Learning to learn: An Introduction to Meta Learning
Slides PDF: https://drive.google.com/file/d/1DuHyotdwEAEhmuHQWwRosdiVBVGm8uYx/view
Abstract:
In recent years, high-capacity models, such as deep neural networks, have enabled very powerful machine learning techniques in domains where data is plentiful. However, domains where data is scarce have proven challenging for such methods because high-capacity function approximators critically rely on large datasets for generalization. This can pose a major challenge for domains ranging from supervised medical image processing to reinforcement learning where real-world data collection (e.g., for robots) poses a major logistical challenge. Meta-learning or few-shot learning offers a potential solution to this problem: by learning to learn across data from many previous tasks, few-shot meta-learning algorithms can discover the structure among tasks to enable fast learning of new tasks.
The objective of this tutorial is to provide a unified perspective of meta-learning: teaching the audience about modern approaches, describing the conceptual and theoretical principles surrounding these techniques, presenting where these methods have been applied previously, and discussing the fundamental open problems and challenges within the area. We hope that this tutorial is useful for both machine learning researchers whose expertise lies in other areas, while also providing a new perspective to meta-learning researchers. All in all, we aim to provide audience members with the ability to apply meta-learning to their own applications, and develop new meta-learning algorithms and theoretical analyses driven by the current challenges and limitations of existing work.
We will provide a unified perspective of how a variety of meta-learning algorithms enable learning from small datasets, an overview of applications where meta-learning can and cannot be easily applied, and a discussion of the outstanding challenges and frontiers of this sub-field.
Видео Learning to learn: An Introduction to Meta Learning канала Machine Learning TV
Abstract:
In recent years, high-capacity models, such as deep neural networks, have enabled very powerful machine learning techniques in domains where data is plentiful. However, domains where data is scarce have proven challenging for such methods because high-capacity function approximators critically rely on large datasets for generalization. This can pose a major challenge for domains ranging from supervised medical image processing to reinforcement learning where real-world data collection (e.g., for robots) poses a major logistical challenge. Meta-learning or few-shot learning offers a potential solution to this problem: by learning to learn across data from many previous tasks, few-shot meta-learning algorithms can discover the structure among tasks to enable fast learning of new tasks.
The objective of this tutorial is to provide a unified perspective of meta-learning: teaching the audience about modern approaches, describing the conceptual and theoretical principles surrounding these techniques, presenting where these methods have been applied previously, and discussing the fundamental open problems and challenges within the area. We hope that this tutorial is useful for both machine learning researchers whose expertise lies in other areas, while also providing a new perspective to meta-learning researchers. All in all, we aim to provide audience members with the ability to apply meta-learning to their own applications, and develop new meta-learning algorithms and theoretical analyses driven by the current challenges and limitations of existing work.
We will provide a unified perspective of how a variety of meta-learning algorithms enable learning from small datasets, an overview of applications where meta-learning can and cannot be easily applied, and a discussion of the outstanding challenges and frontiers of this sub-field.
Видео Learning to learn: An Introduction to Meta Learning канала Machine Learning TV
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![iMAML: Meta-Learning with Implicit Gradients (Paper Explained)](https://i.ytimg.com/vi/u5BkO8XMS2I/default.jpg)
![CS 182: Lecture 21: Part 1: Meta-Learning](https://i.ytimg.com/vi/h7qyQeXKxZE/default.jpg)
![Unleash Your Super Brain To Learn Faster | Jim Kwik](https://i.ytimg.com/vi/uT_GcOGEFsk/default.jpg)
![How to learn any language in six months | Chris Lonsdale | TEDxLingnanUniversity](https://i.ytimg.com/vi/d0yGdNEWdn0/default.jpg)
![](https://i.ytimg.com/vi/fdC8wz9tndQ/default.jpg)
![Zero-Shot Learning - Dr. Timothy Hospedales](https://i.ytimg.com/vi/jBnCcr-3bXc/default.jpg)
![Few-Shot Learning (1/3): Basic Concepts](https://i.ytimg.com/vi/hE7eGew4eeg/default.jpg)
![Ilya Sutskever: OpenAI Meta-Learning and Self-Play | MIT Artificial General Intelligence (AGI)](https://i.ytimg.com/vi/9EN_HoEk3KY/default.jpg)
![Hyperparameter Optimization: This Tutorial Is All You Need](https://i.ytimg.com/vi/5nYqK-HaoKY/default.jpg)
![Meta Learning](https://i.ytimg.com/vi/2z0ofe2lpz4/default.jpg)
![Graph Representation Learning (Stanford university)](https://i.ytimg.com/vi/YrhBZUtgG4E/default.jpg)
![Elon Musk’s 2 Rules For Learning Anything Faster](https://i.ytimg.com/vi/xLkC-ODKQSc/default.jpg)
![WDS 2019 Main Stage Keynote - Scott Young on the Best Way to Learn Anything](https://i.ytimg.com/vi/J2JGiJNUsD4/default.jpg)
![Yonglong Tian - Contrastive Learning: A General Self-supervised Learning Approach](https://i.ytimg.com/vi/7YBwnc9D2d4/default.jpg)
![Hierarchical reinforcement learning - Doina Precup](https://i.ytimg.com/vi/e8b0yC6COJ8/default.jpg)
![Brain Expert Jim Kwik on Why Meta Learning Is the Most Important Skill | Inc.](https://i.ytimg.com/vi/uZ2LFZxwvuw/default.jpg)
![Intro to Deep Learning (ML Tech Talks)](https://i.ytimg.com/vi/AhE8RhPGH1A/default.jpg)
![A Neural Network Model That Can Reason - Prof. Christopher Manning](https://i.ytimg.com/vi/24AX4qJ7Tts/default.jpg)
![Metacognition: The Skill That Promotes Advanced Learning](https://i.ytimg.com/vi/elZFL4FLVLE/default.jpg)
![BayLearn2020 Keynote: Meta-learning for Robustness to the Changing World - Prof. Chelsea Finn](https://i.ytimg.com/vi/RNxCRf9h9xM/default.jpg)