Recent years have seen rapid progress in meta-learning methods, which transfer knowledge across tasks and domains to efficiently learn new tasks, optimize the learning process itself, and even generate new learning methods from scratch. Meta-learning can be seen as the logical conclusion of the arc that machine learning has undergone in the last decade, from learning classifiers, to learning representations, and finally to learning algorithms that themselves acquire representations, classifiers, and policies for acting in environments. In practice, meta-learning has been shown to yield new state-of-the-art automated machine learning methods, novel deep learning architectures, and substantially improved one-shot learning systems. Moreover, to improve one’s own learning capabilities through experience can also be viewed as a hallmark of intelligent beings, and neuroscience shows a strong connection between human and reward learning and the growing sub-field of meta-reinforcement learning.
| Introduction and opening remarks (Live speech) | |
| Ying Wei (Invited talk) | |
| Ying Wei Q&A (Q&A) | |
| Meta-Learning Reliable Priors in the Function Space (Contributed talk & Poster) | |
| Poster session 1 (Poster session) | |
| Carlo Ciliberto (Invited talk) | |
| Carlo Ciliberto Q&A (Q&A) | |
| Mihaela Van Der Schaar (Invited talk) | |
| Mihaela Van Der Schaar Q&A (Q&A) | |
| Break | |
| Panel Discussion | |
| Bootstrapped Meta-Learning (Contributed talk & Poster) | |
| Nan Rosemary Ke (Invited talk) | |
| Nan Rosemary Ke Q&A (Q&A) | |
| Poster session 2 (Poster session) | |
| Luke Metz (Invited talk) | |
| Luke Metz Q&A (Q&A) | |
| Eleni Triantafillou (Invited talk) | |
| Eleni Triantafillou Q&A (Q&A) | |
| Offline Meta-Reinforcement Learning with Online Self-Supervision (Contributed talk & Poster) | |
| Poster session 3 (Poster session) | |
| On the Practical Consistency of Meta-Reinforcement Learning Algorithms (Poster) | |
| Understanding Catastrophic Forgetting and Remembering in Continual Learning with Optimal Relevance Mapping (Poster) | |
| Few Shot Image Generation via Implicit Autoencoding of Support Sets (Poster) | |
| Transfer Learning for Bayesian HPO with End-to-End Landmark Meta-Features (Poster) | |
| How to distribute data across tasks for meta-learning? (Poster) | |
| Introducing Symmetries to Black Box Meta Reinforcement Learning (Poster) | |
| Task Attended Meta-Learning for Few-Shot Learning (Poster) | |
| One Step at a Time: Pros and Cons of Multi-Step Meta-Gradient Reinforcement Learning (Poster) | |
| Meta-learning from sparse recovery (Poster) | |
| A Preliminary Study on the Feature Representations of Transfer Learning and Gradient-Based Meta-Learning Techniques (Poster) | |
| Neural Processes with Stochastic Attention: Paying more attention to the context dataset (Poster) | |
| Curriculum Meta-Learning for Few-shot Classification (Poster) | |
| Successor Feature Neural Episodic Control (Poster) | |
| FedMix: A Simple and Communication-Efficient Alternative to Local Methods in Federated Learning (Poster) | |
| Unsupervised Meta-Learning via Latent Space Energy-based Model of Symbol Vector Coupling (Poster) | |
| A Nested Bi-level Optimization Framework for Robust Few Shot Learning (Poster) | |
| A Meta-Gradient Approach to Learning Cooperative Multi-Agent Communication Topology (Poster) | |
| Efficient Automated Online Experimentation with Multi-Fidelity (Poster) | |
| Sign-MAML: Efficient Model-Agnostic Meta-Learning by SignSGD (Poster) | |
| Meta-learning inductive biases of learning systems with Gaussian processes (Poster) | |
| DARTS without a Validation Set: Optimizing the Marginal Likelihood (Poster) | |
| Studying BatchNorm Learning Rate Decay on Meta-Learning Inner-Loop Adaptation (Poster) | |
| Effect of diversity in Meta-Learning (Poster) | |
| Skill-based Meta-Reinforcement Learning (Poster) | |
| On the Role of Pre-training for Meta Few-Shot Learning (Poster) | |
| Contrastive Embedding of Structured Space for Bayesian Optimization (Poster) | |
| Hierarchical Few-Shot Generative Models (Poster) | |
| Variational Task Encoders for Model-Agnostic Meta-Learning (Poster) | |
| Open-Ended Learning Strategies for Learning Complex Locomotion Skills (Poster) | |
| Transformers Can Do Bayesian-Inference By Meta-Learning on Prior-Data (Poster) | |