# I-SPLIT: Deep Network Interpretability for Split Computing

arXiv: https://arxiv.org/abs/2209.11607 · HTML: https://mlsanity.com/paper/2209.11607/

## Key facts

- Cited by 7 papers in the map (data through 7 Oct 2026).
- Builds on 18 of its 36 references that are in the map.
- First posted on arXiv 23 Sep 2022.

## Builds on (most cited first, 18 of 18)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Sanity Checks for Saliency Maps](https://mlsanity.com/paper/1810.03292/) (2018)
- [Striving for Simplicity: The All Convolutional Net](https://mlsanity.com/paper/1412.6806/) (2014)
- [Not Just a Black Box: Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1605.01713/) (2016)
- [Interpretable Deep Learning: Interpretation, Interpretability, Trustworthiness, and Beyond](https://mlsanity.com/paper/2103.10689/) (2021)
- [Split Computing and Early Exiting for Deep Learning Applications: Survey and Research Challenges](https://mlsanity.com/paper/2103.04505/) (2021)
- [BottleNet++: An End-to-End Approach for Feature Compression in Device-Edge Co-Inference Systems](https://mlsanity.com/paper/1910.14315/) (2019)
- [JointDNN: An Efficient Training and Inference Engine for Intelligent Mobile Cloud Computing Services](https://mlsanity.com/paper/1801.08618/) (2018)
- [Interpretability of machine learning based prediction models in healthcare](https://mlsanity.com/paper/2002.08596/) (2020)
- [BottleNet: A Deep Learning Architecture for Intelligent Mobile Cloud Computing Services](https://mlsanity.com/paper/1902.01000/) (2019)
- [Deep feature compression for collaborative object detection](https://mlsanity.com/paper/1802.03931/) (2018)
- [Supervised Compression for Resource-Constrained Edge Computing Systems](https://mlsanity.com/paper/2108.11898/) (2021)
- [Neural Compression and Filtering for Edge-assisted Real-time Object Detection in Challenged Networks](https://mlsanity.com/paper/2007.15818/) (2020)
- [Auto-tuning Neural Network Quantization Framework for Collaborative Inference Between the Cloud and Edge](https://mlsanity.com/paper/1812.06426/) (2018)
- [Lightweight compression of neural network feature tensors for collaborative intelligence](https://mlsanity.com/paper/2105.06002/) (2021)

## Built on by (most cited first, 7 of 7)

- [Markerless human pose estimation for biomedical applications: a survey](https://mlsanity.com/paper/2308.00519/) (2023)
- [HiDP: Hierarchical DNN Partitioning for Distributed Inference on Heterogeneous Edge Platforms](https://mlsanity.com/paper/2411.16086/) (2024)
- [Split-Et-Impera: A Framework for the Design of Distributed Deep Learning Applications](https://mlsanity.com/paper/2303.12524/) (2023)
- [MTL-Split: Multi-Task Learning for Edge Devices using Split Computing](https://mlsanity.com/paper/2407.05982/) (2024)
- [Enhancing Split Computing and Early Exit Applications through Predefined Sparsity](https://mlsanity.com/paper/2407.11763/) (2024)
- [Multi-Camera Industrial Open-Set Person Re-Identification and Tracking](https://mlsanity.com/paper/2409.03879/) (2024)
- [Learning based Ge'ez character handwritten recognition](https://mlsanity.com/paper/2411.13350/) (2024)

## Abstract (arXiv)

This work makes a substantial step in the field of split computing, i.e., how to split a deep neural network to host its early part on an embedded device and the rest on a server. So far, potential split locations have been identified exploiting uniquely architectural aspects, i.e., based on the layer sizes. Under this paradigm, the efficacy of the split in terms of accuracy can be evaluated only after having performed the split and retrained the entire pipeline, making an exhaustive evaluation of all the plausible splitting points prohibitive in terms of time. Here we show that not only the architecture of the layers does matter, but the importance of the neurons contained therein too. A neuron is important if its gradient with respect to the correct class decision is high. It follows that a split should be applied right after a layer with a high density of important neurons, in order to preserve the information flowing until then. Upon this idea, we propose Interpretable Split (I-SPLIT): a procedure that identifies the most suitable splitting points by providing a reliable prediction on how well this split will perform in terms of classification accuracy, beforehand of its effective implementation. As a further major contribution of I-SPLIT, we show that the best choice for the splitting point on a multiclass categorization problem depends also on which specific classes the network has to deal with. Exhaustive experiments have been carried out on two networks, VGG16 and ResNet-50, and three datasets, Tiny-Imagenet-200, notMNIST, and Chest X-Ray Pneumonia. The source code is available at https://github.com/vips4/I-Split.