What Is R&D Theme Exploration and Creation?

Discovering New Technological Possibilities and Research Hypotheses Beyond Established Domains

R&D Theme Exploration and Creation is the process of identifying new technological possibilities and developing research challenges and hypotheses based on technological capabilities, market needs, and societal challenges. It is a critical activity that determines the direction of research and development, including improvements to existing products, the development of new technologies, the creation of innovative products, and expansion into new business domains.

In corporate R&D, research themes are generally established according to technology strategies and business plans. However, when organizations explore potential themes exclusively within their existing business domains or areas of technical expertise, they risk overlooking opportunities beyond those boundaries.

In particular, principles and methods developed in one technological field may provide solutions to problems in entirely different disciplines. Discovering such possibilities requires looking beyond established technological terminology and classifications to examine functions, structures, operating principles, and relationships.

The central challenge of R&D Theme Exploration and Creation is not merely to select promising themes from known alternatives, but to discover possibilities that have not yet been recognized.

Where Do New R&D Themes Come From?

Approaches to R&D theme creation can be broadly classified into technology-driven, needs-driven, and societal or environmental change-driven exploration.

Technology-Driven Exploration begins with a company’s technological capabilities, research outcomes, proprietary technologies, or specialized expertise and seeks new applications and markets for them.

For example, materials, measurement technologies, or manufacturing processes originally developed for a particular industry may have potential applications in entirely different markets when reconsidered in terms of their fundamental functions and characteristics.

In this approach, the technological capability serves as the starting point, while new problems, applications, and customer needs are explored. As a result, the target market or customer group may change significantly even when the core technology remains substantially the same.

Needs-Driven Exploration begins with customer problems, product performance requirements, manufacturing constraints, or unresolved technical challenges.

Rather than limiting the investigation to improvements using existing technologies, the problem can be redefined in terms of functions and structures, allowing solutions from other technological fields to be considered.

In this approach, the problem or customer need serves as the starting point, while alternative technological solutions are explored. This may require capabilities beyond the company’s existing expertise, including the acquisition of external technologies or collaboration with other organizations.

Societal and Environmental Change-Driven Exploration examines future technological requirements arising from demographic changes, energy transitions, environmental regulations, resource constraints, and changes in industrial structures.

Methods such as horizon scanning and scenario analysis help identify possible future developments and translate them into directions for new research and development.

These approaches are closely interconnected. A technological capability may reveal previously unrecognized market needs, while a deeper understanding of customer problems may lead to the discovery of unexpected technological solutions.

Furthermore, neither needs nor technological solutions are necessarily fixed from the beginning. Discovering a new technology may change how a customer problem is understood, while a better understanding of the problem may alter the direction of technological exploration.

R&D theme creation is therefore an iterative process in which technological knowledge, problem understanding, and potential applications continuously influence one another.

Conventional Methods of R&D Theme Exploration and Their Limitations

Various methods have been developed to support R&D theme creation, including brainstorming, technology roadmapping, TRIZ, morphological analysis, technology scouting, patent and scientific literature analysis, and open innovation.

Brainstorming and other ideation methods help generate ideas from multiple perspectives. However, their effectiveness depends heavily on the knowledge and experience of participants. They may not adequately incorporate knowledge from technological fields unfamiliar to the group.

Systematic problem-solving methods such as TRIZ provide frameworks for considering technical contradictions and general principles of problem solving. These methods can help researchers move beyond conventional approaches, but assessing their practical applications still requires specific technological knowledge and further investigation.

Patent analysis, scientific literature analysis, and technology scouting are effective approaches for identifying technologies and knowledge developed outside an organization. However, when investigations begin with predetermined keywords or technological categories, the scope of exploration may be constrained by existing classifications.

Technology maps and technology landscapes can reveal the structure of technological domains, concentrations of research activity, and relatively unexplored areas. Nevertheless, visualizing the distribution of technological information is different from developing new technological solutions based on that information.

Each of these methods serves an important purpose. Rather than expecting a single method to address every challenge, organizations need to connect information acquisition, problem reformulation, cross-domain exploration, hypothesis generation, and technical verification into a coherent process.

Technological Abstraction and Cross-Domain Exploration

Connecting knowledge from different technological fields requires examining technologies beyond their specific names and applications. Their underlying functions, structures, operating principles, and relationships must be considered at a more abstract level.

For example, suppose the fundamental function of a technology is to detect the condition of an object and deliver necessary resources to the appropriate location.

Similar functions may exist in biological systems, logistics networks, manufacturing equipment, and information networks, even though the specific materials, devices, and applications differ considerably.

