CoLLAs

Accepted Papers

[P1] Few-Shot Learning by Dimensionality Reduction in Gradient Space

Martin Gauch (Johannes Kepler University Linz), Maximilian Beck (Johannes Kepler University Linz), Thomas Adler (Johannes Kepler University Linz), Dmytro Kotsur (Anyline GmbH), Stefan Fiel (Anyline GmbH), Hamid Eghbal-zadeh (Johannes Kepler University Linz), Johannes Brandstetter (Johannes Kepler University Linz), Johannes Kofler (Johannes Kepler University Linz), Markus Holzleitner (Johannes Kepler University Linz), Werner Zellinger (Austrian Academy of Sciences, Linz), Daniel Klotz (Johannes Kepler University Linz), Sepp Hochreiter (Johannes Kepler University Linz), Sebastian Lehner (Johannes Kepler University Linz)

[P2] Meta-Gradients in Non-Stationary Environments

Jelena Luketina (University of Oxford), Sebastian Flennerhag (DeepMind), Yannick Schroecker (DeepMind), David Abel (DeepMind), Tom Zahavy (DeepMind), Satinder Singh (DeepMind)

[P3] Neural Distillation as a State Representation Bottleneck in Reinforcement Learning

Valentin Guillet (ISAE-Supaero, Université de Toulouse), Dennis Wilson (ISAE-Supaero, Université de Toulouse), Carlos Aguilar-Melchor (Sandbox AQ), Emmanuel Rachelson (ISAE-Supaero, Université de Toulouse)

[P4] CORA: Benchmarks, Baselines, and Metrics as a Platform for Continual Reinforcement Learning Agents

Sam Powers* (Carnegie Mellon University), Eliot Xing* (Georgia Institute of Technology), Eric Kolve (Allen Institute for AI), Roozbeh Mottaghi (Allen Institute for AI), Abhinav Gupta (Carnegie Mellon University)

[P5] Test Sample Accuracy Scales with Training Sample Density in Neural Networks

Xu Ji (Mila), Razvan Pascanu (DeepMind), Devon Hjelm (Mila, MSR), Balaji Lakshminarayanan (Google), Andrea Vedaldi (Oxford University)

[P6] What Should I Know? Using Meta-Gradient Descent for Predictive Feature Discovery in a Single Stream of Experience

Alex Kearney (University of Alberta), Anna Koop (University of Alberta), Johannes Günther (University of Alberta, Alberta Machine Intelligence Institute), Patrick Pilarski (University of Alberta, Alberta Machine Intelligence Institute)

[P7] Inherent Limitations of Multi-Task Fair Representations

Tosca Lechner (University of Waterloo), Shai Ben-David (University of Waterloo)

[P8] Predictive learning enables neural networks to learn complex working memory tasks

Thijs L van der Plas (University of Oxford), Tim P Vogels (Institute of Science and Technology Austria), Sanjay G Manohar (University of Oxford)

[P9] A Dataset Perspective on Offline Reinforcement Learning

Kajetan Schweighofer (Johannes Kepler University Linz), Andreas Radler (Johannes Kepler University Linz), Marius-Constantin Dinu (Johannes Kepler University Linz, Dynatrace Research Austria), Markus Hofmarcher (Johannes Kepler University Linz), Vihang Patil (Johannes Kepler University Linz), Angela Bitto-Nemling (Johannes Kepler University Linz, Institute of Advanced Research in Artificial Intelligence), Hamid Eghbal-zadeh (Johannes Kepler University Linz), Sepp Hochreiter (Johannes Kepler University Linz, Institute of Advanced Research in Artificial Intelligence)

[P10] Practical tradeoffs between memory, compute, and performance in learned optimizers

Luke Metz, C. Daniel Freeman, James Harrison|Niru Maheswaranathan|Jascha Sohl-Dickstein

[P11] Disentanglement and Generalization Under Correlation Shifts

Christina Funke (University of Tubingen), Paul Vicol (University of Toronto), Kuan-Chieh Wang (University of Toronto), Matthias Kummerer (University of Tubingen), Richard Zemel (University of Toronto), Matthias Bethge (University of Tubingen)

[P12] Continual Learning with Foundation Models: An Empirical Study of Latent Replay

Oleksiy Ostapenko (Mila, University of Montreal, ServiceNow), Timothee LESORT(Mila, University of Montreal), Pau Rodriguez(ServiceNow), Md Rifat Arefin(Mila, University of Montreal), Arthur Douillard(Sorbonne University, Heuritech), Irina Rish(Mila, University of Montreal,Canada CIFAR AI Chair), Laurent Charlin(Mila, University of Montreal,Canada CIFAR AI Chair)

[P13] How Does the Task Landscape Affect MAML Performance?

