MemoNav: Selecting Informative Memories for Visual Navigation
Not on the map. Why?
Abstract (arXiv)
Image-goal navigation is a challenging task, as it requires the agent to navigate to a target indicated by an image in a previously unseen scene. Current methods introduce diverse memory mechanisms which save navigation history to solve this task. However, these methods use all observations in the memory for generating navigation actions without considering which fraction of this memory is informative. To address this limitation, we present the MemoNav, a novel memory mechanism for image-goal navigation, which retains the agent's informative short-term memory and long-term memory to improve the navigation performance on a multi-goal task. The node features on the agent's topological map are stored in the short-term memory, as these features are dynamically updated. To aid the short-term memory, we also generate long-term memory by continuously aggregating the short-term memory via a graph attention module. The MemoNav retains the informative fraction of the short-term memory via a forgetting module based on a Transformer decoder and then incorporates this retained short-term memory and the long-term memory into working memory. Lastly, the agent uses the working memory for action generation. We evaluate our model on a new multi-goal navigation dataset. The experimental results show that the MemoNav outperforms the SoTA methods by a large margin with a smaller fraction of navigation history. The results also empirically show that our model is less likely to be trapped in a deadlock, which further validates that the MemoNav improves the agent's navigation efficiency by reducing redundant steps.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 23 of its 43 references that are in the map. First posted on arXiv 20 Aug 2022.
Builds on
23 of 43 references.
References · 23
- 2017Attention Is All You Needcited by 60,361
- 2017Proximal Policy Optimization Algorithmscited by 15,215
- 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
- 2020Longformer: The Long-Document Transformercited by 3,413
- 2019Habitat: A Platform for Embodied AI Researchcited by 1,191
- 2021How Attentive are Graph Attention Networks?cited by 835
- 2018On Evaluation of Embodied Navigation Agentscited by 547
- 2018Gibson Env: Real-World Perception for Embodied Agentscited by 481
- 2020Learning to Explore using Active Neural SLAMcited by 316
- 2018Semi-parametric Topological Memory for Navigationcited by 209
- 2020Neural Topological SLAM for Visual Navigationcited by 195
- 2021Episodic Transformer for Vision-and-Language Navigationcited by 156
11 more
- 2019Learning Exploration Policies for Navigationcited by 116
- 2020Topological Planning with Transformers for Vision-and-Language Navigationcited by 116
- 2019Scene Memory Transformer for Embodied Agents in Long-Horizon Taskscited by 107
- 2020MultiON: Benchmarking Semantic Map Memory using Multi-Object Navigationcited by 95
- 2021Memory-Augmented Reinforcement Learning for Image-Goal Navigationcited by 66
- 2021Towards mental time travel: a hierarchical memory for reinforcement learning agentscited by 40
- 2021Not All Memories are Created Equal: Learning to Forget by Expiringcited by 32
- 2020Learning to plan with uncertain topological mapscited by 29
- 2019Working Memory Graphscited by 23
- 2020Rapid Task-Solving in Novel Environmentscited by 23
- 2018Visual Memory for Robust Path Followingcited by 16
Built on by
Citing papers per year: 2023: 1.
1 paper.