データサイエンティスト研究領域

OVERVIEW -データサイエンス研究領域-

MTI Technology AIラボのデータサイエンティストは、データサイエンスに関わる様々な研究を行っており、それらの研究を通して、お客様のAI開発を支えております。

具体的には、MTI Technology AIラボではデータサイエンティストの基礎となる数学及び統計学をコア知識として、以下のようなデータサイエンスに関する技術を日々研究しております。

-Object Character Recognition (OCR)

-自然言語処理(NLP)

-データ分析、データ予測

-Operational Research

-Intelligent Information Retrieval

-機械学習/ディープラーニング/ニューラルネットワーク


1.Object Character Recognition (OCR)

OCR(Object Character Recognition) is an AI technology combining computer vision, natural language and domain knowledge to extract information from image-based documents, such as scanned document, camera-taken document, or PDF, into machine-readable text. Extracted information can be then classified into structured-data format depending on demand of application.In AI Lab, our Data Scientist have developed our OCR system for reading documents that are used in activities of health care sector like prescription.Another OCR system for reading receipts and invoices has been also developed by our AI Lab in the recent years.


2.自然言語処理(NLP)

NLP(Natural Language Processing) is an AI technology which combines both computer and human natural languages. NLP(Natural Language Processing) algorithms help to use the computers to process and analyze large amounts of natural language data. As a result, the computer can help to analyze information from language data as for instance in extracting opinions from product reviews.  This technology also supports humans in automating some processes relating to communication, as for instance in creating dialog systems for smart technologies.

In the MTI Technology AI Lab, our Data Scientist in Vietnam have been working on a natural-language generation system.


3.データ分析、データ予測

In today’s information-rich world, business analysis transforms into business intelligence, in which peoples use advanced data analytics and modeling skills, to get important insight information and make decision. The more the data is big and complex, the more analytics and modeling skill are sophisticated. Nowadays, beside statistical models, machine learning and deep neural networks are also used to analyze the big data, help user to make more efficient decisions and more accurate predictions.

In MTI Technology AI Lab in Vietnam, our Data Scientist have been analyzing a large amount of health care data from projects in cooperation with hospitals and research institutes in Japan. Our results support medicine doctors or health care divisions in diagnosing the disease, through medical applications and recommendation systems.


4.Operational Research

Operational Research is the discipline of using models to aid decision-making in complex implementation problems, mostly for operational or logistical processes. Operational Research uses the sophisticated mathematical algorithms to optimize the operations, reduce the operating cost, but still improve the accuracy and efficiency of the system when compare with humans.


5.Intelligent Information Retrieval

A large amount of information is being endlessly created and cumulated over time, especially the knowledge and skills of experts in specific domains. The information can be structured data such as a well-defined database, but mostly is unstructured data such as raw text from documents or scraped text from websites. This has led to a demand for intelligent and efficient algorithms in research and text retrieval. Today, most techniques developed in AI have been applied to retrieval systems with more or less success, but they have been certainly contributing to creating novel value-added services.

In AI Lab, our data scientist initialized an intelligent information retrieval system called “Smart Manual”. This system is designed to deal firstly with a small and specific domain knowledge, such as a system of company rules, procedures and services. Our further target is an expert system that cans professionally assist medicine doctors in examination, diagnosis and treatment.


6.機械学習/ディープラーニング/ニューラルネットワーク

Today, weather data is increasingly collected with high resolution and high frequency, thanks to modern measurement equipment such as ground radars, air-bone radars, satellites. In the recent years, besides traditional methods, machine learning and deep learning have been also intensively researched to help improve the accuracy of forecasting.

We AI Lab has competences in processing data from various type of weather radars. In the last recent years, we had an opportunity to work on the data of radars network. Our main achievement is building a system for processing, analyzing and forecasting weather radar images by using Machine Learning and Deep Neural Networks.

Please check more Technical info from the our Technical Blog.

Please check Actual Project Example from Project Example.