AI KnowledgeIntermediateUpcoming
SECURE CHAIN: Unstructured-to-KG Pipeline for Secure Chain
The Challenge
This challenge focuses on gathering blog texts and tweets, parsing their unstructured content, and generating new knowledge that can be added to the Secure Chain KG. Participants will use NLP techniques and LLMs to extract meaningful relationships and convert them into triplets. They will also compare outputs from different LLMs. Web scraping through APIs or crawlers is required to collect the data. This is a high-difficulty challenge involving advanced NLP tasks.
About This Challenge
SUMMARY: Participants will collect unstructured textual data—such as blog posts and tweets—and transform it into structured knowledge suitable for the Secure Chain Knowledge Graph (KG). Using NLP techniques and LLMs, participants will detect meaningful entities and relationships, then represent them as triplets to add to the KG. This challenge is high difficulty due to the focus on unstructured text transformation and NLP. A key goal is to compare the performance of different LLMs in extracting accurate and relevant triplets to enhance the Secure Chain KG. CHALLENGE OBJECTIVES: * Collect unstructured textual data using APIs or web crawlers (e.g., blogs, social media). * Parse the text using NLP and LLMs to detect entities and relationships. * Transform extracted information into triplets suitable for the Secure Chain KG. * Compare LLMs’ performance in terms of accuracy, relevance, and completeness. * Add high-quality, automatically extracted knowledge to the Secure Chain KG. SKILLS NEEDED: NLP and text processing LLM usage and evaluation Data extraction via APIs or web crawling Knowledge graph creation and RDF/triplet modeling Python programming EXPECTED OUTCOME: A pipeline that transforms unstructured textual data into structured triplets, evaluates multiple LLMs for extraction quality, and enriches the Secure Chain KG with high-quality knowledge.
