# Unifying and Merging Well-trained Deep Neural Networks for Inference Stage

arXiv: https://arxiv.org/abs/1805.04980 · HTML: https://mlsanity.com/paper/1805.04980/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 May 2018.

## Built on by (most cited first, 8 of 8)

- [Omni-Video: Democratizing Unified Video Understanding and Generation](https://mlsanity.com/paper/2507.06119/) (2025)
- [Collaborative Intelligence: Challenges and Opportunities](https://mlsanity.com/paper/2102.06841/) (2021)
- [On Cross-Layer Alignment for Model Fusion of Heterogeneous Neural Networks](https://mlsanity.com/paper/2110.15538/) (2021)
- [RED: A Systematic Real-Time Scheduling Approach for Robotic Environmental Dynamics](https://mlsanity.com/paper/2308.15368/) (2023)
- [Structure Is Not Enough: Leveraging Behavior for Neural Network Weight Reconstruction](https://mlsanity.com/paper/2503.17138/) (2025)
- [Representation Consolidation for Training Expert Students](https://mlsanity.com/paper/2107.08039/) (2021)
- [MIMONet: Multi-Input Multi-Output On-Device Deep Learning](https://mlsanity.com/paper/2307.11962/) (2023)
- [Model Composition: Can Multiple Neural Networks Be Combined into a Single Network Using Only Unlabeled Data?](https://mlsanity.com/paper/2110.10369/) (2021)

## Abstract (arXiv)

We propose a novel method to merge convolutional neural-nets for the inference stage. Given two well-trained networks that may have different architectures that handle different tasks, our method aligns the layers of the original networks and merges them into a unified model by sharing the representative codes of weights. The shared weights are further re-trained to fine-tune the performance of the merged model. The proposed method effectively produces a compact model that may run original tasks simultaneously on resource-limited devices. As it preserves the general architectures and leverages the co-used weights of well-trained networks, a substantial training overhead can be reduced to shorten the system development time. Experimental results demonstrate a satisfactory performance and validate the effectiveness of the method.