18021089. LEARNING APPARATUS, TRAINED MODEL GENERATION METHOD, CLASSIFICATION APPARATUS, CLASSIFICATION METHOD, AND COMPUTER READABLE RECORDING MEDIUM simplified abstract (NEC Corporation)
LEARNING APPARATUS, TRAINED MODEL GENERATION METHOD, CLASSIFICATION APPARATUS, CLASSIFICATION METHOD, AND COMPUTER READABLE RECORDING MEDIUM
Organization Name
Inventor(s)
LEARNING APPARATUS, TRAINED MODEL GENERATION METHOD, CLASSIFICATION APPARATUS, CLASSIFICATION METHOD, AND COMPUTER READABLE RECORDING MEDIUM - A simplified explanation of the abstract
This abstract first appeared for US patent application 18021089 titled 'LEARNING APPARATUS, TRAINED MODEL GENERATION METHOD, CLASSIFICATION APPARATUS, CLASSIFICATION METHOD, AND COMPUTER READABLE RECORDING MEDIUM
Simplified Explanation
The learning apparatus described in the patent application is designed to improve the performance of a score function through machine learning. It achieves this by selecting training data with positive or negative labels and calculating scores for each data point using the score function. The apparatus then determines the minimum score for the positive-labeled data and the maximum score for the negative-labeled data.
The apparatus includes a pair generation unit that selects training data with scores equal to or higher than the minimum score and equal to or lower than the maximum score. It generates pairs consisting of a positive example and a negative example. The optimization unit updates a parameter of the score function through machine learning to increase the number of pairs where the score of the positive-labeled data is higher than the score of the negative-labeled data.
- The learning apparatus calculates scores for training data using a score function.
- It determines the minimum score for positive-labeled data and the maximum score for negative-labeled data.
- The pair generation unit selects training data within the score range and generates pairs of positive and negative examples.
- The optimization unit updates a parameter of the score function through machine learning to increase the number of pairs where the positive example has a higher score than the negative example.
Potential applications of this technology:
- Improving the performance of recommendation systems by learning from user preferences and feedback.
- Enhancing the accuracy of sentiment analysis in natural language processing tasks.
- Optimizing search engine ranking algorithms by learning from user interactions and relevance feedback.
Problems solved by this technology:
- Addressing the challenge of accurately ranking and recommending items based on user preferences.
- Overcoming the limitations of traditional score functions in capturing complex patterns and relationships in data.
- Improving the discrimination power of machine learning models in binary classification tasks.
Benefits of this technology:
- Enhanced accuracy and performance of score functions in various applications.
- Improved user experience through more personalized recommendations and search results.
- Increased efficiency and effectiveness of machine learning models in binary classification tasks.
Original Abstract Submitted
A learning apparatus includes: a score calculation unit that calculates scores by inputting training data with positive or negative labels to a score function; a score specification unit that specifies the lowest one of the scores for the training data with positive labels as a minimum score, and specifies the highest one of the scores for the training data with negative labels as a maximum score; a pair generation unit that selects training data for which the scores are equal to or higher than the minimum and equal to or lower than the maximum, and generates pairs of a positive example and a negative example, and an optimization unit that updates a parameter of the score function through machine learning so as to increase the number of pairs in which a score of training data with positive label is higher than a score of training data with negative label.
