0.25
0.5
0.75
1.25
1.5
1.75
2
LLMs4SchemaDiscovery: A Human-in-the-Loop Workflow for Scientific Schema Mining with Large Language Models
Related categories
Presentation
LLMs4SchemaDiscovery: A Human-in-the-Loop Workflow for Scientific Schema Mining with Large Language Models00:00
Agenda00:16
1. Motivation00:47
Motivation00:51
What is Schema?02:05
Schemas: Toward AI-Ready Databases of Science03:20
We Proposed Schema-Miner04:59
2. Methodology: LLMs4SchemaDiscovery06:05
LLMs4SchemaDiscovery Workflow06:10
Stage 1: Initial Schema Mining06:43
Stage 1: Initial Schema Mining07:16
Stage 2: Preliminary Schema Refinement07:54
Stage 2: Preliminary Schema Refinement08:47
Expert Feedback Guidelines09:14
Stage 3: Finalize Schema Refinement10:22
3. Application in Material Sciences11:40
Use Case: Schema Discovery in Atomic Layer Deposition (ALD)11:46
What is Atomic Layer Deposition (ALD)?12:18
4. Experiments and Results13:10
Experimental Settings13:11
Effect of Feedback in Schema Refinement14:16
LLM Performance across Schema-Miner Stages15:52
ALD Extracted Schema16:53
5. Conclusion and Future Work17:10
Conclusion17:16
Tool Repository17:38