Martínez Soriano: “When I grow up I want to be a data scientist”
Ignacio Martínez Soriano combines his position as a data analyst at the Rafael Méndez Hospital (Murcia) with his work at MMG. Here he is part of the Artificial Intelligence team and is in charge of representing documents with a system of vectors to obtain other semantically similar ones. To do this, he applies the Word2Vec algorithm with the SNOMED CT concepts. This allows him to obtain the meaning of the document, which in turn allows each document to be associated with a normalized code of clinical terminology. The aim is to obtain information that complements DeepAIMed, MMG’s search engine.
“I love the latest approach of DeepAIMed, as it can be integrated as a service.”
Graduated with the first batch of Computer Science at the University of Granada, Ignacio Martínez Soriano can boast of having won the Inforsalud 2012 award. In this competition, where almost all public hospitals usually participate, he obtained the recognition for the Best Oral Communication.
His contribution consisted of presenting a syntactic search engine whose function was to find certain terms in the discharge reports. First, he made a web page and then he added all the high reports, later he routed these reports via an algorithm importance ranking. “I approached it like Google, which made it easier to search for references,” explained the computer scientist.
His true vocation, Artificial Intelligence
On June 7, Ignacio Martínez Soriano will attend the third edition of the CBMS (International Symposium on Computer-Based Medical Systems) to be held in Cordoba. There he will carry out development originated through MMG based on the use of vector generation with an application in the clinical terminology of SNOMED CT. “I created a new concept, the SNOMED 2Vec,” he said.
Soriano has taken the clinical terminations from the descriptions of SNOMED CT concepts and vectorized them. This vectorization has been carried out in a way that allows the search of information to be more efficient than it is in a normal syntactic search. Through this method, he can associate concepts with precision depending on where he searches.
Ignacio Martínez Soriano worked for 18 years as head of Information Systems at the Rafael Méndez Hospital. His interest in acquiring knowledge led him to take all of the Stanford University courses offered at Coursera on data analysis. This is how he realized that it was his true vocation and became an analyst in the same hospital, a position he currently occupies. From there, he focused on Artificial Intelligence, Deep Learning and Machine Learning with a specialization in Natural Language Processing.
“I specialized in a specific technique, Word Embedding, to analyze words and see the relationships with their context in order to identify semantics. With this, you can discover what you’re talking about thanks to the words around you.”
Areas of interest
Ignacio Martínez Soriano has lived the evolution of hospital technology. “I introduced the first information system a HIS (hospital information system) because when I first arrived there wasn’t one at all,” he said. Looking back, he explained that they only had a small file management program, similar to a patient agenda. He then took the software from the Virgen de Arrixaca Hospital, studied it and took it to Lorca to incorporate the HIS.
“Artificial Intelligence is not a matter of the future, it is already a reality with DeepAIMed, always looking for solutions.”
His goals today are to create predictive algorithms using Machine Learning techniques, Business Intelligence and Open Source systems. In addition, data mining and semantic information extraction using natural language processing techniques occupy his areas of interest.