3 papers
cs.LG2022
Breaking the Architecture Barrier: A Method for Efficient Knowledge Transfer Across Networks
Maciej A. Czyzewski, Daniel Nowak, Kamil Piechowiak
Transfer learning is a popular technique for improving the performance of neural networks. However, existing methods are limited to transferring parameters between networks with sa…
cs.LG2021
Transfer Learning Between Different Architectures Via Weights Injection
Maciej A. Czyzewski
This work presents a naive algorithm for parameter transfer between different architectures with a computationally cheap injection technique (which does not require data). The prim…
cs.LG2020
batchboost: regularization for stabilizing training with resistance to underfitting & overfitting
Maciej A. Czyzewski
Overfitting & underfitting and stable training are an important challenges in machine learning. Current approaches for these issues are mixup, SamplePairing and BC learning. In our…