Mohamed Ben Haddou, CEOWithin enterprises, the speed, variety, and volume at which data is gathered drives the need for semantic approaches to establish connections and correlate information and documents to automate decision making process. Semantic technologies extract meaning from quantitative data and text as well as from video, voice, and images. However, the primary concern lies in analyzing the text, gaining a thorough understanding and keeping dialog with the machines. Headquartered in Brussels, Belgium, Mentis is powered by groundbreaking technology and data scientists that help in scrutinizing data and turning them into perceptive knowledge.
Founded in 2005 as a by-product of the Artificial Intelligence Laboratory of the University of Brussels, the company has etched out a name in an array of industries including media and publishing, healthcare, finance, offshore, and energy. To ease the burden of comprehending the complex structure of documents and texts, Mentis developed the AIDA semantic technology, designed to analyze complex document structure. “Traditional semantic technologies examine and convert text into words, without having any advanced analysis that often create inconsistencies in the sentence structure and meaning in the text,” says Mohamed Ben Haddou, CEO, Mentis. The big data platform, AIDA is built on the concept that it needs to retain the complexity in the document’s structure in the analysis process as well. The platform helps in understanding the words—location, relation with the next word, sentence structure, and drags these comprehensions thoroughly into the analysis pipeline.
“AIDA efficiently converts text from document space into concept space, where words are grouped, based on semantic or syntactic content,” adds Haddou. The platform provides an abstract framework, allowing development teams to create their own text analytics systems, along with a powerful implementation of general purpose methods for structuring and analyzing documents. In addition, AIDA extracts common entities such as names and places and advanced entities like specific fields from unstructured data using proprietary extraction methods. “We also leverage various online resources to search, prioritize, and identify the accurate entity,” explains Haddou.
AIDA efficiently converts text from document space into concept space, where words are grouped, based on semantic or syntactic content
In one instance, one of the clients in the digital media industry required an effective system to interact and recommend movies to audiences. The client wanted their users to answer three questions to make the whole recommendation process easier. Although, the customer had the idea of the recommendation process, they lacked the relevant resources to implement this in reality. By approaching Mentis and deploying their AIDA platform, the digital media firm was able to analyze movies and convert it into semantic space, while renovating the whole set of questions to put forward to their users. The next process was to fathom the user behavior and their likes and dislikes of the movies based on the ratings given. This information was used by Mentis to help the firm in making spontaneous recommendation for new movies.
Mentis is also moving their AIDA platform to the cloud for driving better opportunities. Through these enhancements, Mentis is providing semantic search in news, and will soon be launching a mobile application to serve smart news to customers. “AIDA is also designed to assist in web mining and analysis of social media—blogs, forums as well as news sites,” points out Haddou.
Amalgamating artificial intelligence with smart option, the company plans to enhance their AIDA platform by adding deep learning capabilities. “We also plan to convert AIDA into an embedded platform that can analyze all kind of devices, which will initiate the plan of getting rid of unwanted data in order to solve the big data problem,” concludes Haddou.