, we gathered a base dataset of around 1,400 audio clips, labeled as angry or not angry, augmented the dataset using a generative model and then used transfer learning with a pre-trained Yamnet model to predict whether the clip was angry or not, ultimately achieving accuracy of greater than 90 percent. This work is important because it shows that deep-learning approaches work very well with audio data and that not much data is needed because it's possible to augment datasets.

Research Presentation: Strategies for the Development and Commercialization of an Antibody-Drug Conjugate as a New Therapy for Solid Tumors of Hypopharyngeal Cancer

Presented by: Anam Khalid — M.S. in Biotechnology Management and Entrepreneurship

Abstract: Grit Bio seeks to transform healthcare for oncology patients by developing safer cancer therapeutics. Grit Bio is focused on treating squamous cell carcinoma (SCC), with a priority on a rare type of throat cancer, hypopharyngeal cancer, which is typically diagnosed at a late stage and has a very poor prognosis. Grit Bio’s strategy is to target dormant cancer cells to prevent recurrent cancer. The company is developing two products, a peptide inhibitor and an antibody-drug conjugate (ADC), both targeting a specific metabolic process in dormant cancer cells. This project focuses on one product, the novel ADC (LIGRECA, pronounced lig-re-sa). The project aims to assist the company in understanding the market potential and commercialization feasibility of LIGRECA via market analysis and risk assessment and proposes a risk-mitigation strategy for the successful development of new ADC as new therapy for solid tumors, based on the market of hypopharyngeal and exemplified by two FDA approved ADCs used for solid tumors (PADCEV and KADCYLA TDM-1).

Research Presentation: Force-Indentation Curves Of Spheroidal Objects

Presented by: Benjamin Goykadosh — M.A. in Physics

Abstract: Currently, there are many methods to determine the forces and energy necessary to deform cells. These methods, however, are computationally slow and require intensive testing outside of the body. In this project, we propose a theoretical method to recover the energy and force versus indentation curves produced by the deformation using an Atomic Force Microscope, which is necessary to understand cell contraction and deformation. The theory depends on parameters, such as Young’s modulus, object length and Poisson’s ratio, and is obtained by fitting our mathematical expressions to experimental force versus indentation curves. Thus, our results provide a systematic way to measure those material parameters, in general, but particularly in a soft matter where the materials are highly heterogeneous and where their properties are often dependent on external stresses.

Research Presentation: Adbot: Integrated Communications Go To Market Plan

Presented by: Hannah Kraitberg, Tomer Mendler, Linyu Zhen, Panxinyue Zhang, and Jingyuan Wang — M.S in Digital Marketing and Media

Abstract: In this collaborative project, the Katz School's digital marketing and media team developed a comprehensive brand and marketing strategy for Adbot, a startup company aiming to change the way YouTube creators monetize. The resulting strategy leverages deep-customer and competitor insights, as well as the company’s business purpose and differentiators. The strategy also combines all relevant marketing channels. It provides a roadmap to steer Adbot’s successful brand development and a clear framework for reaching Adbot’s target audience, including details on planned activity and the relevant methods of communication.

Research Presentation: A Proposed Novel Two-Drug Combination Topical Treatment of Actinic Keratosis

Presented by: Jonathan Taub — M.S. in Biotechnology Management & Entrepreneurship

Abstract: Actinic keratosis (AK) and cutaneous squamous cell carcinoma (cSCC) incidences have increased exponentially over the last half-century, even as methods of diagnosis and diagnostic definitions themselves remain malleable. An estimated two-thirds of all AK and cSCC patients are over the age of 65, possibly presenting a public health concern. One group of AK and cSCC sufferers that might find themselves with more restricted options in a crowded therapeutic field are immunocompromised patients, particularly patients with transplants. GritBIO proposes a combination of two drugs—miltefosine and carmofur—in a novel composition for topical administration, which would provide the requisite cytotoxicity without the attendant adverse effects on immune cells. This project aims to understand the market, IP constrains, regulatory requirements and the potential for fast-track designation for this treatment.

Research Presentation: Capture the Flag

Presented by: Kevin Suckiel — M.S. in Cybersecurity

Research Presentation: In the classroom, technical skills can sometimes be overshadowed by theoretical knowledge. Capture the Flag (CTF) events remove those barriers by allowing students to use any tools in their arsenal to complete challenges related to the subject matter. CTF challenges range in difficulty and give students a snapshot of problems they may face in the real world. To test the effectiveness of CTF events in facilitating cybersecurity skills development, 40 students were asked to participate in CTF events across several categories. The outcome, based on student feedback, suggests that students learned more when engaged with CTF events and felt better prepared to handle real-life cybersecurity opportunities and challenges.

Research Presentation: Financial Impact of Tropical Cyclones on U.S. Real Estate Sector

Presented by: Atreish V. Ramlakhan, Aishwarya Singh, and Brian Livian — M.S. in Artificial Intelligence and M.A. in Mathematics

Abstract: This project seeks to quantify the financial effects of tropical cyclones on the U.S. real estate market. The project is a collaboration between students in the Artificial Intelligence and Mathematics programs at the Katz School of Science and Health and S&P Global Market Intelligence. Using hurricane data from the National Oceanic and Atmospheric Administration (NOAA) and the Intergovernmental Panel on Climate Change (IPCC), as well as proprietary real estate trust data from S&P databases, we analyzed the growing impact of tropical cyclones in the United States to show the relationship between increased hurricane frequency due to climate change and the financial losses of Real Estate Investment Trusts (REIT).

Research Presentation: Regulation of Testicular Sertoli Cell by SUMOylation

Presented by: Shanza Baseer and Manveet Singh Nanda — M.S. in Biotechnology Management and Entrepreneurship

Abstract: SUMO proteins are essential for normal development and are expressed in mouse and human Sertoli cells. However, the cell-specific role of sumoylation in those cells has only started to be elucidated. In other cell types, including granulosa cells, sumoylation is regulated by a SUMO ligase KAP1/Trim28. Deletion of KAP1 in Sertoli cells causes testicular degeneration, but the role of KAP1 in those cells has not been identified. In addition, the molecular regulation of Sertoli cells and their crosstalk with germ cells has not been fully characterized. Here, we show that both mouse and human Sertoli undergo apoptosis upon inhibition of sumoylation with a chemical inhibitor or via a siRNA technology. We have additionally detected changes in the Sertoli cell proteome upon the inhibition of sumoylation, and our data suggest that among others, the expression of ER/stress-related proteins is highly affected by this inhibition. Sumoylation may also regulate the NOTCH signaling which is important for the maintenance of the developing germ cells. Furthermore, we show that a siRNA-down-regulation of KAP1 in a Sertoli-derived cell line causes an almost complete inactivation of sumoylation. In conclusion, sumoylation regulates important survival and signaling pathways in Sertoli cells, and KAP1 can be a major regulator of sumoylation in these cells.