Karolina Dziugaite
Senior Research Scientist, Google Brain
Adjunct Professor, McGill University
PACBayes Meets Interactive Learning
Hawaii Convention Center, Honolulu, Hawaii
Meeting Room 314
28th July, 2023
Interactive learning encompasses online learning, continual learning, active learning, bandits, reinforcement learning, and other settings where an algorithm must learn while interacting with a continual stream of data. Such problems often involve explorationexploitation dilemmas, which can be elegantly handled with probabilistic and Bayesian methods. Deep interactive learning methods leveraging neural networks are typically used when the setting involves rich observations, such as images. As a result, both probabilistic and deep interactive learning methods are growing in popularity. However, acquiring observations in an interactive fashion with the environment can be costly. There is therefore great interest in understanding when sampleefficient learning with probabilistic and deep interactive learning can be expected or guaranteed. Within statistical learning theory, PACBayesian theory is designed for the analysis of probabilistic learning methods. It has recently been shown to be wellsuited for the analysis of deep learning methods. This workshop aims to bring together researchers from the broad Bayesian and interactive learning communities in order to foster the emergence of new ideas that could contribute to both theoretical and empirical advancement of PACBayesian theory in interactive learning settings.
Senior Research Scientist, Google Brain
Adjunct Professor, McGill University
Assistant Professor, Université Laval
Canada CIFAR AI Chair
Assistant Professor
Princeton University, USA
Assistant Professor
Carnegie Mellon University, Amazon
PhD Student
ETH Zürich, Switzerland
Time  Title  Speaker 

09:0009:15  Welcome  Organisers 
09:1510:00 
PACBayes Tutorial

Pascal Germain 
10:0010:30  Posters + break 1  
10:3011:15 
Invited Talk 1: Lessons Learned from Studying PACBayes and Generalization

Gintare Karolina Dziugaite 
11:1512:00 
Invited Talk 2: A unified recipe for deriving (timeuniform) PACBayes bound

Aaditya Ramdas 
12:0013:00  Lunch break  
13:0013:10 
Contributed Talk: Improved TimeUniform PACBayes Bounds using Coin Betting

Kyoungseok Jang 
13:1013:20 
Contributed Talk: PACBayesian Offline Contextual Bandits with Guarantees

Otmane Sakhi 
13:2013:30 
Contributed Talk: PACBayes bounds’ parameter optimization via events’ space discretization: new bounds for losses with general tail behaviors

Borja Rodríguez Gálvez 
13:3014:15 
Invited Talk 3: MetaLearning Reliable Priors for Interactive Learning

Jonas Rothfuss 
14:1515:00 
Invited Talk 4: Generalization Theory for Robot Learning

Anirudha Majumdar 
15:0015:30  Posters + break 2  
15:3015:40 
Contributed Talk: Anytime Model Selection in Linear Bandits

Parnian Kassraie 
15:4015:50 
Contributed Talk: PACBayesian Error Bound, via Rényi Divergence, for a Class of Linear TimeInvariant StateSpace Models

Deividas Eringis 
15:5016:00 
Contributed Talk: Experiment Planning with Function Approximation

Aldo Pacchiano 
16:0016:45  Panel discussion 
Canada CIFAR AI Chair
Assistant Professor, Université Laval
PhD Student, TU Darmstadt, Bosch Center for Artificial Intelligence
Associate Professor, University College London
Tenured Research Scientist, Inria
Turing Fellow, The Alan Turing Institute
PhD Student, Université Laval, Thales Group
Associate Professor, University of Southern Denmark
Please contact the organisers at: pacbayes.interactivelearning@gmail.com