斯坦福大学的CS 330课程Deep Multi-Task and Meta Learning（深度多任务学习与元学习）正在进行中，官方网站中部分Notes已放出。
虽然深度学习在有监督学习和强化学习问题（如图像分类、语音识别、游戏等）中获得了卓越的成果，这些模型在很大程度上都是针对单向任务来进行训练的。斯坦福大学的CS 330课程Deep Multi-Task and Meta Learning将会覆盖需要解决多项任务的场景，学习如何有效和高效地利用多任务模型。
成为VIP会员查看完整内容
Course introduction, problem definitions, applications
Supervised multi-task learning, black-box meta-learning
Meta-learning problem statement, black-box meta-learning
Optimization-based meta-learning
Few-shot learning via metric learning
Advanced meta-learning topics
Variational inference tutorial
Bayesian meta-learning
Reinforcement learning tutorial
RL review
Multi-task RL and goal-conditioned RL
Model-based RL for multi-task learning
Meta-RL 1: Adaptable models and policies
Meta-RL 2: Learning to explore
Offline RL and offline MTRL
Hierarchical RL and skill discovery
What do language models learn from language modeling?
Learned optimizers: why they're the future, why they're hard, and what they can do now
Lifelong learning: problem statements, forward & backward transfer
Frontiers and open problems
Learning to edit pre-trained models
BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learning
Language Conditioned Imitation Learning Over Unstructured Data
Learning Language-Conditioned Robot Behavior from Offline Data and Crowd-Sourced Annotation
Autonomous Reinforcement Learng
Learning Generalizable Robotic Reward Functions from “In-The-Wild” Human Videos
Reinforcement Learning as Sequence Modeling
Lifelong Robotic Reinforcement Learning by Retaining Experiences
相关内容
Arxiv
20+阅读 · 2020年3月10日
Arxiv
33+阅读 · 2019年10月24日
Arxiv
4+阅读 · 2018年4月22日
相关VIP内容
相关资讯
相关论文
Arxiv
20+阅读 · 2020年3月10日
Arxiv
33+阅读 · 2019年10月24日
Arxiv
4+阅读 · 2018年4月22日
微信扫码咨询专知VIP会员