This document is a Japanese translation of a passage entitled "Artificial intelligence for curation of information and knowledge acquisition " from Chapter 3, "Examples of corporate use of materials informatics," in the book "Next-generation materials development using materials informatics, quantum computers , natural language processing and autonomous experimental systems" published by AndTech and written by Dr. Christopher L. Farrow and Dr. Alexandre Chabot-Leclerc of Ensort.

overview

As competition for new material development intensifies, knowledge acquisition to promote research and development is becoming increasingly important. This document explains how artificial intelligence can contribute to knowledge acquisition.
The process of turning information into knowledge can be divided into two stages: "curation" and "knowledge acquisition." AI can support both stages by fusing and interpreting information and helping researchers access it. Some of the
scientific search systems introduced in this document have been developed by our company in collaboration with clients and are actually in operation.

Table of contents

Introduction

1.  Curation using technology
   • 1.1 Curation as search
     • 1.1.1 NLP-powered search
     • 1.1.2 Image Search
     • 1.1.3 Table Search and Specialty Search
     • 1.1.4 Extracting Data from a Graph
   • 1.2 Search Limitations
2. 2. Generative AI for curation
   • 2.1 All information can be converted into text
   • 2.2 Text can be treated as data
   • 2.3 Beyond Text Search: Multimodal Embeddings
3. 3. Knowledge Acquisition by Generative AI
   • 3.1 Answering Questions with Retrieval-Augmented Generation
   • 3.2 Agents Performing Tasks
   • 3.3 Association by Embedding

Conclusion

References