Semantic Audio-Visual Navigation
Open accessAbstract
Recent work on audio-visual navigation assumes a constantly-sounding target and restricts the role of audio to signaling the target’s position. We introduce semantic audio-visual navigation, where objects in the environment make sounds consistent with their semantic meaning (e.g., toilet flushing, door creaking) and acoustic events are sporadic or short in duration. We propose a transformer-based model to tackle this new semantic AudioGoal task, incorporating an inferred goal descriptor that captures both spatial and semantic properties of the target. Our model’s persistent multimodal memory enables it to reach the goal even long after the acoustic event stops. In support of the new task, we also expand the SoundSpaces audio simulations to provide semantically grounded sounds for an array of objects in Matterport3D. Our method strongly outperforms existing audio-visual navigation methods by learning to associate semantic, acoustic, and visual cues.1
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The reference SAVi agent achieved an average success rate of 25%.
“Despite these additions, the agent was only able to achieve an average success rate of 25%.”
From Design and evaluation of a global workspace agent embodied in a realistic multimodal environment (2024) · cites this paper as Chen et al., 2021
The SAVi task features two semantic concepts that can be queried from the features learned by the agent.
“The SAVi task (Chen et al., 2021) features two semantic concepts that can be queried from the features learned by the agent.”
From Design and evaluation of a global workspace agent embodied in a realistic multimodal environment (2024) · cites this paper as Chen et al., 2021
The target object category is provided to the agent via the acoustic modality observation and is available only for a variable duration from the start of the episode.
“First, the target object category, which varies across episodes, is provided to the agent via the acoustic modality observation, and is available only for a variable duration from the start of the episode.”
From Design and evaluation of a global workspace agent embodied in a realistic multimodal environment (2024) · cites this paper as Chen et al., 2021
There are 21 classes of target object categories in total.
“There are 21 classes of target object categories in total, covering various daily life objects such as chairs, tables, cabinets, sinks, and more (Chen et al., 2021).”
From Design and evaluation of a global workspace agent embodied in a realistic multimodal environment (2024) · cites this paper as Chen et al., 2021
The original SAVi agent trained with RL only reached 25% success on average.
“while the original SAVi agent trained with RL only reached 25% success on average (Chen et al., 2021).”
From Design and evaluation of a global workspace agent embodied in a realistic multimodal environment (2024) · cites this paper as Chen et al., 2021
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · see citing works
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · see citing works
Proximal Policy Optimization Algorithms
Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017) · arXiv · doi:10.48550/arxiv.1707.06347 · Cited by 10,625 · see citing works
Matterport3D: Learning from RGB-D Data in Indoor Environments
Chang, A., Dai, A., Funkhouser, T., Halber, M., Niebner, M., Savva, M., Song, S., Zeng, A., & Zhang, Y. (2017) · 2017 International Conference on 3D Vision (3DV) · doi:10.1109/3dv.2017.00081 · Cited by 1,858 · see citing works
SoundSpaces: Audio-Visual Navigation in 3D Environments
Chen, C., Jain, U., Schissler, C., Gari, S. V. A., Al-Halah, Z., Ithapu, V. K., Robinson, P., & Grauman, K. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58539-6_2 · Cited by 202 · see citing works
Self-citationLearning to Set Waypoints for Audio-Visual Navigation
Chen, C., Majumder, S., Al-Halah, Z., Gao, R., Ramakrishnan, S. K., & Grauman, K. (2020) · arXiv · doi:10.48550/arxiv.2008.09622 · Cited by 53 · see citing works
Self-citation
Flagged first, then by citations
Cited by
Sorted by citations
Multimodal learning with transformers: A survey
Xu, P., Zhu, X., & Clifton, D. A. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3275156 · Cited by 1,072 · see citing works
PONI: Potential Functions for ObjectGoal Navigation with Interaction-free Learning
Ramakrishnan, S. K., Chaplot, D. S., Al-Halah, Z., Malik, J., & Grauman, K. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01832 · Cited by 154 · see citing works
Self-citationZero Experience Required: Plug & Play Modular Transfer Learning for Semantic Visual Navigation
Al-Halah, Z., Ramakrishnan, S. K., & Grauman, K. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.01652 · Cited by 58 · see citing works
Self-citationTransformer Memory for Interactive Visual Navigation in Cluttered Environments
Li, W., Hong, R., Shen, J., Yuan, L., & Lu, Y. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3241803 · Cited by 25 · see citing works
Design and evaluation of a global workspace agent embodied in a realistic multimodal environment
Dossa, R. F. J., Arulkumaran, K., Juliani, A., Sasai, S., & Kanai, R. (2024) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2024.1352685 · Cited by 5 · see citing works
“There are 21 classes of target object categories in total, covering various daily life objects such as chairs, tables, cabinets, sinks, and more (Chen et al., 2021).”
BackgroundMethodInfluentialCAVEN: An Embodied Conversational Agent for Efficient Audio-Visual Navigation in Noisy Environments
Liu, X., Paul, S., Chatterjee, M., & Cherian, A. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i4.28167 · Cited by 6 · see citing works
Audio-Guided Visual Perception for Audio-Visual Navigation
Wang, Y., Yu, Y., Sun, F., Wang, L., & Zheng, W. (2025) · 2025 International Conference on Virtual Reality and Visualization (ICVRV) · doi:10.1109/icvrv67992.2025.00184
“SAVi [4] enables continuous navigation after sound cessation through a pre-trained sub-module that predicts target sound semantics.”
- Authors
- Chen, C., Al-Halah, Z., & Grauman, K.
- Affiliations
- Meta (United States) (US)
- The University of Texas at Austin (US)
- Venue
- 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2021-06
- Pages
- 15511-15520
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W3176232375
- Cited by
- 96
- References
- 6
- Citations by year
- 2026: 7 · 2025: 18 · 2024: 14 · 2023: 31 · 2022: 20 · 2021: 5 · 2020: 1
- FWCI
- 8.08
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Science Foundation
- University of Texas at Austin
- Defense Advanced Research Projects Agency
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Engineering › Automotive Engineering › Spatial Cognition and Navigation
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Keywords
- multimodal memory, Matterport3D
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
