Professors and students from The Machine Learning Center at Georgia Tech (ML@GT) are kicking off the New Year presenting some of their latest research at the 33rd AAAI Conference on Artificial Intelligence (AAAI-19) in Honolulu, Hawaii, Jan. 27 through Feb. 1.
Focused solely on artificial intelligence, the conference brings together more than 2,000 artificial intelligence (AI) researchers from academia and industry.
One of the highlights of the conference for Georgia Tech will be the recognition of Charles Isbell as a 2019 AAAI Fellow. Isbell, Associate Executive Dean for the College of Computing, is being recognized for his more than two decades of significant and sustained technical contributions to the field of AI.
Also well known for his contributions to AI, ML@GT’s Ashok Goel is one of the conference’s invited speakers and will be discussing Experiments in Teaching AI. In his talk, Goel will present several experiments on teaching cognitive systems in online and blended learning settings. Goel – a professor in the School of Interactive Computing at Georgia Tech – will also share results and draw out some general principles for teaching AI, as well as using AI to teach AI.
AAAI-19 will also feature five Georgia Tech-led research papers. According to the conference website, accepted papers touch on a variety of topics within the field including natural language processing, robotics, deep learning, and knowledge representation, and can be applied to transportation, commerce, sustainability, healthcare, and other important industries.
Georgia Tech’s five papers are:
- Composable Modular Reinforcement Learning
- Understanding Story Characters, Movie Actors and Their Versatility with Gaussian Representations
- Revisiting Projection-Free Optimization For Strongly Convex Constraint Sets
- A Novel Framework for Robustness Analysis of Visual QA Models
- GAMENet: Graph Augmented MEmory Networks for Recommending Medication Combination
The Knowledge Extraction from Games workshop taking place on Jan. 27 was organized by Georgia Tech Computer Science Ph.D. candidate Matthew Guzdial and his peers from Pomona College and Drexel University. The workshop will explore approaches and questions to the automated extraction of design elements, music, character graphics, and other “knowledge” from games.