Profile & History
Kunihiro Tada
An independent path of research,
translated into original AI systems.
My work on concept formation and AI knowledge systems has grown out of planning advanced-technology seminars, editing technical reference materials, researching and writing about emerging technologies, localizing and distributing international software, more than 25 years of practical work in self-organizing systems and data mining, and the development of original software systems.
Profile
A researcher and systems designer who does not fit neatly into conventional categories
My field is not simply AI, data science, or software development. I study how humans and machines form concepts, organize knowledge, and use experience in judgment—both as an epistemological question and as a problem of system design.
As a student, I was strongly influenced by Hoshino’s writings on the theory of technology. They taught me to view technology not merely as machinery or technique, but in relation to people, society, and industry. This became the intellectual starting point for my later work in advanced-technology research, concept formation, and AI systems.
The foundation of my professional skills was formed at a publishing and seminar company, where I worked from 1982. By planning and operating advanced-technology seminars and editing technical reference collections, I learned how to investigate unfamiliar fields quickly, identify the important questions and the right specialists, and organize complex material into a coherent form that others could understand. The basic pattern of my present commissioned R&D work—research, structure, and implementation—originated in this period.
My subsequent research path did not begin in a university laboratory either. It developed through new-business development, technical writing on multimedia and computer graphics, business development during the early commercialization of the Internet, a cross-disciplinary series of studies on concepts spanning philosophy, sociology, and psychology, the localization and distribution of international software, and the introduction of analytical technologies into business practice.
Since the early 2000s, I have worked continuously with self-organizing maps, data mining, statistical analysis, and Bayesian networks in practical applications. Through collaboration and exchange with university researchers, I also co-authored several peer-reviewed papers on self-organizing maps. Today, AI-assisted coding allows me to implement ideas developed over many years directly in ConceptMiner and ThinkNavi.
My research approach is not merely to criticize established methods, but to design alternatives based on different principles and demonstrate them as working systems.
Research Focus
Concept Formation and AI Knowledge Systems
I approach AI not only as a computational technology, but as an epistemological problem: how a system distinguishes the world, what it treats as a concept, and how it organizes knowledge and experience.
Concepts are formed, not simply discovered
I do not treat classifications and meanings as fixed entities. I treat them as provisional cognitive models formed through attention, purpose, context, and the structure of the observer.
Retrieval is not the same as understanding
Knowledge formation cannot be explained by vector similarity or document retrieval alone. I examine conceptual structure, memory, routing, and judgment as parts of a single system.
Turn ideas into working systems
I combine LLMs, self-organizing machine learning, statistical analysis, probabilistic models, and databases to test research hypotheses in systems that actually run.
Move between practical implementation and research
In addition to implementing self-organizing maps in practice, I co-authored several peer-reviewed papers through collaboration with university researchers. The experience of examining implementation results and expressing them in a reproducible form continues to inform my current R&D.
Research Path
From researching emerging technologies to building original systems
My professional starting point was the work of investigating advanced technologies quickly, structuring the key questions, and communicating them through seminars and technical publications. The same method continues today in the design and prototyping of AI systems.
Influence of Hoshino’s theory of technology
I was an ardent fan of the technology critic Yoshiro Hoshino. He tought me the theory of technology that viewed technology not as isolated machinery or technique, but in relation to people, society, and industry. This perspective led me to examine not only what is new in a technology, but also its assumptions and meaning. *Perception and Patterns* by Satosi Watanabe (published in the Iwanami Shinsho series)—a book I picked up casually at the time—also became an important volume for me in later years.
Began my career as a technical writer
I began my professional career by interpreting technical information and turning it into writing and structure that users could understand. More important than the details of the employer was the fact that my work from the outset involved understanding technology and explaining it clearly.
Planning advanced-technology seminars and editing technical reference collections
At a publishing and seminar company, I planned and operated seminars on advanced technologies, selected speakers, and edited technical reference collections. This was where I developed the ability to investigate unfamiliar fields, identify the essential questions and experts, and edit complex knowledge into a coherent structure.
