18466469. ELECTRONIC APPARATUS AND CONTROL METHOD THEREOF simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

ELECTRONIC APPARATUS AND CONTROL METHOD THEREOF

Organization Name

SAMSUNG ELECTRONICS CO., LTD.

Inventor(s)

Krzysztof Arendt of Warszawa (PL)

Grzegorz Stefanski of Warszawa (PL)

Piotr Masztalski of Warszawa (PL)

Artur Szumaczuk of Warszawa (PL)

Jakub Tkaczuk of Warszawa (PL)

Mateusz Matuszewski of Warszawa (PL)

Michal Swiatek of Warszawa (PL)

Tymoteusz Oleniecki of Warszawa (PL)

ELECTRONIC APPARATUS AND CONTROL METHOD THEREOF - A simplified explanation of the abstract

This abstract first appeared for US patent application 18466469 titled 'ELECTRONIC APPARATUS AND CONTROL METHOD THEREOF

Simplified Explanation

The abstract describes an electronic apparatus that uses artificial intelligence to classify data. Here are the key points:

• The electronic apparatus includes a memory and a processor.
• The processor acquires training data and embedding vectors for the data.
• An artificial intelligence model is trained to classify the training data based on the embedding vectors.
• The model identifies a misclassified embedding vector and finds the closest embedding vector in the embedding space.
• A synthetic embedding vector is created to connect the misclassified vector to the closest vector.
• The artificial intelligence model is re-trained by adding the synthetic embedding vector to the training data.

Potential applications of this technology:

• Improving the accuracy of artificial intelligence models in various fields such as image recognition, natural language processing, and data analysis.
• Enhancing the performance of recommendation systems by refining the classification of user preferences.
• Assisting in fraud detection by identifying patterns in data that may indicate fraudulent behavior.

Problems solved by this technology:

• Addressing misclassifications by identifying and correcting errors made by artificial intelligence models.
• Improving the interpretability and explainability of artificial intelligence models by visualizing the embedding space and the paths connecting misclassified vectors.

Benefits of this technology:

• Increased accuracy and reliability of artificial intelligence models.
• Enhanced understanding of the decision-making process of artificial intelligence models.
• Improved efficiency in training and re-training artificial intelligence models.

Original Abstract Submitted

An electronic apparatus is provided. The electronic apparatus includes a memory and a processor, wherein the processor is configured to, by executing the at least one instruction, acquire a plurality of training data; acquire a plurality of embedding vectors that are mappable to an embedding space for the plurality of training data, respectively; train an artificial intelligence model classifying the plurality of training data based on the plurality of embedding vectors, identify an embedding vector misclassified by the artificial intelligence model among the plurality of embedding vectors, identify an embedding vector closest to the misclassified embedding vector in the embedding space, acquire a synthetic embedding vector corresponding to a path connecting the misclassified embedding vector to the embedding vector closest to the misclassified embedding vector in the embedding space, and re-train the artificial intelligence model by adding the synthetic embedding vector to the training data.