Research map / Multi-agent systems and games / Human-AI and goal-conditioned
WhisperNet: A Scalable Solution for Bandwidth-Efficient Collaboration
Gong Chen, Chaokun Zhang, Xinyan Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Channel Merit Distributor (ablation) | WhisperNet baseline · OPV2V | AP@0.7: 0.8333 -> 0.8532 | Improved |
| add Spatial Focus Engine (ablation) | WhisperNet (CMD) · OPV2V | AP@0.7: 0.8532 -> 0.8683 | Improved |
| add Collaborative Feature Routing (ablation) | WhisperNet (CMD+SFE) · OPV2V | AP@0.7: 0.8683 -> 0.8764 | Improved |
| add Collaborative Feature Routing (ablation) | WhisperNet baseline · OPV2V | AP@0.7: 0.8333 -> 0.8672 | Improved |
| add Collaborative Feature Routing (ablation) | WhisperNet (CMD+SFE) · DAIR-V2X | AP@0.7: 0.6271 -> 0.6480 | Improved |
| increase experts from 1 to 4 (ablation) | CFR module · OPV2V | AP@0.7: 0.8512 -> 0.8764 | Improved |
| increase experts from 4 to 8 (ablation) | CFR module · OPV2V | AP@0.7: 0.8764 -> 0.8435 | Hurt |
| use patch size 1 instead of 4 at 5% comm rate (ablation) | WhisperNet · OPV2V | AP@0.7: 0.7732 -> 0.8511 | Improved |
| use patch size 1 instead of 4 at 50% comm rate (ablation) | WhisperNet · OPV2V | AP@0.7: 0.8752 -> 0.8764 | No clear effect |
| use Laplacian instead of Jacobian channel score (ablation) | WhisperNet (multi-scale, weighted) · OPV2V | AP@0.7: 0.8481 -> 0.8729 | Improved |
| use Laplacian instead of Mean channel score (ablation) | WhisperNet (multi-scale, weighted) · OPV2V | AP@0.7: 0.8474 -> 0.8729 | Improved |
| use group-weighted instead of max aggregation (ablation) | WhisperNet (Laplacian, multi-scale) · OPV2V | AP@0.7: 0.8709 -> 0.8729 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- CoSDH — CoSDH: Communication-Efficient Collaborative Perception via Supply-Demand Awareness and Intermediate-Late Hybridization (CVPR 2025) · better on 2 of 2 shared results
- ERMVP — ERMVP: Communication-Efficient and Collaboration-Robust Multi-Vehicle Perception in Challenging Environments (CVPR 2024) · better on 2 of 2 shared results
- How2Comm — How2comm: Communication-Efficient and Collaboration-Pragmatic Multi-Agent Perception (NeurIPS 2023) · better on 2 of 2 shared results
- Where2comm — Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps (NeurIPS 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- InfoCom: Kilobyte-Scale Communication-Efficient Collaborative Perception with Information Bottleneck (AAAI 2026 · similarity 0.89)
- IFTR: An Instance-Level Fusion Transformer for Visual Collaborative Perception (ECCV 2024 · similarity 0.89)
- How2comm: Communication-Efficient and Collaboration-Pragmatic Multi-Agent Perception (NeurIPS 2023 · similarity 0.89)
- Privacy-Preserving V2X Collaborative Perception Integrating Unknown Collaborators (AAAI 2025 · similarity 0.88)
- STAMP: Scalable Task- And Model-agnostic Collaborative Perception (ICLR 2025 · similarity 0.88)
- CoSDH: Communication-Efficient Collaborative Perception via Supply-Demand Awareness and Intermediate-Late Hybridization (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
