18668637. METHOD AND APPARATUS FOR SEARCHING FOR NEURAL NETWORK ENSEMBLE MODEL, AND ELECTRONIC DEVICE simplified abstract (Huawei Technologies Co., Ltd.)
Contents
METHOD AND APPARATUS FOR SEARCHING FOR NEURAL NETWORK ENSEMBLE MODEL, AND ELECTRONIC DEVICE
Organization Name
Inventor(s)
Pedro Esperanca of London (GB)
Fabio Maria Carlucci of London (GB)
METHOD AND APPARATUS FOR SEARCHING FOR NEURAL NETWORK ENSEMBLE MODEL, AND ELECTRONIC DEVICE - A simplified explanation of the abstract
This abstract first appeared for US patent application 18668637 titled 'METHOD AND APPARATUS FOR SEARCHING FOR NEURAL NETWORK ENSEMBLE MODEL, AND ELECTRONIC DEVICE
The abstract describes a method for searching for a neural network architecture ensemble model using a distributional neural network architecture search algorithm.
- Obtaining a dataset with samples and annotations in a classification task.
- Determining hyperparameters of a neural network architecture distribution.
- Sampling valid neural network architectures from the distribution.
- Training and evaluating the architectures on the dataset to obtain performance indicators.
- Identifying neural network architecture distributions that share hyperparameters to form a candidate pool of base learners.
- Determining a surrogate model to predict test performance of base learners.
- Forming an ensemble model with diverse base learners that meet task scenario requirements.
Potential Applications: - Improving the efficiency and accuracy of neural network architecture search. - Enhancing the performance of classification tasks in various fields such as image recognition, natural language processing, and more.
Problems Solved: - Streamlining the process of finding optimal neural network architectures. - Reducing the time and resources required for architecture search.
Benefits: - Increased accuracy and efficiency in classification tasks. - Enhanced performance of neural network models. - Facilitates the creation of ensemble models for improved results.
Commercial Applications: "Optimizing Neural Network Architecture Search for Enhanced Classification Performance"
Questions about the technology: 1. How does this method compare to traditional neural network architecture search approaches? 2. What are the potential limitations of using a distributional neural network architecture search algorithm?
Original Abstract Submitted
Disclosed is a method for searching for a neural network architecture ensemble model. The method includes: obtaining a dataset, where the dataset includes a sample and an annotation in a classification task; performing search by using a distributional neural network architecture search algorithm, including: determining a hyperparameter of a neural network architecture distribution; sampling a valid neural network architecture from the architecture distribution defined by the hyperparameter; training and evaluating the neural network architecture on the dataset, to obtain a performance indicator; determining, based on the performance indicator, neural network architecture distributions that share the hyperparameter, to obtain a candidate pool of base learners; and determining a surrogate model; and predicting test performance of the base learner in the candidate pool by using the surrogate model, and determining that k diverse base learners that meet a task scenario requirement form an ensemble model.
