Service

Organizing Text Data into Knowledge

There are many values in a text because our thoughts are written into it. We organize text data into knowledge and utilize knowledge to boost data-driven systems and data-driven decisions.

Cohaku

COre HAck tools for Knowledge Unifying

Cohaku Logo

We develop knowledge processing engine, Cohaku (Core Hack tools for Knowledge Unifying), the concept of which is to organize text data into knowledge and utilize knowledge to build AI system, even when there are small amounts of data. Cohaku engine supports following functionalities.

Data Collection

Collect particular data efficiently and find such effective samples as o enhance models

Information Extraction

Extract information such as named-entities from both structured and unstructured data

Transfer Learning

Utilize related data and models to target domain and target tasks, and build high performance models

Service

Data Collection and Analytics

Collect specific data from the web, organize collected data into databases, and make analytics reports.

Consulting on AI Technologies

Consult on issues about a text, and provide technical supports on natural language processing, text mining, and machine learning.

Cohaku Functions

Provide some functions of our knowledge engine, Cohaku depending upon the situation.

Case Studies

Trend Analysis for Branding

Analyze self-strength and weaknesses, competitors and trends in the market from several resources such as official websites, news articles, and media quantitatively, the results of which are useful for self-branding and decision-making.

Fast Development of Order-made Classifier

Fast development of high-performance text classifiers based on transfer learning, whenever there are only small amounts of training data. Deal with a wide range of text such as not only general documents but also patents, tradename, and POS data

Text Summarization and Title Generation

Develop such technologies as to extract important sentences and generate title, headline, and summarization from documents like news articles, a part of which has been granted by Tokyo Metropolitan Small and Medium Enterprise Support Center in 2016-2017.

Automatically Building Database from Web Data

Collect specific data from web efficiently, extract and organize particular entities and information such as corporates, products, and services into a database, for the purpose of corporate analysis and sales supports.