Knowledge-infused learning (KIL) is a class of AI techniques that enhance machine learning models by integrating structured external knowledge, such as facts, rules, and relationships from various sources. The method aims to improve a model's ability to understand complex concepts, generalize from limited data, and provide interpretable decisions. By combining the pattern recognition capabilities of traditional data-driven models and large language models (LLMs) with the reasoning power of knowledge-based systems, KiL seeks to create more robust, capable, and understandable AI systems, with applications including natural language processing, robotics, and healthcare applications.

The UMBC Training Centers Center for Applied AI LunchBox Series will hold a session 12-12:30 pm ET on March 21, 2024 on the newest Google Gemini AI systems.

Introducing Google's latest LLM and Chatbot, Gemini Advanced: During a productive 30-minute lunch break, you will experience the power of Google's cutting-edge AI.

Discover enhanced text generation, image understanding, and seamless integration with Google Apps, including Maps, Gmail, and Drive.

Speaker: Liam Echlin 

Register here and get a link to the free online session

UMBC Center for Applied AI Lunchbox Series

We'll dig into data this Thursday with Insightful Data Analysis Through AI: A New Era of Understanding! Our presenter, Dhruv Ratra, is an experienced consultant and trainer in data analysis and visualization.

Together we'll take a look at how to upload and ask questions about your data, then let ChatGPT do the heavy lifting by turning complex information into clear, easy-to-understand insights.

Register for the Lunchbox Series here.

Details:

• Date: Thursday, March 7th
• Time: 12:00pm - 1:00pm ET
• Presenter: Dhruv Ratra
• Platform: Zoom (link in confirmation email after registration

2:15-3:15 pm ET March 5, 2024

Reasoning about real-life events is a unifying challenge in AI and NLP that has profound utility in a variety of domains, while any fallacy in high-stake applications like law, medicine, and science could be catastrophic. Able to work with diverse text in these domains, large language models (LLMs) have proven capable of answering questions and solving problems. In this talk, I demonstrate that end-to-end LLMs still systematically fail on reasoning tasks of complex events. Moreover, their black-box nature gives rise to little interpretability and user control. To address these issues, I propose two general approaches to use LLMs in conjunction with a structured representation of events. The first is a language-based representation involving relations of sub-events that can be learned by LLMs via fine-tuning. The second is a symbolic representation involving states of entities that can be leveraged by either LLMs or deterministic solvers. On a suite of event reasoning tasks, I show that both approaches outperform end-to-end LLMs in terms of performance and trustworthiness.

Li "Harry" Zhang is a 5th-year PhD student working on Natural Language Processing (NLP) and artificial intelligence at the University of Pennsylvania advised by Prof. Chris Callison-Burch. He earned his Bachelor's degree at the University of Michigan mentored by Prof. Rada Mihalcea and Prof.  Dragomir Radev. He has published more than 20 papers in NLP conferences that have been cited more than 1,000 times. He has reviewed more than 50 papers in those venues and has served as Session Chair and Program Chair in many conferences and workshops. Being a musician, producer, content creator of over 50,000 subscribers, he is also passionate in the research of AI music.