View source for 17551533. FABRICATING DATA USING CONSTRAINTS TRANSLATED FROM TRAINED MACHINE LEARNING MODELS simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)

=FABRICATING DATA USING CONSTRAINTS TRANSLATED FROM TRAINED MACHINE LEARNING MODELS=


==Organization Name==

[[:Category:INTERNATIONAL BUSINESS MACHINES CORPORATION|INTERNATIONAL BUSINESS MACHINES CORPORATION]]


[[Category:INTERNATIONAL BUSINESS MACHINES CORPORATION]]

==Inventor(s)==

[[:Category:Oleg Blinder of Haifa (IL)|Oleg Blinder of Haifa (IL)]][[Category:Oleg Blinder of Haifa (IL)]]

[[:Category:Michael Vinov of Nofit (IL)|Michael Vinov of Nofit (IL)]][[Category:Michael Vinov of Nofit (IL)]]

[[:Category:Omer Yehuda Boehm of Haifa (IL)|Omer Yehuda Boehm of Haifa (IL)]][[Category:Omer Yehuda Boehm of Haifa (IL)]]

[[:Category:Eyal Bin of Haifa (IL)|Eyal Bin of Haifa (IL)]][[Category:Eyal Bin of Haifa (IL)]]

==FABRICATING DATA USING CONSTRAINTS TRANSLATED FROM TRAINED MACHINE LEARNING MODELS - A simplified explanation of the abstract==


This abstract first appeared for US patent application 17551533 titled 'FABRICATING DATA USING CONSTRAINTS TRANSLATED FROM TRAINED MACHINE LEARNING MODELS

==Simplified Explanation==

The abstract of the patent application describes a system that uses machine learning to generate fabricated data based on given constraints. Here is a simplified explanation of the abstract:

* The system includes a processor that receives a data set for training a machine learning model.
* The processor trains the machine learning model using the provided data set.
* The trained machine learning model is then translated into constraint satisfaction problem (CSP) variables and constraints.
* Based on these CSP variables and constraints, the processor generates fabricated data.

Potential Applications:

* Data Augmentation: The generated fabricated data can be used to augment existing datasets, providing more diverse and representative samples for training machine learning models.
* Synthetic Data Generation: The system can be used to generate synthetic data that closely resembles real-world data, which can be useful for various applications such as testing and simulation.
* Privacy Protection: By generating fabricated data, sensitive information can be protected while still allowing the training of machine learning models.

Problems Solved:

* Limited Data Availability: The system addresses the problem of limited data availability by generating additional data that can be used for training machine learning models.
* Data Privacy Concerns: By generating fabricated data, the system helps address privacy concerns by reducing the need for sharing sensitive real-world data.

Benefits:

* Improved Model Performance: The generated fabricated data can help improve the performance of machine learning models by providing more diverse and representative training samples.
* Enhanced Data Privacy: By generating fabricated data, the system reduces the need for sharing sensitive real-world data, thus enhancing data privacy.
* Cost and Time Savings: Instead of collecting or acquiring additional real-world data, the system can generate fabricated data, saving both time and cost.


==Original Abstract Submitted==

An example system includes a processor to receive a data set for training a machine learning model. The processor can train the machine learning model on the data set. The processor can also translate the machine learning model into constraint satisfaction problem (CSP) variables and constraints. The processor can generate fabricated data based on the CSP variables and constraints.



[[Category:G06N3/08]]
[[Category:G06N3/04]]
[[Category:G06N5/00]]