Skip to content

A.I-Patent Granted

Artificial Intelligence-Patent Granted #10540439 & #10073890 & #10621499

Technologies we use

A Technology That Will Make Your Life Easier

CSC Beyond has achieved a published patent in the USPTO (United States Patent and Trademark Office): Patent Name:

SYSTEMS AND METHODS FOR IDENTIFYING EVIDENTIARY INFORMATION

Publication Number: 20170300470 Patent Abstract: Systems and methods for semantically analyzing digital information.

A cognitive engine is configured to determine useful evidentiary information from large digital content data sets. Further, the cognitive engine can analyze or manipulate the evidentiary information to derive data needed to solve problems, identify issues, and identify patterns.

The results can then be applied to any application, interface, or automation as appropriate.

Systems and methods for semantically analyzing digital information. A cognitive engine is configured to determine useful evidentiary information from large digital content data sets. Further, the cognitive engine can analyze or manipulate the evidentiary information to derive data needed to solve problems, identify issues, and identify patterns. The results can then be applied to any application, interface, or automation as appropriate.


Historically, in order to understand or appreciate a particular topic, one would need to read a myriad of resources and manually synthesize the contents of the resources. Conclusions or theories or broadly-categorized “results” could then be made based on this synthesis. This, of course, is a time-intensive and user-specific process.

However, as digital information becomes more and more prevalent and an increasing number of resources become available in a digital format in online databases, there is an opportunity to automate the reading and understanding of resources in order to derive useful knowledge across a wide variety of topics and for any generic user.

For example, the article Joint Learning of Ontology and Semantic Parser from Text by Starc and Mladenic from Jozef Stefan International Postgraduate School and published November 2015, describes a semantic parsing approach to analysis of digital content. However, this approach uses an ontology direction, which is the basis of supervised learning. The ontology defines the pathways or steps for reading the content. However, this approach requires supervisor intervention for the ontology direction.

In another example, the article Natural Language Processing (Almost) from Scratch by Collobert et al. from Journal of Machine Learning Research and published August 2011, describes an approach that utilizes neural networks based on supervised learning and requires a prior data set with predefined results for semantic understanding.

In another example, the article A Unified Architecture for Natural Language Processing: Deep Neural Networks with Multitask Learning by Collobert and Weston of NEC Labs America, describes an approach that also utilizes neural networks based on supervised learning.

Therefore, there is a need for systems and methods that provide for real-time, accurate, and verifiable identification and analysis of digital content that is more sophisticated than a basic key word analysis and which requires less supervision than existing systems.

Need Help? Call Our Support Team

We’re looking for people who share our passion for bringing Qatar online