Data analysis efforts to connect buried value to results
Our mission at Generic Solutions is to give computers intelligence through unique solutions. Specifically, we are developing a service that analyzes the customer's database and provides valuable data. It was founded in 2006 when I was a student at Keio University SFC (Shonan Fujisawa Campus). It has been called "data analysis" since that time, but it is now the base technology for what is called "big data analysis" or "AI", and to put it simply, it means a huge collection of data to humans. It refers to retrieving certain information.
During college, he was involved in research such as information retrieval and data mining (a method of digging up correlations and values that are difficult to find by human power from data) under Professor Yasushi Kiyoki of the Faculty of Environment and Information Studies, and imitated human memory and cognitive functions. While experiencing various applied research with semantic associative search (mathematical model of meaning) as the core technology, I decided to start a business with the conviction that "this technology has future potential and opportunities".
Over the next 10 years, the terms data analysis and big data analysis have become quite commonplace, but it is difficult to add value simply by analyzing and visualizing data, and the case ends with "just analyzing the data and gaining awareness." Is often seen. Among them, our aim is to dig up hidden value from the data that companies have in advance and contribute to the creation of a mechanism that leads to new profits by more practical data expression.
That's where the proprietary software package "GS8" comes into play. While many other companies that claim AI start using data by collecting and organizing it, or starting learning after introduction, GS8 can produce results immediately from existing data. The feature is that you can experience the results of the introduction and let the machine learn the results so that you can connect to the next proposal.
As for the field of application, the research theme in college was to improve the learning effect by data analysis, so it started with the introduction to major cram schools and triggered by a contract with a food delivery net supermarket centering on organic vegetables etc. We have embarked on a retail business initiative. Specifically, we analyze information about "who bought what when" and use it to predict "who will buy what when". It is possible to grasp the changes in the needs of each customer in chronological order and maximize the sales of each customer at the best timing.
Develop research at Keio University and contribute to practical problem solving
It's only been a few years since I've been involved in initiatives in various industries that the tailwind has finally blown. I feel that the opportunities for major companies to become interested have increased dramatically since they started using the terms "big data" and "AI" all at once. To give an easy-to-understand example, the "Personal Flyer", which is currently used by the "Protecting the Earth", creates different leaflets for each person. Based on data analysis, the table lists products that are not in the customer's purchase history but may be of interest to them, and the back side is structured with a focus on reminders so that there are no omissions. By automatically utilizing the data analysis results, it is possible to save a lot of labor and waste by manually relying on intuition and experience.
On the other hand, in response to the growing attention to "FinTech", the number of transactions is increasing in the financial field. For example, in the case of a bank, we analyze the movement of deposit accounts for individuals and make optimal recommendations for each customer, such as investment trusts for this person and loans for individuals. For corporations, we have achieved results such as effective support by utilizing the same know-how for lending business feasibility.
One of the hardware features of these software is that the actual machine is installed in the customer's office instead of the cloud-based infrastructure that is currently the mainstream. In addition to building an on-premises (installed) infrastructure with a unique hardware configuration, we also provide our own cloud services. This machine is also an original one that we assembled from parts, but the unique infrastructure built by the optimal cooperation of hardware and software makes it possible to provide data processing services that can flexibly respond to the purpose. Is one of our great strengths.
In addition, many ventures that have "AI" solutions on their signboards try to forcibly apply one mathematical model such as deep learning (deep learning) or kernel method, or solve general problems with technologies that are biased toward specific areas. There are some cases of trying to apply to. The biggest question is how to fit such technology to real-world challenges. I myself learned from Professor Kiyoki that I express problems and solutions as distances on the same information space, select a mathematical model or information space that suits the purpose, and create a mathematical model that automates the algorithm. From that experience, we have established a method to build a unique algorithm system by flexibly incorporating deep learning, neural networks, pattern recognition, etc. according to the problem while utilizing various mathematical models. I think this attitude has led to results that deliver valuable results to our customers.
We will continue to practice and research for the optimal society without waste.
In recognition of the significance and success of these efforts, we will receive KII support from August 2017. In addition to marketing advice, we will set up a research team with our teacher, Professor Kiyoki, from October to take on the challenge of new stages with the aim of developing elemental technologies.
As a personal motivation, do you want more people to use the new technology that you have come up with? Since the advent of computers, the principle itself has not changed significantly, but we are aiming to create a "thinking computer" as a tool to create the future, not just a computer that plays the role of data processing. ..
On the other hand, what is needed is to promote the customer's understanding of utilizing information. While the technology that makes computers smarter by accumulating data and machine learning is advancing, we must also deepen our customers' understanding of how to actually apply the technology. In addition, a specialized organization that not only trains data scientists in-house and mechanizes the intuition and experience of the person in charge, but also enhances the predictability of various events and creates a "thinking computer" that can flexibly respond to changes. I think that is required.
In the future, we would like to expand the "thinking function" of computers by enhancing the interface for acquiring stream data such as sensors and communication such as image analysis and natural language processing. Furthermore, we would like to speedily create new innovations by combining the big data analysis and artificial intelligence technologies that we have practiced so far with Professor Kiyoki's "semantic associative search technology" that can handle sensibilities and emotions.
I would like to continue to make efforts under the slogan "Heaven helps those who help themselves" together with the thoughts of the seven falls and eight rises in the letter "8" of GS8 and the engineers who support it.
[Link to official website]