We approach product development using a variety of design thinking frameworks, including interviews.
The information obtained from interviews, surveys, and prototyping is visualized and analyzed using HCD (Human-Centered Design) methods. We seek the best approach to the issues from the perspective of the actual user of the product.
This prevents a mismatch between the actual problem and the technology applied, and enables us to create products that solve the problem more effectively.
We can let the machine learning model learn unlabeled data and determine if it deviates significantly from the data it has learned. It can be used to detect abnormal patterns.
It is also possible to build prediction models for time series data without labeling and annotation work. This system is expected to be used in the fields of disaster prevention and medical care.
Sparse modeling, or sparse coding, and more efficient model architectures allow you to build higher performance models with less data than before.
Data saving also reduces the training time of the model, and the computational resources and costs required can be reduced.
As demonstrated in our case study of osteoporosis determination, a higher prediction accuracy can be achieved by combining data in a complex manner.
This combination of multiple data (e.g., combination of electronic medical record information and CT) is called multimodal, and it is one of the essential techniques to make deep learning models learn to make more sophisticated decisions based on the situation.
We have a wealth of experience in handling large data sets such as X-ray images, CT and MRI in the medical field and have the know-how to process large data sets at high speed.
Our engineers are familiar with the key aspects of algorithm acceleration and can design for computational reduction without simply relying on hardware or external software libraries.
We seek to create an organisation and culture that allows us to work efficiently and quickly.
For example, there is no such thing as a mere review of progress or regular meetings, and we accept people from all over the country and abroad to work with us. Instead of being expected to be professional, members are given free rein to carry out their work.
Since our foundation, we have been able to maintain an organization that can deliver value in product development projects by not having everyone in one place, and by constantly updating to allow for quick decision making.
We use a collaborative team approach to deliver projects, with a highly technical lead engineer and a highly skilled project manager (PM) sharing the workload.
With a lead engineer looking after the technical details and a PM looking after the project as a whole, we are able to combine breadth with depth of detail. With a lead engineer looking after the technical details and a PM looking after the project as a whole, we are able to combine breadth with depth of detail.
We have successfully completed a number of projects with this organisation.