# D-CAT: Decoupled Cross-Attention Transfer between Sensor Modalities for Unimodal Inference

arXiv: https://arxiv.org/abs/2509.09747 · HTML: https://mlsanity.com/paper/2509.09747/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 29 references that are in the map.
- First posted on arXiv 11 Sep 2025.

## Builds on (most cited first, 8 of 8)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [VGGSound: A Large-scale Audio-Visual Dataset](https://mlsanity.com/paper/2004.14368/) (2020)
- [PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition](https://mlsanity.com/paper/1912.10211/) (2019)
- [Towards Continual Egocentric Activity Recognition: A Multi-modal Egocentric Activity Dataset for Continual Learning](https://mlsanity.com/paper/2301.10931/) (2023)
- [Audio-Based Activities of Daily Living (ADL) Recognition with Large-Scale Acoustic Embeddings from Online Videos](https://mlsanity.com/paper/1810.08691/) (2018)
- [Learning Modality Knowledge Alignment for Cross-Modality Transfer](https://mlsanity.com/paper/2406.18864/) (2024)
- [Human Activity Recognition in RGB-D Videos by Dynamic Images](https://mlsanity.com/paper/1807.02947/) (2018)
- [UAC: Uncertainty-Aware Calibration of Neural Networks for Gesture Detection](https://mlsanity.com/paper/2504.02895/) (2025)

## Abstract (arXiv)

Cross-modal transfer learning is used to improve multi-modal classification models (e.g., for human activity recognition in human-robot collaboration). However, existing methods require paired sensor data at both training and inference, limiting deployment in resource-constrained environments where full sensor suites are not economically and technically usable. To address this, we propose Decoupled Cross-Attention Transfer (D-CAT), a framework that aligns modality-specific representations without requiring joint sensor modality during inference. Our approach combines a self-attention module for feature extraction with a novel cross-attention alignment loss, which enforces the alignment of sensors' feature spaces without requiring the coupling of the classification pipelines of both modalities. We evaluate D-CAT on three multi-modal human activity datasets (IMU, video, and audio) under both in-distribution and out-of-distribution scenarios, comparing against uni-modal models. Results show that in in-distribution scenarios, transferring from high-performing modalities (e.g., video to IMU) yields up to 10% F1-score gains over uni-modal training. In out-of-distribution scenarios, even weaker source modalities (e.g., IMU to video) improve target performance, as long as the target model isn't overfitted on the training data. By enabling single-sensor inference with cross-modal knowledge, D-CAT reduces hardware redundancy for perception systems while maintaining accuracy, which is critical for cost-sensitive or adaptive deployments (e.g., assistive robots in homes with variable sensor availability). Code is available at https://github.com/Schindler-EPFL-Lab/D-CAT.