# CalexNet: Soft Cascade-Aligned Training and Calibration for Lightweight Early-Exit Branches

arXiv: https://arxiv.org/abs/2509.08318 · HTML: https://mlsanity.com/paper/2509.08318/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Early exit and edge inference, one of 12 directions in Efficient LLM inference and serving.
- Builds on 24 of its 40 references that are in the map.
- First posted on arXiv 10 Sep 2025.

## Placed in

- Efficient LLM inference and serving / [Early exit and edge inference](https://mlsanity.com/inference/early-exit-and-edge-inference/)

## Builds on (most cited first, 24 of 24)

- [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)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Learning both Weights and Connections for Efficient Neural Networks](https://mlsanity.com/paper/1506.02626/) (2015)
- [Identity Mappings in Deep Residual Networks](https://mlsanity.com/paper/1603.05027/) (2016)
- [Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference](https://mlsanity.com/paper/1712.05877/) (2017)
- [Cascade R-CNN: Delving into High Quality Object Detection](https://mlsanity.com/paper/1712.00726/) (2017)
- [Class-Balanced Loss Based on Effective Number of Samples](https://mlsanity.com/paper/1901.05555/) (2019)
- [BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks](https://mlsanity.com/paper/1709.01686/) (2017)
- [Dynamic Neural Networks: A Survey](https://mlsanity.com/paper/2102.04906/) (2021)
- [LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding](https://mlsanity.com/paper/2404.16710/) (2024)
- [Pruning and Quantization for Deep Neural Network Acceleration: A Survey](https://mlsanity.com/paper/2101.09671/) (2021)
- [What is the Effect of Importance Weighting in Deep Learning?](https://mlsanity.com/paper/1812.03372/) (2018)
- [A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendations](https://mlsanity.com/paper/2308.06767/) (2023)
- [Split Computing and Early Exiting for Deep Learning Applications: Survey and Research Challenges](https://mlsanity.com/paper/2103.04505/) (2021)
- [Adaptive Inference through Early-Exit Networks: Design, Challenges and Directions](https://mlsanity.com/paper/2106.05022/) (2021)
- [Improved Techniques for Training Adaptive Deep Networks](https://mlsanity.com/paper/1908.06294/) (2019)
- [Predictive Exit: Prediction of Fine-Grained Early Exits for Computation- and Energy-Efficient Inference](https://mlsanity.com/paper/2206.04685/) (2022)
- [Jointly-Learned Exit and Inference for a Dynamic Neural Network : JEI-DNN](https://mlsanity.com/paper/2310.09163/) (2023)
- [Computation-efficient Deep Learning for Computer Vision: A Survey](https://mlsanity.com/paper/2308.13998/) (2023)
- [Boosted Dynamic Neural Networks](https://mlsanity.com/paper/2211.16726/) (2022)
- [PTEENet: Post-Trained Early-Exit Neural Networks Augmentation for Inference Cost Optimization](https://mlsanity.com/paper/2501.02508/) (2025)
- [DistrEE: Distributed Early Exit of Deep Neural Network Inference on Edge Devices](https://mlsanity.com/paper/2502.15735/) (2025)

## Abstract (arXiv)

Early-exit cascades over a frozen convolutional backbone enable adaptive inference but suffer from three sources of train-inference mismatch: branches train on samples they will never see at inference, their per-class precision thresholds are calibrated on the wrong distribution, and the standard cross-entropy target on backbone argmax labels discards the backbone's uncertainty signal. We close all three gaps with CalexNet (Cascade-Aligned Early eXits), a training-recipe-only modification: branches train under continuously-weighted importance sampling that matches the cascade-survivor distribution; per-class precision thresholds are calibrated on the actual cascade-survivor subset of the validation set; and the classification head is trained against the backbone's full softmax via a temperature-scaled KL objective. Combined with an augmented prototype-pooling branch head, CalexNet is evaluated on ResNet18 and ResNet50 backbones across CIFAR-100 (20-superclass coarse, the harder primary setting) and CINIC-10 (10-class, the easier cross-validation counterpart). On the accuracy-FLOPs Pareto frontier, CalexNet matches or exceeds three published baselines (PTEEnet, ZTW, BoostNet) and a within-paper "no-alignment, no-KD" reference. The largest gains appear in the practically relevant 30-70% FLOPs-reduction regime and are stable across n=3 training seeds. CalexNet requires no inference-time architectural change and is a drop-in for any frozen-backbone early-exit cascade.