Liam Collins (University of Texas at Austin), Aryan Mokhtari (University of Texas at Austin), Sanjay Shakkottai (University of Texas at Austin)

[P14] Energy-Based Models for Continual Learning

Shuang Li (MIT), Yilun Du (MIT), Gido M.van de Ven (Baylor College of Medicine), Igor Mordatch (Google Brain)

Poster

[P15] Probing Transfer in Deep Reinforcement Learning without Task Engineering

Andrei Alex Rusu , Sebastian Flennerhag , Dushyant Rao, Razvan Pascanu, Raia Hadsell

[P16] Sim-To-Real Transfer of Visual Grounding for Human-Aided Ambiguity Resolution

Georgios Tziafas (University of Groningen), Hamidreza Kasaei (University of Groningen)

[P17] Consistency is the key to further mitigating catastrophic forgetting in continual learning

Prashant Bhat (Navinfo Europe B.V), Elahe Arani (Navinfo Europe B.V), Bahram Zonooz (Navinfo Europe B.V)

[P18] Learning Skills Diverse in Value-Relevant Features

Matthew J. A. Smith (University of Oxford), Jelena Luketina (University of Oxford), Kristian Hartikainen (University of Oxford), Maximilian Igl (University of Oxford), Shimon Whiteson (University of Oxford)

[P19] Improved Policy Optimization for Online Imitation Learning

Jonathan Wilder Lavington (University of British Columbia), Sharan Vaswani (Simon Fraser University), Mark Schmidt (University of British Columbia)

[P20] Trustworthiness Evaluation and Trust-Aware Design of CNN Architectures

Mingxi Cheng (University of Southern California), Tingyang Sun (University of Southern California), Shahin Nazarian (University of Southern California), Paul Bogdan (University of Southern California)

[P21] SHELS: Exclusive Feature Sets for Novelty Detection and Continual Learning Without Class Boundaries

Meghna Gummadi (University of Pennsylvania), David Kent (University of Pennsylvania), Jorge A. Mendez (University of Pennsylvania), Eric Eaton (University of Pennsylvania).

[P22] InBiaseD: Inductive Bias Distillation to Improve Generalization and Robustness through Shape-awareness

Shruthi Gowda(Advanced Research Lab, NavInfo Europe), Bahram Zonooz (Advanced Research Lab, NavInfo Europe) , Elahe Arani (Advanced Research Lab, NavInfo Europe)

[P23] CompoSuite: A Compositional Reinforcement Learning Benchmark

Jorge A. Mendez (University of Pennsylvania), Marcel Hussing (University of Pennsylvania), Meghna Gummadi (University of Pennsylvania), Eric Eaton (University of Pennsylvania)

[P24] Optimal transport meets noisy label robust loss and MixUp regularization for domain adaptation

Kilian Fatras (Mila, Mcgill University), Hiroki Naganuma (Mila, Université de Montréal), Ioannis Mitliagkas (Mila, Université de Montréal)

[P25] SYNERgy between SYNaptic consolidation and Experience Replay for general continual learning

Fahad Sarfraz (Advanced Research Lab, NavInfo Europe), Elahe Arani (Advanced Research Lab, NavInfo Europe), Bahram Zonooz (Advanced Research Lab, NavInfo Europe)

[P26] Hierarchical Kickstarting for Skill Transfer in Reinforcement Learning

Michael Matthews (University College London), Mikayel Samvelyan (University College London, Meta AI), Jack Parker-Holder (University of Oxford), Edward Grefenstette (University College London), Tim Rocktaschel (University College London)

[P27] Lifelong Robotic Reinforcement Learning by Retaining Experiences

Annie Xie (Stanford University), Chelsea Finn (Stanford University)

