Surpassing the Limits of LLMs! Free Webinar 'The future of technology can be 'estimated', not 'predicted' - Next R&D themes and business opportunities derived from papers, patents, and grants, unseen by LLMs' Held
Astamuse will hold a free webinar on April 23, 2026, for R&D and IP professionals. It introduces a unique method combining patent/paper data, TF-IDF analysis, and LLMs to overcome AI hallucinations and estimate future technology trends.
📋 Article Processing Timeline
- 📰 Published: April 9, 2026 at 20:49
- 🔍 Collected: April 9, 2026 at 12:30
- 🤖 AI Analyzed: April 20, 2026 at 08:57 (260h 26m after Collected)
Astamuse Co., Ltd. (Headquarters: Chiyoda-ku, Tokyo; President and CEO: Ayumu Nagai) will hold a free webinar titled "The future of technology can be 'estimated', not 'predicted' - Next R&D themes and business opportunities derived from papers, patents, and grants, unseen by LLMs" starting at 12:00 on Thursday, April 23, 2026. This content reconstructs the presentation "Statistical analysis of physics literature databases and future estimation of research trends using TF-IDF" given at the 73rd JSAP Spring Meeting, tailored for corporate R&D, intellectual property, and new business development professionals.
Register for participation (free) here
https://www.astamuse.co.jp/event/37249/
* Even if you are unavailable on the day, if you register, you can watch the archive video at a later date.
Additionally, those who complete the questionnaire can download the slide materials.
Overview of this webinar
With the spread of generative AI (LLMs) such as ChatGPT, Copilot, and Claude, literature research and grasping technology trends have become dramatically more efficient. However, on the other hand, new challenges are emerging on the frontlines of corporate decision-making. Anxiety about "plausible but incorrect information (hallucinations)," and the structural limitation of LLMs where information accuracy is lower in emerging and cutting-edge fields—as a result, a situation has arisen where it is difficult to directly utilize LLM outputs for selecting R&D themes and judging new businesses.
In this webinar, we will introduce a new approach: "The future of technology is not something to be predicted, but can be estimated from data." Using an analysis method developed originally by Astamuse, by integrating three types of data—papers, patents, and grants—we structurally understand technology on a timeline of "present (patents)," "near future = 3-5 years later (grants)," and "far future = 5-10 years later (papers)." We will explain using real data a method to objectively derive upcoming technology areas and business opportunities by combining quantitative analysis based on TF-IDF with LLMs.
The case study covered will be the nanomaterials field (over 140,000 papers). We will specifically show how trends are extracted and the future is estimated. This content is reconstructed based on the presentation "Statistical analysis of physics literature databases and future estimation of research trends using TF-IDF" at the 73rd JSAP Spring Meeting, with an eye toward application in corporate decision-making.
Recommended for the following people:
- Those in research and development/research planning departments who cannot see the next R&D theme to tackle
- Those in charge of technology strategy/R&D theme selection who cannot explain or persuade regarding technology trends using data
- Those in charge of corporate planning/management of technology (MOT) who use LLMs but are anxious about accuracy and reliability
- Those in new business development/business planning departments who want to bring objective rationale to the direction of new businesses
- Those in IP/patent departments who want to utilize paper and patent information more strategically and at a higher level
Event Overview
[Title] The future of technology can be 'estimated', not 'predicted' - Next R&D themes and business opportunities derived from papers, patents, and grants, unseen by LLMs
Register for participation (free) here
https://www.astamuse.co.jp/event/37249/
* Even if you are unavailable on the day, if you register, you can watch the archive video at a later date.
Additionally, those who complete the questionnaire can download the slide materials.
Overview of this webinar
With the spread of generative AI (LLMs) such as ChatGPT, Copilot, and Claude, literature research and grasping technology trends have become dramatically more efficient. However, on the other hand, new challenges are emerging on the frontlines of corporate decision-making. Anxiety about "plausible but incorrect information (hallucinations)," and the structural limitation of LLMs where information accuracy is lower in emerging and cutting-edge fields—as a result, a situation has arisen where it is difficult to directly utilize LLM outputs for selecting R&D themes and judging new businesses.
In this webinar, we will introduce a new approach: "The future of technology is not something to be predicted, but can be estimated from data." Using an analysis method developed originally by Astamuse, by integrating three types of data—papers, patents, and grants—we structurally understand technology on a timeline of "present (patents)," "near future = 3-5 years later (grants)," and "far future = 5-10 years later (papers)." We will explain using real data a method to objectively derive upcoming technology areas and business opportunities by combining quantitative analysis based on TF-IDF with LLMs.
The case study covered will be the nanomaterials field (over 140,000 papers). We will specifically show how trends are extracted and the future is estimated. This content is reconstructed based on the presentation "Statistical analysis of physics literature databases and future estimation of research trends using TF-IDF" at the 73rd JSAP Spring Meeting, with an eye toward application in corporate decision-making.
Recommended for the following people:
- Those in research and development/research planning departments who cannot see the next R&D theme to tackle
- Those in charge of technology strategy/R&D theme selection who cannot explain or persuade regarding technology trends using data
- Those in charge of corporate planning/management of technology (MOT) who use LLMs but are anxious about accuracy and reliability
- Those in new business development/business planning departments who want to bring objective rationale to the direction of new businesses
- Those in IP/patent departments who want to utilize paper and patent information more strategically and at a higher level
Event Overview
[Title] The future of technology can be 'estimated', not 'predicted' - Next R&D themes and business opportunities derived from papers, patents, and grants, unseen by LLMs