Joint Research Project
R&D Theme Exploration & Hypothesis Generation
using AI and Conceptual Structure Models
Discovering the next related themes from technological continuity and customer commonality
Request Proposal PDFThe True Dilemma in the R&D Field
“Information gathers. AI can summarize it. Yet, it doesn’t lead to specific R&D themes or hypotheses.”
In exploring R&D themes, it is necessary to consider where the company is currently positioned in terms of technology, product, and customer domains, what adjacent domains surround it, and what intermediate technologies, applications, and customer touchpoints are required to advance to the target domain.
On the other hand, even if strategies like “expanding existing technologies to new markets” are discussed company-wide, they are merely rough analyses like cross-charts, falling short of the detailed connections the field wants to know: “Which technology, via which intermediate application, should be deployed to which customer?”
3 Features of This Project
1. R&D Application of the Established Core Technology “ConceptMiner”
“ConceptMiner” and conceptual structure models (machine learning algorithms like GNG) developed by Mindware Research Institute are already established, powerful analysis platforms. The purpose of this project is to conduct joint research to apply and specialize this established technology for “theme exploration and hypothesis generation” in R&D departments.
We model vast amounts of external text information such as papers, patents, and news, alongside the company’s technological seeds and customer needs. Rather than simple keyword searches, detailed technical information is mapped as fine-grained areas on a conceptual structure, making the current position, adjacent areas, boundaries, and bridging areas computable.
2. Not Separating Strategy from Technical Details
While the R&D theme itself is a matter of micro-level technical details, it must not be separated from business strategy. In this project, instead of making rough strategic decisions, strategic theories such as “technological continuity” and “customer commonality” are applied to explore the advancement paths (routes) connecting micro-level technical information.
Furthermore, beyond investigating already known themes, we consider attempts to discover new themes beyond existing search boundaries through “conceptual research.” This involves setting abstract concepts (e.g., “our company’s winning pattern” or “it would be great if this existed” as a structure, not a specific idea) and matching them with unknown themes.
3. Formation of “Testable Hypotheses” Accompanied by Evidence
We do not stop at mere idea generation via AI brainstorming.
Based on detailed paths between technologies, applications, and needs derived from the conceptual structure model, the LLM texts a “testable proposition (theme hypothesis)” equipped with the target, establishing conditions, expected effects, unconfirmed items, and verification methods. Because it is always linked to the original evidence text (papers, patents, etc.), experts can accurately evaluate its scientific and technical validity.
Target Departments
For companies with challenges in the following areas:
- Research & Development
- New Product / Business Development
- Technology Strategy
- Research Planning
- Intellectual Property
- Open Innovation
Request the Joint Research Project Proposal
We will send you a proposal (PDF) detailing the research and verification steps, integration with existing technologies like ConceptMiner, and deliverables of the project. Please request it using the form below.