18045664. DETERMINATION OF EVENT TYPES FROM AUTOENCODER-BASED UNSUPERVISED EVENT DETECTION simplified abstract (Dell Products L.P.)
Contents
- 1 DETERMINATION OF EVENT TYPES FROM AUTOENCODER-BASED UNSUPERVISED EVENT DETECTION
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 DETERMINATION OF EVENT TYPES FROM AUTOENCODER-BASED UNSUPERVISED EVENT DETECTION - 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
DETERMINATION OF EVENT TYPES FROM AUTOENCODER-BASED UNSUPERVISED EVENT DETECTION
Organization Name
Inventor(s)
Paulo Abelha Ferreira of Rio de Janeiro (BR)
Vinicius Michel Gottin of Rio de Janeiro (BR)
[[:Category:Herberth Birck Fr�hlich of Florianópolis (BR)|Herberth Birck Fr�hlich of Florianópolis (BR)]][[Category:Herberth Birck Fr�hlich of Florianópolis (BR)]]
DETERMINATION OF EVENT TYPES FROM AUTOENCODER-BASED UNSUPERVISED EVENT DETECTION - A simplified explanation of the abstract
This abstract first appeared for US patent application 18045664 titled 'DETERMINATION OF EVENT TYPES FROM AUTOENCODER-BASED UNSUPERVISED EVENT DETECTION
Simplified Explanation
Unsupervised event detection is disclosed in this patent application. The process involves clustering reconstruction data generated from input samples using a machine learning model. By labeling one or more samples within a cluster, all samples in that cluster can be assigned the same label. During inference, any input sample that produces a similar reconstructed sample can be given the label previously assigned to the cluster.
- Clustering reconstruction data from input samples
- Labeling samples within a cluster to assign the same label to all samples in that cluster
- Using machine learning model for event detection
- Applying labels to input samples based on similarity to reconstructed samples
Potential Applications
The technology described in this patent application could be applied in various fields such as:
- Anomaly detection in data streams
- Fraud detection in financial transactions
- Monitoring and alerting systems for abnormal events in industrial processes
Problems Solved
This technology addresses the following issues:
- Identifying and categorizing events in large datasets
- Automating the process of event detection without the need for manual labeling
- Improving the efficiency and accuracy of event detection systems
Benefits
The benefits of this technology include:
- Faster and more accurate event detection
- Reduction in false positives and false negatives
- Scalability for processing large volumes of data
Potential Commercial Applications
The technology could have commercial applications in:
- Cybersecurity for detecting unusual network activity
- Healthcare for monitoring patient data and detecting anomalies
- Retail for fraud detection in online transactions
Possible Prior Art
One possible prior art for this technology could be:
- Clustering algorithms used in unsupervised machine learning for data analysis
What are the limitations of this technology in real-world applications?
The limitations of this technology in real-world applications may include:
- Sensitivity to noise in input data
- Dependence on the quality of the machine learning model
- Interpretability of the clustering results
How does this technology compare to supervised event detection methods in terms of accuracy and efficiency?
This technology offers advantages in terms of efficiency by eliminating the need for manual labeling of data, but it may have limitations in accuracy compared to supervised methods that require labeled training data.
Original Abstract Submitted
Unsupervised event detection is disclosed. Reconstruction data resulting from processing input samples with a machine learning model is clustered. By labeling one or more samples of a cluster, all of the samples in the same cluster can be labeled the same. During inference, any input sample generating a similar reconstructed sample can be given the label previously applied to the cluster.