UMBC ebiquity research group

05/16/2025

https://ai.umbc.edu/news/post/150206/
UMBC Prof. Anupam Joshi was interviewed by aecc india on the evolving roles of AI, data science, machine learning, and cybersecurity, their impact on jobs and careers in many fields, and opportunities to master them at UMBC.

03/03/2025

Dr. Virginia Dignum discusses the often contradictory nature of AI, exploring how its advancements highlight the irreplaceable qualities of human intelligence and the importance of governance. She will use key paradoxes, such as the Agreement Paradox, which questions why the more we discuss AI, the less we seem to agree on what it is. We'll also examine the Intelligence Paradox, revealing how AI's capabilities underscore what makes human intelligence unique. Furthermore, she'll tackle the Justice Paradox, addressing the challenge of achieving true fairness with AI, and the Regulation Paradox, which focuses on balancing innovation and oversight in the AI era. All in all, an exploration of how paradoxes can help us uncover how AI shapes our world and how we can ensure it serves humanity ethically and equitably. https://ai.umbc.edu/news/post/147750/

01/28/2025

https://ai.umbc.edu/news/post/146793/
Shan Huang (UIUC) talks about her ongoing work evaluating how well today’s LLM systems understand and can reason about cybersecurity concepts using the CCI and CCA sets of evaluation problems. Her talk will be online from 12-1pm ET on Friday, January 31.

10/23/2024

Kevin Coakley from UC San Diego talks on ML Reproducibility: Sources of Algorithmic, Implementation, and Observational Variability, 4-5pm EDT Tues., 29 Oct. 2024, online. Achieving reproducibility is difficult in 𝐦𝐚𝐜𝐡𝐢𝐧𝐞 𝐥𝐞𝐚𝐫𝐧𝐢𝐧𝐠 due to variability in algorithms, implementations, and observational factors. This talk explores key contributors to irreproducibility in ML, including algorithmic factors like hyperparameter tuning and random weight initialization, implementation differences in software and hardware, and observational factors such as dataset bias and data preprocessing. Sponsored by the iHARP Institute.
https://ai.umbc.edu/news/post/144995/

04/24/2024

New paper: Exploring the Impact of Increased Health Information Accessibility in Cyberspace on Trust and Self-care Practices, ACM SaT-CPS: ACM Workshop on Secure and Trustworthy Cyber-Physical Systems, June 2024. https://ebiquity.umbc.edu/paper/html/id/1176/

04/07/2024

UMBC's Anupam Joshi will discuss how opinions can be shaped through narrative construction to influence societies & spread mis/disinformation. He will also show how new AI technologies like LLMs can detect this. 12-1 pm ET Friday, 12 April 2024 via WebEx.

Dr. Anupam Joshi on misinformation, disinformation & LLMs The UMBC Cyber Defense Lab presentsShaping Opinion and Influencing Societies through Narrative ConstructionAnupam JoshiCSEE Professor and Acting CoEIT Dean12–1pm, Friday, April 12, 2024 via WebExThere has been a significant body of work tha...

03/19/2024

New paper on GenAIPABench, a benchmark for evaluating Generative AI-based Privacy Assistants.https://ebiquity.umbc.edu/paper/html/id/1173/ It includes 1) A set of curated questions about privacy policies with annotated answers for various organizations and regulations; 2) Metrics to assess response accuracy, relevance, & consistency; 3) A tool to generate prompts to introduce privacy policies and paraphrased variants of the curated questions. The system was evaluated using three leading systems, ChatGPT-4, Bard, and Bing AI.

03/18/2024

A new paper on how well generative AI systems like ChatGPT-4, Bard, and Bing AI understand today's privacy policies, GenAIPABench: A Benchmark for Generative AI-based Privacy Assistants, will appear in Proceedings on Privacy Enhancing Technologies. https://ebiquity.umbc.edu/paper/html/i

02/15/2024

Enhancing Knowledge Graph Consistency through Open Large Language Models: A Case Study is an AAAI-MAKE paper on the results of a study of how well different LLM models can detect inconsistencies between facts extracted from text and the text from which they were extracted. https://ebiquity.umbc.edu/paper/html/id/1169/Enhancing-Knowledge-Graph-Consistency-through-Open-Large-Language-Models-A-Case-Study

02/06/2024

UMBC Cyber Defense Lab presents Anonymized Data Can Still Tell Tales: 2024 UMBC SFS Research Study Christian Badolato, CSEE PhD Student 12-1pm ET Fri, 9 Feb. 2024 via WebEx Joint...