We are a company that provides technology to free people around the world from all kinds of risks.
We detect early signs of machine failures and diseases that cannot captured by the human eyes,
and support measures and preventive actions to avoid serious accidents and diseases before they happen.
With this technology, we contribute to the eradication of serious machine failures and diseases.
Data generated by machines, such as log data and IoT data, can sometimes be terabytes in size, and genomic data can be over 100,000 dimensions. Our company has a lot of experience in analyzing such multidimensional big data and possesses various data transformation and data modeling technologies.
Training-type AI is extremely effective in fields where high-quality training data is available. However, the fields where high-quality training data can be obtained are limited. In such fields, quickly grasping the degree of modulation of the target is effective in preventing major risks such as failures, diseases, fraud, and even catastrophic breakdowns in advance. D’Insight is a ‘risk visualization AI’ that quantifies risks while allowing humans to recognize risks at a glance.
There are various definitions of risk, but we define it as ‘the degree of deviation from the normal state equals the size of the risk’. We believe that some signs of abnormality are hidden in states that deviate from the normal state. The larger the deviation, the more we think the abnormal elements are expanding. Therefore, to detect abnormalities early and prevent major problems, it is important to detect risks while they are still small.
Our approach is very straightforward. First, we define what a normal situation is. Usually, it’s no problem to consider the initial situation as the normal state. Next, we visualize and index how much the target deviates from the normal situation. The data used at this time varies, such as current, vibration, sound, temperature, and concentration for machines. For humans, it could be heart rate, respiration, movement with walking, various test results, etc. In finance, various economic indicators and each transaction record data can be used.
With the remarkable development of AI technology in recent years, many people believe that AI can solve
any kind of problems. But this is a great misunderstanding. Indeed, in fields where high-quality training data is abundantly available, training-type AI is an extremely powerful tool. On the other hand, “D’Insight”, a visualization-type AI, can quickly capture slight changes in the target without training, so high-quality training data is not necessary. In reality, the fields where high-quality training data can be obtained are limited in the world. In such an environment, “D’Insight”, which can quickly grasp the degree of modulation of the target, is extremely effective in preventing major risks such as failures, diseases, fraud, and even catastrophic breakdowns in advance.
Training-type AI attempts to understand the distribution of multidimensional data through training; data for training is initially required. On the other hand, visualization-type AI reduces the dimensions of multidimensional data distribution using mathematical methods and displays it in the form of a two-dimensional (or three-dimensional) map that humans can understand; it does not require training data. The former is suitable for categorization, such as distinguishing between dogs and cats. On the other hand, the latter is suitable to illustrate degree of changes, such as the degree of degradation. We believe you will understand that visualization-type AI is suitable for the visualization of risk that we analyze.
The change in the state of a bearing from its initial stable state, gradually deteriorating until eventually failing, can be understood by observing the shift towards the bottom left of the plot.
The way in which the coronavirus has mutated and spread from Wuhan to all over the world is visualized. It is understood that in early 2020, the virus was directly transmitted from China to Japan, and in summer, another strain was transmitted via Europe.
Machine Failure Prediction
Quality Control of Products
Road Deterioration Diagnosis
Driving Risk Analysis
Fraudrant Access Analysis of Computer Centers
Air Conditioner Failure Data Analysis
Patent Data (Providing Search Services),
Job-Matching Services
Simplified Sleep Apnea Test Using a Smartphone
Air Pressure Analysis Based on Percussive Sounds
Heart Rate and Respiration Analysis,
Stress Analysis
Mutation and Spread Route Analysis of a New Virus
Extraction of Infectious Genes from Camels
Infectivity Analysis of a Virus Through Protein (Antibody) Data Analysis
©2023 d-isum. All Rights Reserved