17952577. METHODS AND SYSTEMS FOR AUTOMATED CREATION OF ANNOTATED DATA AND TRAINING OF A MACHINE LEARNING MODEL THEREFROM simplified abstract (Robert Bosch GmbH)
Contents
- 1 METHODS AND SYSTEMS FOR AUTOMATED CREATION OF ANNOTATED DATA AND TRAINING OF A MACHINE LEARNING MODEL THEREFROM
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 METHODS AND SYSTEMS FOR AUTOMATED CREATION OF ANNOTATED DATA AND TRAINING OF A MACHINE LEARNING MODEL THEREFROM - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Original Abstract Submitted
METHODS AND SYSTEMS FOR AUTOMATED CREATION OF ANNOTATED DATA AND TRAINING OF A MACHINE LEARNING MODEL THEREFROM
Organization Name
Inventor(s)
Yuheng Wang of Centereach NY (US)
Bingqing Wang of San Jose CA (US)
Zhe Feng of Mountain View CA (US)
METHODS AND SYSTEMS FOR AUTOMATED CREATION OF ANNOTATED DATA AND TRAINING OF A MACHINE LEARNING MODEL THEREFROM - A simplified explanation of the abstract
This abstract first appeared for US patent application 17952577 titled 'METHODS AND SYSTEMS FOR AUTOMATED CREATION OF ANNOTATED DATA AND TRAINING OF A MACHINE LEARNING MODEL THEREFROM
Simplified Explanation
The systems and methods described herein are directed to a Co-Augmentation framework that may learn new rules and labels simultaneously from unlabeled data with a small set of seed rules and a few manually labeled training data. The augmented rules and labels are further used to train supervised neural network models. Specifically, the systems and methods described herein include two major components: a rule augmenter, and a label augmenter. The rule augmenter is directed to learning new rules, which can be used to obtain weak labels from unlabeled data. The label augmenter is directed to learning new labels from unlabeled data. The Co-Augmentation framework is an iterative learning process which generates and refines a high precision set. At each iteration, both the rule augmenter and label augmenter will contribute new and more accurate labels to the high precision set, which is in turn used to train both the rule augmenter and label augmenter.
- Rule augmenter learns new rules from unlabeled data
- Label augmenter learns new labels from unlabeled data
- Co-Augmentation framework iteratively generates and refines a high precision set
- High precision set is used to train both the rule augmenter and label augmenter
Potential Applications
This technology can be applied in:
- Natural language processing
- Data mining
- Information retrieval
Problems Solved
This technology helps in:
- Improving the accuracy of supervised neural network models
- Learning new rules and labels from unlabeled data
- Enhancing the efficiency of training processes
Benefits
The benefits of this technology include:
- Increased accuracy in classification tasks
- Reduced manual labeling efforts
- Enhanced performance of machine learning models
Potential Commercial Applications
This technology can be commercially be used in:
- Sentiment analysis tools
- Recommendation systems
- Fraud detection algorithms
Possible Prior Art
One possible prior art for this technology could be:
- Co-training algorithms in machine learning
Unanswered Questions
How does the Co-Augmentation framework handle noisy data?
The abstract does not mention how the system deals with noisy data during the learning process. Noise in the data can affect the accuracy of the rules and labels generated.
Are there any limitations to the size of the seed rules and manually labeled training data?
The abstract does not specify any limitations on the size of the seed rules or training data. It would be important to understand if there are any constraints on the amount of initial data required for the system to function effectively.
Original Abstract Submitted
The systems and methods described herein are directed to a Co-Augmentation framework that may learn new rules and labels simultaneously from unlabeled data with a small set of seed rules and a few manually labeled training data. The augmented rules and labels are further used to train supervised neural network models. Specifically, the systems and methods described herein include two major components: a rule augmenter, and a label augmenter. The rule augmenter is directed to learning new rules, which can be used to obtain weak labels from unlabeled data. The label augmenter is directed to learning new labels from unlabeled data. The Co-Augmentation framework is an iterative learning process which generates and refines a high precision set. At each iteration, both the rule augmenter and label augmenter will contribute new and more accurate labels to the high precision set, which is in turn used to train both the rule augmenter and label augmenter.