[P28] Lifelong DP: Consistently Bounded Differential Privacy in Lifelong Machine Learning

Phung Lai (New Jersey Institute of Technology), Han Hu (New Jersey Institute of Technology), NhatHai Phan (New Jersey Institute of Technology), Ruoming Jin (Kent State University), My T. Thai (University of Florida), An M. Chen (Qualcomm Incorporated)

[P29] Heat-RL: Online Model Selection for Streaming Time-Series Anomaly Detection

Yujing Wang (Peking University), Luoxing Xiong (Microsoft), Mingliang Zhang (Peking University), Hui Xue (Microsoft Research Asia), Qi Chen (Microsoft Research Asia), Yaming Yang (Microsoft), Yunhai Tong (Peking University), Congrui Huang (Microsoft), Bixiong Xu (Microsoft)

[P30] Online Continual Learning for Embedded Devices

Tyler L. Hayes (Rochester Institute of Technology), Christopher Kanan (University of Rochester; Rochester Institute of Technology)

[P31] Self-Activating Neural Ensembles for Continual Reinforcement Learning

Sam Powers (Carnegie Mellon University), Eliot Xing (Georgia Institute of Technology), Abhinav Gupta (Carnegie Mellon University)

[P32] Simulation-Acquired Latent Action Spaces for Dynamics Generalization

Nicholas Corrado (University of Wisconsin--Madison), Yuxiao Qu (University of Wisconsin--Madison), Josiah P. Hanna (University of Wisconsin--Madison)

[P33] A Multi-Head Model for Continual Learning via Out-of-Distribution Replay

Gyuhak Kim (University of Illinois at Chicago), Zixuan Ke (University of Illinois at Chicago), Bing Liu (University of Illinois at Chicago)

[P34] Benchmarking Learning Efficiency in Deep Reservoir Computing

Hugo Cisneros (Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University, WILLOW Inria & ENS PSL), Josef Sivic (Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University), Tomas Mikolov (Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University)

[P35] Reactive Exploration to Cope With Non-Stationarity in Lifelong Reinforcement Learning

Christian Alexander Steinparz (Visual Data Science Lab, Institute of Compute Graphics, Johannes Kepler University), Thomas Schmied (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University), Fabian Paischer (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University), Marius-Constantin Dinu (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University, and Dynatrace Research), Vihang Prakash Patil (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University), Angela Bitto-Nemling (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University, and Institute of Advanced Research in Artificial Intelligence IARAI), Hamid Eghbal-zadeh (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University), Sepp Hochreiter (ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University, and Institute of Advanced Research in Artificial Intelligence IARAI)

[P36] CLActive: Episodic Memories for Rapid Active Learning

Sri Aurobindo Munagala (International Institute of Information Technology Hyderabad), Sidhant Subramanian (International Institute of Information Technology Hyderabad), Shyamgopal Karthik (Eberhard-Karls-Universität Tübingen), Ameya Prabhu (University of Oxford), Anoop Namboodiri (International Institute of Information Technology Hyderabad)

[P37] Zipfian Environments for Reinforcement Learning

Stephanie Chan (DeepMind), Andrew Lampinen (DeepMind), Pierre Richemond (DeepMind), Felix Hill (DeepMind)

[P38] Task Agnostic Representation Consolidation: a Self-supervised based Continual Learning Approach

Prashant Bhat (Navinfo Europe B.V), Elahe Arani (Navinfo Europe B.V), Bahram Zonooz (Navinfo Europe B.V)

[P39] TAG: Task-based Accumulated Gradients for Lifelong learning

Pranshu Malviya (Polytechnique Montréal), Balaraman Ravindran (Indian Institute of Technology Madras), Sarath Chandar (Polytechnique Montréal)

[P40] Learning Object-Centered Autotelic Behaviors with Graph Neural Networks

Ahmed Akakzia (Sorbonne University), Olivier Sigaud (Sorbonne University)

[P41] Continual Learning of Dynamical Systems With Competitive Federated Reservoir Computing

Leonard Bereska, Efstratios Gavves

[P42] Inexperienced RL agents can’t get it right: lower bounds on regret at finite sample complexity

Vincent Létourneau (university of ottawa), Maia Fraser (university of ottawa))

