SiG-DML-L1.5
When Can Sequence Modelling Approaches Recover the Target Policy In Offline Reinforcement Learning? a Statistical Analysis
abdelghani ghanem, International University of Rabat, Morocco; Philippe Ciblat, Telecom Paris, Institut Polytechnique de Paris, France; Mounir Ghogho, University Mohammed VI Polytechnic, Morocco
Session:
SiG-DML-L1: Sequential & Neural Network Learning Lecture
Track:
SiG-DML - Signal and Data Analytics for Machine Learning
Location:
Sala Massimo
Presentation Time:
Tue, 9 Sep, 15:20 - 15:40 Italy Time (UTC +2)
Session Co-Chairs:
Lucio Marcenaro, University of Genova and Diego Valsesia,
Presentation
Discussion
Resources
No resources available.
Session SiG-DML-L1
SiG-DML-L1.1: RIEMANNIAN Q-FUNCTIONS FOR POLICY ITERATION IN REINFORCEMENT LEARNING
Minh Vu, Konstantinos Slavakis, Institute of Science Tokyo, Japan
SiG-DML-L1.2: LEVERAGING DYNAMIC INTERACTION MODELS FOR ANOMALOUS BEHAVIOR DETECTION IN 3D ENVIRONMENTS
Saleemullah Memon, Ali Krayani, Pamela Zontone, Lucio Marcenaro, University of Genova, Italy; David Martin Gomez, University Carlos III de Madrid, Spain; Carlo Regazzoni, University of Genova, Italy
SiG-DML-L1.3: TRANSFORMER ENCODER AND MULTI-FEATURES TIME2VEC FOR FINANCIAL PREDICTION
Nguyen Kim Hai Bui, Chien Nguyen Duy, Péter Kovács, Gergő Bognár, ELTE Eötvös Loránd University, Hungary
SiG-DML-L1.4: Early Exiting Predictive Coding Neural Networks for Edge AI
Alaa Zniber, International University of Rabat, Morocco; Mounir Ghogho, University Mohammed VI Polytechnic, Morocco; Ouassim Karrakchou, Mehdi Zakroum, International University of Rabat, Morocco
SiG-DML-L1.5: When Can Sequence Modelling Approaches Recover the Target Policy In Offline Reinforcement Learning? a Statistical Analysis
abdelghani ghanem, International University of Rabat, Morocco; Philippe Ciblat, Telecom Paris, Institut Polytechnique de Paris, France; Mounir Ghogho, University Mohammed VI Polytechnic, Morocco