Research map / Vision-language and multimodal / Cross-modal learning
Cross-Tactile Sensor Representation Learning
Yan Zhang, Zheng WANG, Pengpeng Zeng, Xing Xu, Jingkuan Song, Heng Tao Shen
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 Cross-Sensor Modulator (CSM) (ablation) | CTSRL Stage 1 · SITR Bench DIGIT-Mini | accuracy: 78.16 -> 80.62 | Improved |
| add cross-sensor alignment loss (ablation) | CTSRL Stage 1 · SITR Bench DIGIT-Mini | accuracy: 74.34 -> 80.62 | Improved |
| add intra-sensor loss (ablation) | CTSRL Stage 1 · SITR Bench DIGIT-Mini | accuracy: 73.70 -> 80.62 | Improved |
| add loss normalization (ablation) | CTSRL Stage 1 · SITR Bench DIGIT-Mini | accuracy: 26.02 -> 80.62 | Improved |
| add cross-sensor alignment on top of normalized loss versus plain Barlow Twins (ablation) | CTSRL Stage 1 · SITR Bench DIGIT-Mini | accuracy: 23.84 -> 80.62 | Improved |
| add intra-sensor loss (ablation) | CTSRL Stage 1 · SITR Bench no-transfer | accuracy: 99.62 -> 99.58 | No clear effect |
| add CSM (ablation) | CTSRL Stage 2 · OF Real | accuracy: 48.70 -> 50.14 | Improved |
| add Stage 1 cross-sensor pretraining (ablation) | CTSRL Stage 2 · OF Real | accuracy: 46.88 -> 50.14 | Improved |
| freeze the CLIP vision encoder (ablation) | CTSRL Stage 2 · Cloth material | accuracy: 41.98 -> 48.33 | Improved |
| add language modality alignment (ablation) | CTSRL Stage 2 · OF Real | accuracy: 34.09 -> 50.14 | Improved |
| add vision modality alignment (ablation) | CTSRL Stage 2 · OF Real | accuracy: 46.02 -> 50.14 | Improved |
| add Stage 2 cross-modal semantic alignment after Stage 1 | CTSRL · Touch and Go material | accuracy: 56.42 -> 87.70 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AnyTouch+ — AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors (ICLR 2025) · better on 5 of 6 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
- AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors (ICLR 2025 · similarity 0.91)
- Sensor-Invariant Tactile Representation (ICLR 2025 · similarity 0.90)
- Touch in the Wild: Learning Fine-Grained Manipulation with a Portable Visuo-Tactile Gripper (NeurIPS 2025 · similarity 0.89)
- Binding Touch to Everything: Learning Unified Multimodal Tactile Representations (CVPR 2024 · similarity 0.88)
- Universal Visuo-Tactile Video Understanding for Embodied Interaction (NeurIPS 2025 · similarity 0.87)
- Touch-R1: Reinforcing Touch Reasoning in MLLMs (NeurIPS 2026 · 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.