Different technological systems may share common structures or relationships in the way they solve problems.

A theoretical foundation for this type of cross-domain analogy can be found in Structure-Mapping Theory, developed by cognitive scientist Dedre Gentner.

Structure-Mapping Theory emphasizes the importance of correspondences between relational structures rather than merely similarities in the superficial characteristics of individual objects.

Applied to R&D theme exploration, this perspective suggests that technologies with different names, materials, or applications may still be relevant to one another when they share similar functional or relational structures.

However, structural similarity does not guarantee technological transferability. Material properties, operating environments, scale, energy requirements, costs, and other constraints may differ substantially.

Ideas generated through analogy must therefore be examined against the technical conditions of the target application.

The value of cross-domain exploration lies not simply in combining technologies from different disciplines, but in acquiring knowledge that allows researchers to reconsider how problems are formulated and which principles might be used to solve them.

Creating R&D Themes Through Concept Exploration

Since the 1990s, Mindware Research Institute has pursued an approach called Concept Exploration, originally developed under the Japanese term Gainen Chosa (Concept-Based Investigation).

Concept Exploration investigates knowledge across different disciplines by focusing on the characteristics, functions, structures, and relationships of technologies and problems.

The process begins by analyzing the target technology or problem and identifying its essential characteristics. These characteristics are then abstracted to formulate exploration conditions that are not limited to conventional technological terminology or industrial classifications.

By using these abstracted conditions, researchers can search for relevant principles, structures, and problem-solving approaches in fields outside the original technological domain.

Consider, for example, the development of a material capable of repeatedly repairing damage.

A conventional investigation might focus primarily on existing self-healing materials and repair agents. A concept-oriented investigation, however, could examine the more general problem of continuously delivering repair substances to damaged locations.

This reformulation suggests possible connections with biological circulatory systems, fluid distribution networks, and distributed material transport systems.

The functions and operating conditions of these systems could then be investigated to develop hypotheses concerning material-embedded supply networks, damage detection and controlled delivery, or mechanisms for maintaining repeated repair capabilities.

These examples illustrate possible directions for exploration rather than demonstrating the technical feasibility of particular solutions.

Nevertheless, they show how abstracting a problem can reveal alternatives that might be difficult to identify when an investigation remains within its original technological field.

From Information Exploration to R&D Hypothesis Generation

Discovering an interesting technological possibility is not the same as establishing a research theme that can be investigated systematically.

Identifying a technology or principle from another domain is only an initial step. Researchers must determine which problem it could solve, which operating principle would be applied, under what conditions the solution might work, and what evidence would be required to assess its feasibility.

For this reason, promising ideas must be developed into testable R&D hypotheses.

For example, the general idea of applying biological circulatory systems to self-healing materials does not yet define a specific research objective.

A more concrete hypothesis might be formulated as follows:

An internal supply network that repeatedly delivers repair agents to damaged locations may maintain repair performance under repeated damage more effectively than a system relying on a finite local reserve of repair agents.

This formulation provides a clearer basis for identifying technical challenges, establishing experimental conditions, and comparing alternative approaches.

An R&D hypothesis should therefore go beyond a simple combination of technologies or a product concept. It should clarify the proposed mechanism, the conditions under which it is expected to operate, its anticipated effects, and the observations that could challenge or falsify the hypothesis.

At the same time, an R&D theme involves more than a technical hypothesis. Potential applications, strategic relevance, required resources, and expected outcomes must also be considered.

Once a technical hypothesis has been formulated, it needs to be examined for practical relevance and developed into a concrete research plan.

R&D Theme Exploration Using LLMs and Machine Learning

Advances in large language models and machine learning have created opportunities to support cross-domain exploration across a broader range of technological knowledge than was previously practical.

LLMs can analyze technical literature, extract technological characteristics and functions, and help reformulate problems at different levels of abstraction.

They can also use abstracted characteristics to explore related knowledge in other disciplines and generate candidate technological hypotheses.

Machine learning techniques, including semantic similarity analysis and clustering, can help organize technological documents and generated hypotheses, revealing the structure of technological domains and relationships among alternative research ideas.

By combining these technologies, R&D theme exploration can be organized into a systematic process.