[P43] Streaming Inference for Infinite Non-Stationary Clustering

Rylan Schaeffer (Stanford), Gabrielle Kaili-May Liu (MIT), Yilun Du (MIT), Scott Linderman (Stanford), Ila Rani Fiete (MIT)

[P44] Continual Learning and Private Unlearning

Bo Liu (The University of Texas at Austin), Qiang Liu (The University of Texas at Austin), Peter Stone (The University of Texas at Austin)

[P45] Improving Meta-Learning Generalization with Activation-Based Early-Stopping

D46 Simon Guiroy (Mila - Quebec AI Institute, Université de Montréal), Christopher Pal (Mila - Quebec AI Institute, Polytechnique Montréal), Gonçalo Mordido (Mila - Quebec AI Institute, Polytechnique Montréal), Sarath Chandar (Mila - Quebec AI Institute, Polytechnique Montréal)

[P46] Continual Unsupervised Learning for Optical Flow Estimation with Deep Networks

Simone Marullo (University of Florence), Matteo Tiezzi (University of Siena), Alessandro Betti (Inria, CNRS, Lab I3S, Maasai Team, Université Côte d'Azur), Lapo Faggi (University of Florence), Enrico Meloni (University of Florence), Stefano Melacci (University of Siena)

[P47] On Anytime Learning at Macroscale

Lucas Caccia (MILA, McGill University / FAIR), Jing Xu (FAIR), Myle Ott (FAIR), Marc'Aurelio Ranzato (FAIR), Ludovic Denoyer (FAIR)

[P48] EFL: Elastic Federated Learning on Non-IID Data

Zichen Ma (The Chinese University of Hong Kong, Shenzhen; JD AI Research), Yu Lu (The Chinese University of Hong Kong, Shenzhen; JD AI Research), Wenye Li (The Chinese University of Hong Kong, Shenzhen; Shenzhen Research Institute of Big Data), Shuguang Cui (The Chinese University of Hong Kong, Shenzhen; Shenzhen Research Institute of Big Data)

[P49] Model-Free Generative Replay for Lifelong Reinforcement Learning: Application to Starcraft-2

[P50] Adapting Pre-trained Language Models to Low-Resource Text Simplification: The Path Matters

Cristina Garbacea|Qiaozhu Mei

[P51] Continual Novelty Detection

Rahaf Aljundi|Daniel Olmeda Reino|Nikolay Chumerin|Richard E Turner

[P52] Differencing based Self-supervised pretraining for Scene Change Detection

Vijaya Raghavan T Ramkumar|Elahe Arani|Bahram Zonooz

[P53] Curbing Task Interference using Representation Similarity-Guided Multi-Task Feature Sharing

Naresh Kumar Gurulingan|Elahe Arani|Bahram Zonooz

[P54] MO2: Model-Based Offline Options

[P55] OPEN SET RECOGNITION VIA AUGMENTATION-BASED SIMILARITY LEARNING

Sepideh Esmaeilpour|Lei Shu|Bing Liu

[P56] How do Quadratic Regularizers Prevent Catastrophic Forgetting: The Role of Interpolation

Ekdeep Singh Lubana|Puja Trivedi|Danai Koutra|Robert Dick

[P57] A Theory for Knowledge Transfer in Continual Learning

Diana Benavides Prado|Patricia Riddle

[P58] Sparsity and Heterogeneous Dropout for Continual Learning in the Null Space of Neural Activations

Ali Abbasi|Parsa Nooralinejad|Vladimir Braverman|Hamed Pirsiavash|Soheil Kolouri

[P59] On the Limitations of Continual Learning for Malware Classification

Mohammad Saidur Rahman|Scott Coull|Matthew Wright

[P60] Forgetting and Imbalance in Robot Lifelong Learning with Off-policy Data

[P61] Increasing Model Generalizability for Unsupervised Visual Domain Adaptation

Mohammad Rostami

[P62] Overcoming challenges in leveraging GANs for few-shot data augmentation

[P63] A Rule-based Shield: Accumulating Safety Rules from Catastrophic Action Effects

Shahaf S. Shperberg|Bo Liu|Alessandro Allievi|Peter Stone

[P64] Continual Learning through Hamilton Equations