From technology research to new-business development
I extended the research, editing, and seminar-planning skills I had developed into new-business development. I investigated new media, information technology, and emerging industries, and worked not only to explain technologies but also to formulate their industrial potential as business concepts.
Planned an advanced-technology seminar on AI chips
During the second AI boom, I planned a seminar on dedicated AI hardware and its industrial potential. It focused on the relationship between AI and specialized computing infrastructure roughly four decades before today’s competition in AI semiconductors.
Moved into independent technology research and writing
After leaving the organization, I continued researching advanced technologies as an external contributor and wrote research reports, articles for technical publications, and explanatory materials.
Research and writing on multimedia and computer graphics
I researched and wrote about CD-ROM, CD-I, interactive media, computer graphics, and the new expressive technologies and information industries then grouped under the term “multimedia.”
Entered Internet business development at the start of commercialization
I began developing Internet-based businesses when commercial use of the Internet was first emerging in Japan. I launched www.mindware-jp.com at an early stage, and it appeared alongside major corporate sites in “Japan’s Home Pages,” a directory that listed roughly one hundred Japanese websites at the time. Because the company had not yet been incorporated, I could not obtain a .co.jp domain and added “-jp” to a .com address—a naming decision I later came to regret.
“Concept Research” — the beginning of an epistemological inquiry
I spent an extended period working in libraries, reading widely in philosophy, sociology, psychology, and cognitive science, and wrote a serialized study titled “Concept Research” that examined human classification, meaning, and concept formation across disciplines.
The questions developed during this period later led to my work on self-organizing maps, ConceptMiner, ThinkNavi, and my current research on AI knowledge systems.
Research and commercialization of self-organizing maps
I began working seriously with Kohonen self-organizing maps and launched the localization, introduction, and technical support of Viscovery SOMine in Japan. This was the point at which questions of classification and cognition from my earlier concept research moved into the practical application of machine learning.
In the early 2000s, I also engaged in exchange and joint research with university researchers and co-authored several peer-reviewed papers on self-organizing maps. This gave me experience not only in practical product implementation, but also in organizing research results as academic publications.
Expanded into statistical analysis and Bayesian networks
Through HUGIN, XLSTAT, and related products, I developed a business introducing probabilistic reasoning, statistical analysis, and data mining into corporate practice.
Development of ConceptMiner and ThinkNavi
Using AI-assisted coding, I began designing and developing original systems that integrate years of work on concepts, self-organizing technologies, and LLMs.
Commissioned R&D as an independent research studio
With ConceptMiner and ThinkNavi as representative works, I am expanding into commissioned R&D for technical problems that cannot be solved by LLMs alone.
Company History
History of Mindware Research Institute
The company began by introducing and distributing international technologies and is now evolving into an independent studio that conceives and develops original AI systems.
Mindware partnership established
Founded to provide advanced-technology research, technical information, and software-related services.
Viscovery SOMine business launched
Began localizing, distributing, and supporting data-mining software based on self-organizing maps.
Began distributing HUGIN and XLSTAT
Expanded the business into Bayesian networks and statistical analysis.
Mindware incorporated
Established the corporate foundation that later became Mindware Research Institute, Inc.
Restructured the statistical software business
Concluded the long-running XLSTAT business and accelerated the transition toward original R&D.
Development of ConceptMiner and ThinkNavi begins
Shifted toward developing original systems centered on concept formation, knowledge structures, and AI memory.
Redefined as an independent R&D studio
Refocused the business on concept formation, AI knowledge systems, and a limited number of commissioned R&D projects.
Selected Works
Two systems that represent the current research
ConceptMiner
A system that combines embeddings, self-organizing machine learning, network structures, and LLM-based interpretation to form and explore conceptual structures in text and data.
Explore ConceptMiner →ThinkNavi
An environment integrating research, knowledge bases, conceptual structures, memory, and dialogue to move AI beyond information retrieval toward knowledge production and decision support.
Explore ThinkNavi →Turn unresolved technical questions into working systems.
We undertake commissioned R&D combining LLMs, machine learning, statistical analysis, and databases.