ECE Seminar: Anastasia Angelopoulou

ECE Seminar: Anastasia Angelopoulou

Join Zoom Meeting: https://bostonu.zoom.us/j/97952794519?pwd=lPKABGJhgzYbRf0nHK4vtjviK1EZhg.1&from=addon

Meeting agenda: https://docs.zoom.us/agenda/doc/55a1c7bb-d45c-4409-a1f3-ed49280a9888?from=addon

Meeting ID: 979 5279 4519

Passcode: 678634

Title: Embeddings: The Backbone of Modern AI Systems

Abstract: This lecture will explore how modern AI systems move beyond keyword matching to capture meaning through word embeddings. We will begin with traditional approaches such as Bag-of-Words and TF-IDF, then transition to semantic representations that model relationships between words in vector space.

We will introduce core concepts including the distributional hypothesis, cosine similarity, and models such as Word2Vec, and discuss the evolution toward contextual embeddings in transformer-based architectures. Through examples and real-world applications, we will demonstrate how embeddings support tasks such as semantic search and sentiment analysis. The lecture will also discuss key limitations of embedding-based models.

Bio: Dr. Anastasia Angelopoulou holds a Ph.D. in Modeling and Simulation from the University of Central Florida. She is currently a Visiting Associate Professor of Business Intelligence and Analytics Management at DeVry University and has industry experience as a Senior Data Scientist II at Shipt (Target Corporation). Her research focuses on natural language processing and simulation, with applications in educational data mining, accessibility, and social media analytics.