AI @ UMBC

UMBC has a long history of faculty in Computer Science and Information Systems working on research that would be classified as artificial intelligence (AI), dating back to the mid-1980s.

The UMBC Center for AI has more than 60 faculty members with research interests in AI and related areas, including robotics, machine learning, natural language understanding, data science, image processing, multi-agent systems, large language models, knowledge representation, as well as reasoning, planning, knowledge graphs, and neural networks. These faculty members work in more than 30 laboratories and research centers and teach many AI-related courses across departments and disciplines.

With the recent introduction of generative AI, we now see the opportunity for AI, specifically large-language models (LLM), to be used for research in many different ways: to improve research productivity in organizing documents, as a way to summarize qualitative data, or to be used in the development of research.

AI computation often relies upon the very specialized graphic processor units (GPUs).

The chip-gpu cluster consists of more than 20 distinct server computers or “nodes”. Each is equipped with NVidia GPU Cards from a variety of architectures (RTX 2080Ti, RTX 6000, RTX 8000, L40S, H100). The SIG-GPU subcommittee is a faculty governance group that determines the chip-cpu usage policies and provides advice to DoIT on evolving research needs. For more information on the advanced computing resources made available, visit the UMBC High Performance Computing Facility webpage.

In addition to the chip-gpu cluster, DoIT has purchased a machine with two H100 NVIDIA GPUs that is being configured to run Meta’s open-source Llama LLM. This LLM provides faculty with the opportunity to run LLM-based research without incurring costs when LLMs must be run programmatically.

 

Generative AI @ UMBC

Generative AI, also known as GenAI, is a type of artificial intelligence that can create new content like text, images, videos, and music.

GenAI can also learn from data and generate new data instances. Like any technology, generative AI offers both opportunities and risks to manage. UMBC is actively evaluating and exploring AI and its potential impact on teaching and learning, research and scholarship, administrative, and other functions within our community.

One of the critical issues with GenAI is that some tools use the information provided, such as documents, textual input, or other forms of content, as training data for the GenAI service. Unless this information is considered public material, something you would publish on a website for the Internet to see, you should not use content from UMBC on any GenAI service unless you know that UMBC has verified it is safe to use. Luckily, UMBC has access to several GenAI tools that have been verified as safe to use on UMBC Level 1 & FERPA data, which is data intended to be kept internal to UMBC.

Explore the World of Generative AI!

We have a new place to learn more about software – check out the UMBC Software Catalog! It’s your one-stop shop to easily search, discover, and access the latest Generative AI platforms supported for your work. Dive in to discover which AI platforms are approved, secure, and ready for you to use today.

Find Your AI Software

 

Academic and Instructional AI @ UMBC

UMBC offers three generative AI tools to faculty, staff and students:

The Blackboard AI Course Design Assistant tool offers suggestions for:

  • Generate a suggested course structure including learning module organization
  • Search or generate thumbnail images for learning modules and course banners
  • Generate questions for a test or question bank as well as rubrics
  • Generate prompts for assignments, discussions and journals.

The Blackboard AI Conversations tool provides opportunities for students to engage in Socratic or roleplaying scenarios with a virtual partner — designed and specified by the instructor.

 

Administrative AI Available @ UMBC

We have several GenAI Tools that employees and students can utilize for administrative tasks.

  • Google Gemini Advanced works well with our Google Workspace environment and can help people manage their tasks, calendar, and email.
  • Microsoft Copilot Pro is a powerful tool when integrated with Microsoft applications.
  • Amplify AI product provides employees with access to a variety of generative AI chat interfaces.

The myUMBC Answers service utilizes UMBC’s Amazon Web Services (AWS) environment to run our Portal, redefining the myUMBC Search function. The original search has been extended with Amazon’s Bedrock AI services, where you can ask questions, and myUMBC will provide answers. The initial focus has been on answering questions that students may have, including those specific to them.

Here are some examples of questions it can answer:

  • How many days are left on my parking permit?
  • What is my date that I can register for classes?
  • What is my GPA?
  • What funds are on my campus card?
  • Where is the career center?
  • What are the hours of the tutoring center?

myUMBC Answers utilizes information from the student handbook, some of UMBC’s authoritative websites, and your data that would be summarized on your myUMBC profile.  For staff who answer student questions, we encourage you to try the questions you receive from students in myUMBC Answers and provide us with feedback on how it works.

Over the summer, we will be working to extend this to better support faculty and staff by working with groups to add more content to our answers. Departments interested in working with DoIT on this should reach out to the myUMBC Team.

If you are someone who is responsible for answering RT student support tickets, we are interested in working with a few departments that have good student documentation or FAQs available to expand our testing of providing support staff with access to use GenAI services to provide answers inside RT. DoIT staff have been testing this with an RT enhancement defined by the RT queue to which the ticket is assigned. When enabled, RT will show the AI answer to the question and provide a way to send the AI answer as the RT reply to the ticket. We hope that, over time, as we improve documentation, this could help many overburdened support staff increase their productivity and have time to do other tasks.