19 WSN Faculty Fellows - Preview Page

Organic electronics — devices that replace heavy and brittle silicon with carbon-based molecules — are enabling a future in which televisions can be rolled up like posters, smart sensors can be placed on bridges for real-time structural health monitoring, and paper-thin solar panels can be affixed to rooftops with simple adhesive. These devices are characterized by their inexpensive manufacturability, adaptability, and compatibility with flexible plastic supports, features that promise applications not possible or prohibitively expensive using traditional silicon-based technologies.

Now imagine if these devices also could move and evolve on their own. Mobile and smart structural integrity sensors could scan surfaces for cracks or even heal themselves if damaged. Plantlike solar panels could spontaneously track the sun as it travels across the sky. These are just two examples out of endless possibilities that will be enabled by combining organic electronics with dynamic molecular crystals — materials that can bend, twist, crawl, jump and stretch. Together, Lee and Naumov aim to animate conductive molecular crystals, imparting mechanical motion to crystals that conduct charge, convert sunlight into electricity, and emit light.

Misinformation, and its targeted counterpart disinformation, has become a major concern in the open online world where everyone can contribute information that can be publicly accessed. This has only been aggravated by the recent advances in artificial intelligence (AI), machine learning, and Large Language Models (LLMs) that allow text and information to be created at scale without human effort and thus raise serious concerns about their harmful implications.

This project develops methodologies that aim to evaluate the trustworthiness of automatically generated text from large-scale language models, such as the one used in ChatGPT. We will investigate and suggest methods for evaluating the stability and factual correctness of LLM-generated text and investigate these questions across multiple AI-based systems. Our goal is to better understand the harms that can be caused and develop techniques for their early detection. We will also design measures and develop tools for a more trustworthy use of generative AI-based text tools.

In the aftermath of the Arab Spring, diverse political actors across the MENA region — from armed groups and clerics to activists and politicians — flocked to social media to organize, reach their constituencies, and amplify their causes. As these actors create content and interact with others, they leave behind digital traces of their strategic communication and the networks through which it spreads. Taking advantage of this rich record of online behavior, we plan to develop an infrastructure for tracking shifts in political alliances, including extremism, sectarianization, and polarization, in countries across the Middle East.

This will allow us to monitor emerging threats to regional stability both in the immediate aftermath of events on the ground and over long-range time horizons. To reach this goal, we plan to carry out the following three research objectives: 1) creating a database of tens of thousands of political, economic, religious, and cultural elites’ online communication, eventually across 28 countries of the Middle East, with this first phase of the project focusing on four countries as a proof-of-concept; 2) developing scalable approaches to mapping and measuring shifting polarization, sectarianism, and extremism among these actors and their followers over time; and 3) training the next generation of analysts of Arab social media data to improve and continue scaling our approach to real-time threat monitoring in the MENA region.