Maximum-Entropy Based Categorical Cross-Entropy Loss Function for Noisy Non-Convex Optimization Problems
Social Science Research Network(2023)
Abstract
We propose a novel Maximum Categorical Cross-Entropy (MCCE) loss for Deep Neural Networks (DNNs), specifically Convolutional Neural Networks (CNNs) in image classification tasks. The novelty of our loss is established on the assumption that a priori information of the input dataset in conjunction with entropic distribution of all significant layers in a DNN constrain the problem space enough for enhanced exploration of the solution space. Computing a novel reconstruction loss of the hidden layers through the use of ME measures of the convolutional kernels in real-time and utilizing the difference between the entropic distributions of the convolutional kernels and the input dataset as a monitoring function to penalize overly complex models which in effect, allows our proposed MCCE loss to provide both L1 and L2 regularization effects. Our proposed MCCE loss provides superior performance relative to CCE, Focal and Hinge losses. The average performance gain of models trained with the proposed MCCE loss function is 3.51% and up to 23.1% relative to the traditional CCE loss across 6 unique datasets and 5 different CNN models.
MoreTranslated text
Key words
Backpropagation Learning
求助PDF
上传PDF
View via Publisher
AI Read Science
AI Summary
AI Summary is the key point extracted automatically understanding the full text of the paper, including the background, methods, results, conclusions, icons and other key content, so that you can get the outline of the paper at a glance.
Example
Background
Key content
Introduction
Methods
Results
Related work
Fund
Key content
- Pretraining has recently greatly promoted the development of natural language processing (NLP)
- We show that M6 outperforms the baselines in multimodal downstream tasks, and the large M6 with 10 parameters can reach a better performance
- We propose a method called M6 that is able to process information of multiple modalities and perform both single-modal and cross-modal understanding and generation
- The model is scaled to large model with 10 billion parameters with sophisticated deployment, and the 10 -parameter M6-large is the largest pretrained model in Chinese
- Experimental results show that our proposed M6 outperforms the baseline in a number of downstream tasks concerning both single modality and multiple modalities We will continue the pretraining of extremely large models by increasing data to explore the limit of its performance
Upload PDF to Generate Summary
Must-Reading Tree
Example

Generate MRT to find the research sequence of this paper
Data Disclaimer
The page data are from open Internet sources, cooperative publishers and automatic analysis results through AI technology. We do not make any commitments and guarantees for the validity, accuracy, correctness, reliability, completeness and timeliness of the page data. If you have any questions, please contact us by email: report@aminer.cn
Chat Paper
Summary is being generated by the instructions you defined