# Dynamic Influence-Weighted Distillation for Single-IMU Activity Recognition

arXiv: https://arxiv.org/abs/2608.24904 · HTML: https://mlsanity.com/paper/2608.24904/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 32 references that are in the map.
- First posted on arXiv 17 Jul 2026.

## Builds on (most cited first, 19 of 19)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [FitNets: Hints for Thin Deep Nets](https://mlsanity.com/paper/1412.6550/) (2014)
- [Knowledge Distillation: A Survey](https://mlsanity.com/paper/2006.05525/) (2020)
- [An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling](https://mlsanity.com/paper/1803.01271/) (2018)
- [Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer](https://mlsanity.com/paper/1612.03928/) (2016)
- [Learning to Reweight Examples for Robust Deep Learning](https://mlsanity.com/paper/1803.09050/) (2018)
- [Unifying distillation and privileged information](https://mlsanity.com/paper/1511.03643/) (2015)
- [Deep, Convolutional, and Recurrent Models for Human Activity Recognition using Wearables](https://mlsanity.com/paper/1604.08880/) (2016)
- [Learning to Teach with Dynamic Loss Functions](https://mlsanity.com/paper/1810.12081/) (2018)
- [WEAR: An Outdoor Sports Dataset for Wearable and Egocentric Activity Recognition](https://mlsanity.com/paper/2304.05088/) (2023)
- [Adaptive Multi-Teacher Knowledge Distillation with Meta-Learning](https://mlsanity.com/paper/2306.06634/) (2023)
- [Learning from the Best: Contrastive Representations Learning Across Sensor Locations for Wearable Activity Recognition](https://mlsanity.com/paper/2210.01459/) (2022)
- [MESEN: Exploit Multimodal Data to Design Unimodal Human Activity Recognition with Few Labels](https://mlsanity.com/paper/2404.01958/) (2024)
- [Virtual Fusion with Contrastive Learning for Single Sensor-based Activity Recognition](https://mlsanity.com/paper/2312.02185/) (2023)
- [Learning and Refining of Privileged Information-based RNNs for Action Recognition from Depth Sequences](https://mlsanity.com/paper/1703.09625/) (2017)
- [Tailoring Instructions to Student's Learning Levels Boosts Knowledge Distillation](https://mlsanity.com/paper/2305.09651/) (2023)
- [TSAK: Two-Stage Semantic-Aware Knowledge Distillation for Efficient Wearable Modality and Model Optimization in Manufacturing Lines](https://mlsanity.com/paper/2408.14146/) (2024)

## Abstract (arXiv)

Inertial sensors at multiple body locations can improve activity recognition, but requiring every sensor at inference increases the deployment burden. We study whether four synchronized IMUs available during training can improve a student that uses only the right-arm IMU during fitting and inference. A frozen four-IMU teacher provides logit and feature targets. Fixed-weight knowledge distillation applies each target with the same strength to every fitting sample, although the student may not benefit equally from them. We introduce dynamic influence weighting (DIW), which tests a one-step candidate update on separate fold-internal training participants. DIW then assigns separate sample-wise gates to the logit and feature losses. On WEAR, we evaluate 19 labels and 68,298 complete windows from 22 participants using subject-disjoint five-fold cross-validation. Pooled out-of-fold macro-F1 is 0.561820 for Supervised and 0.571623 for Fixed-weight KD. DIW reaches 0.638451, gains of 7.66 and 6.68 percentage points, respectively. It exceeds Supervised for 18 of 19 labels and 21 of 22 held-out participants. All three routes retain the same 80,915-parameter right-arm student at inference. Under this protocol, DIW converts training-only multi-position information into a stronger single-IMU model without changing deployed sensing or the student forward graph.