Multimodal Aggregation Approach for Memory Vision-Voice Indoor Navigation with Meta-Learning

Yan, Liqi; Liu, Dongfang; Song, Yaoxian; Yu, Changbin

doi:10.48550/arxiv.2009.00402

Search citation statements

Order By: Relevance

Paper Sections

Select...

B Meta Reinforcement Learning1

Citation Types

Supporting

Mentioning

Contrasting

Year Published

2021

Publication Types

Select...

Other1

Relationship

Self Cite0

Independent1

Authors

Journals

Cited by 1 publication

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In real cases, the tasks often share lots of common features, for instance going upstairs and downstairs both involve bending the knee and lifting the legs. Meta learning methods could exploit these common structure by collecting previous tasks and implement parameter generalization on meta-training set [26] [27]. If we hope the agents could quickly adapt to a new task, the meta reinforcement learning algorithm could help alleviate this problem by aggregating prior trajectories [3] [11] [20].…”

Section: B Meta Reinforcement Learningmentioning

confidence: 99%

Adaptive Adversarial Training for Meta Reinforcement Learning

Chen¹,

Chen²,

Wang³

2021

Preprint

View full text Add to dashboard Cite

Meta Reinforcement Learning (MRL) enables an agent to learn from a limited number of past trajectories and extrapolate to a new task. In this paper, we attempt to improve the robustness of MRL. We build upon model-agnostic metalearning (MAML) and propose a novel method to generate adversarial samples for MRL by using Generative Adversarial Network (GAN). That allows us to enhance the robustness of MRL to adversal attacks by leveraging these attacks during meta training process.

show abstract