Stage	Main Activities
1. Knowledge Base Construction	Collect and organize technical literature and structure the technological knowledge used for exploration.
2. Technological Element Extraction	Extract functions, structures, operating principles, performance characteristics, constraints, and other technological elements.
3. Abstraction and Exploration Query Generation	Generalize technological characteristics and formulate questions for cross-domain exploration.
4. Cross-Domain Exploration	Search other disciplines for technologies, principles, and problem-solving methods with similar functions or structures.
5. R&D Hypothesis Generation	Develop candidate technological solutions and research hypotheses based on the knowledge discovered.
6. Hypothesis Structuring and Examination	Organize similar or overlapping hypotheses and examine their supporting evidence, operating conditions, and technical challenges.

This process can generate diverse candidate hypotheses from technological knowledge and organize their relationships for systematic examination.

However, hypotheses generated by LLMs do not necessarily possess technological novelty or validity.

The results may include reformulations of existing technologies, unsupported assumptions, or combinations that are technically infeasible.

Furthermore, reasoning generated from an LLM’s learned knowledge must be distinguished from technological facts supported by traceable sources.

Before adopting a hypothesis as an R&D theme, researchers must investigate prior art, assess technical feasibility, and conduct appropriate experiments or simulations.

The role of AI is not to guarantee that an invention is valid or technically feasible. Rather, it is to expand the alternatives available for consideration and support researchers in knowledge exploration and hypothesis formation.

Hypothesis Integration, Evaluation, and Development into R&D Themes

When cross-domain exploration generates a large number of hypotheses, evaluating each one independently may not be sufficient.

Several hypotheses may share a common technological principle, or they may represent complementary elements that could be integrated into a broader R&D theme.

Analyzing similarities and relationships among hypotheses can help identify common technical challenges and solution principles. Through this process, individual ideas may be integrated into more systematic research themes.

Conversely, different hypotheses may propose alternative solutions to the same technical problem. Comparing these alternatives can clarify the conditions that must be verified and the technological options available for further investigation.

Evaluation of new R&D themes should consider their differences from existing technologies, technical feasibility, potential application value, and resource requirements.

However, hypotheses emerging from unfamiliar technological domains may not have sufficient evidence for reliable quantitative evaluation during their early stages.

In such circumstances, it is important not to prioritize hypotheses solely on the basis of projected business value.

An alternative approach is to consider which experiments or additional investigations would most effectively reduce uncertainty.

By transforming exploratory hypotheses into testable research questions, experimental plans, and R&D projects, organizations can connect theme creation with practical R&D management decisions.

Mindware Research Institute’s Approach

ConceptMiner R&D, a Cross-Domain R&D Concept Exploration System developed by Mindware Research Institute, builds upon the Concept Exploration approach by combining LLMs and machine learning to support the discovery and creation of new research themes.

ConceptMiner R&D constructs knowledge bases from technical literature and uses them to support technological element extraction, conceptual abstraction, exploration prompt generation, cross-domain investigation, and R&D hypothesis generation.

The system also employs concept mapping based on machine learning techniques such as Growing Neural Gas (GNG) to visualize the structure of technological knowledge and exploration results, supporting researchers in understanding relationships and identifying additional directions for investigation.

Beyond generating individual hypotheses, Mindware Research Institute aims to extend this approach toward hypothesis integration, evaluation, and research planning by analyzing relationships among hypotheses and their connections to the technological elements from which they were generated.

We regard R&D theme exploration not merely as a process of generating ideas, but as a knowledge creation process that extracts knowledge from technological information, connects concepts across disciplines, formulates new hypotheses, and examines their validity.

Deep expertise in established technological fields and exploration of unfamiliar possibilities are not opposing approaches.

Specialized knowledge provides the foundation for understanding technical problems. Abstraction allows those problems to be reconsidered from different perspectives. Cross-domain knowledge exploration introduces alternative principles and methods. Expert evaluation and experimentation determine whether the resulting hypotheses are technically meaningful.

Together, these activities provide a systematic foundation for creating new research and development themes.

The essential capability required for R&D Theme Exploration and Creation is not simply the ability to investigate known alternatives efficiently. It is the ability to discover technological possibilities that have not previously been recognized and transform them into testable R&D hypotheses.

ConceptMiner R&D aims to provide a framework through which organizations can develop and apply this capability systematically and continuously.

References and Further Reading

Gentner, D. (1983). Structure-Mapping: A Theoretical Framework for Analogy. Cognitive Science, 7(2), 155–170.
Kornish, L. J., & Hutchison-Krupat, J. (2017). Research on Idea Generation and Selection: Implications for Management of Technology. Production and Operations Management, 26(4), 633–651.
Deliberate Exploratory Search in Technology Innovation: Discovering and Developing Need–Solution Pairs. Research Policy.
OECD (2015). Frascati Manual 2015: Guidelines for Collecting and Reporting Data on Research and Experimental Development.