Maria-Florina Balcan Gives Talk about Machine Learning

Yesterday, Maria-Florina Balcan, Carnegie Mellon University Associate Professor of Computer Science, gave a talk on her research titled “Machine Learning: New Challenges and Connections.” This talk is part of the Cyber Alliance Distinguished Speaker Series and was held at the Kilachand building.

In her presentation, she spoke machine learning’s far-reaching influences. Machine learning is now doing tasks for us more quickly and efficiently. For example, it can help detect what is spam email in a mailbox or can categorize many data points. She also noted that while machine learning is changing the world, our world is also changing machine learning. The high demand for vast amounts of complex data is calling for new learning protocols such as interactive learning, life-long learning, and multi-agent learning. In this fast-paced and sophisticated society. there is a need for Machine Learning to learn more complex objects and process information even faster.

Maria-Florina went on to detail a new learning approach: interactive machine learning. The goal of this approach is to be able to take massive amounts of raw data, use minimal labeled examples, and have the ability to pick out specific, related data to whatever research at hand. While interactive machine learning is very useful in practice, the downside is that as time goes on, sampling bias can occur.

Before closing up her talk, she spoke of the usefulness of learning algorithms for solving combinatorial problems. Algorithms can be applied in various fields, for example in revenue management. A number of items for sale with a number of buyers can be inputted to achieve an assigned price as the output in order to maximize revenue.

This event was a huge success as Maria-Florina delivered her talk in a completely packed auditorium with an actively engaging audience eager to learn more about the growth of machine learning.