Hiroyoshi Toyoshiba, Research Director and Chief Technology Officer of Life Science AI Business Division, FRONTEO“With the aid of our domain experts, we offer consultation and advanced AI-based data analysis services to help bridge the gap between data scientists and health systems. Our service is a complete package of consultations consisting of data, algorithms, outputs, and system integrations,” expresses Hiroyoshi Toyoshiba, research director and CTO of FRONTEO’s Life Science AI Business Division.
The company applies state-of-the-art word and document embedding techniques via its AI engine—Concept Encoder—which allows users to search through databases more quickly with complete transparency and reproducibility. Concept Encoder embeds all sentences and key words used by a physician or administrator into a vector when a patient’s documents are uploaded. “The embedding process is based on the Distributional Hypothesis that states that a word is characterized by the company it keeps,” adds Toyoshiba. Concept Encoder first defines and understands the meaning of each word or phrase by itself and its similarity to other meanings. FRONTEO leverages machine learning (ML) technologies and methodologies to embedded vectors, which significantly increases the accuracy of this process through the AI engine.
The company also offers three distinctive services, where its AI-based Diagnosis Support gathers information from various conversations between patients and their respective physicians. For instance, a doctor can administer a mini-mental state exam (MMSE), and FRONTEO’s ML capabilities provide healthcare professionals with the possibility of quickly detecting Mild Cognitive Impairment (MCI) and dementia. The company’s Healthcare Operational Support service analyses hospital incident reports, including data on patient conditions from electronic folders as well as the text data of free comments found in daily records of examination and treatment. This allows the AI to calculate, predict, and identify patients who have a high fall risk in a hospital. Lastly, as a Pharmaceutical Industry Support service, the company’s AI aids drug discovery and enables the analysis of various text data with the aim of making pharmacovigilance work more efficient, while delivering high quality of data. This system is being further developed to predict causal or responsive genes when a disease name is provided as a query, generating a molecular map linking the gene’s sets. “We also offer a solution that automatically classifies official documents from the government into groups that are most relevant,” says Toyoshiba.
Moving forward, the company aims to penetrate healthcare markets beyond the borders of Japan. Owing to the country’s advanced ageing society, FRONTEO is also going to focus its efforts toward developing dementia diagnosis support systems, based on neuro-linguistic programming (NLP) and have them approved as patented medical devices to help improve drug discovery across the globe.
The company also offers three distinctive services, where its AI-based Diagnosis Support gathers information from various conversations between patients and their respective physicians. For instance, a doctor can administer a mini-mental state exam (MMSE), and FRONTEO’s ML capabilities provide healthcare professionals with the possibility of quickly detecting Mild Cognitive Impairment (MCI) and dementia. The company’s Healthcare Operational Support service analyses hospital incident reports, including data on patient conditions from electronic folders as well as the text data of free comments found in daily records of examination and treatment. This allows the AI to calculate, predict, and identify patients who have a high fall risk in a hospital. Lastly, as a Pharmaceutical Industry Support service, the company’s AI aids drug discovery and enables the analysis of various text data with the aim of making pharmacovigilance work more efficient, while delivering high quality of data. This system is being further developed to predict causal or responsive genes when a disease name is provided as a query, generating a molecular map linking the gene’s sets. “We also offer a solution that automatically classifies official documents from the government into groups that are most relevant,” says Toyoshiba.
Moving forward, the company aims to penetrate healthcare markets beyond the borders of Japan. Owing to the country’s advanced ageing society, FRONTEO is also going to focus its efforts toward developing dementia diagnosis support systems, based on neuro-linguistic programming (NLP) and have them approved as patented medical devices to help improve drug discovery across the globe.
