Collect Literature
Use databases, advisor suggestions, and AI-assisted search prompts to build an initial reading pool, then preserve source metadata in Zotero.
AI-KM Workflow
AI-Assisted Learning & Knowledge Management
This page turns "using AI tools" into a reviewable learning workflow: how literature is collected, read, organized, and converted into research communication.
AI Workflow Zotero Obsidian
Project path
Place AI-KM inside the evidence system
This page explains the workflow boundary without publishing a private knowledge vault. Reviewers can return to projects, open the materials hub, or inspect the claim matrix below.
Workflow Map
A lightweight toolchain connects reading, organization, and output.
Read and Annotate
Separate abstracts, figures, methods, and questions so a PDF becomes reusable understanding instead of a saved file only.
Organize Knowledge
Use Obsidian to connect course concepts, reading notes, terminology cards, and research questions into a revisitable network.
Turn Notes into Output
Use Markdown and LaTeX to turn notes into weekly reviews, research ideas, presentation outlines, and project narratives.
Tool Stack
Tools support a traceable learning process; generated text is not treated as fact by itself.
AI Tools
Draft search strings, decompose paper structure, and support English expression while keeping original reading and fact checks in the loop.
Zotero
Manage bibliographic records, PDFs, tags, and citation metadata so sources remain traceable.
Obsidian
Store concept cards, reading notes, and topic maps as a long-term knowledge base.
Markdown / LaTeX
Convert reading outcomes into clear, versionable research-writing drafts.
Publicly Explainable Outputs
Reading note templateTopic mapWeekly reviewResearch question listPresentation outlineBilingual writing draft
Publication Boundary
The public page explains the workflow, not the private Obsidian vault, raw PDFs, account information, unsanitized note screenshots, or formal application packets. Claims without public certificates remain marked as case-backed or narrative-bound.
Claim-to-Evidence Matrix
Each ability claim shows its evidence status, publication boundary, and related routes so narrative is not mistaken for hard proof.
Modeling Practice
Able to turn open-ended problems into variables, assumptions, models, analysis, and written arguments.
- Boundary
- Supported by public award certificates; full competition papers are not published on the public site.
- Evidence count
- 5 public evidence item(s)
Academic Performance
GPA, ranking, scholarship, and university merit recognition support sustained academic performance.
- Boundary
- The public site shows sanitized outcomes and certificates, not student ID, birth date, or full review files.
- Evidence count
- 2 public evidence item(s)
Chemistry Foundation
Chemistry competition training supports later interest in organic chemistry, computational chemistry, and AI chemistry.
- Boundary
- Evidence is limited to public certificates; undergraduate research work remains framed as direction and training.
- Evidence count
- 2 public evidence item(s)
AI-KM Research Workflow
Uses AI tools, Zotero, Obsidian, Markdown, and LaTeX to manage reading, reflection, and research communication.
- Boundary
- This is a public case narrative; private vault content, raw PDFs, account data, and unsanitized note screenshots stay unpublished.
- Evidence count
- No certificate evidence yet
Public Portfolio System
Organizes resumes, evidence, project narratives, material routing, and automated QA into a maintainable bilingual static site.
- Boundary
- The site itself is the project case; source and deployment records are maintained through the repository and Actions.
- Evidence count
- No certificate evidence yet
Campus Collaboration
Debate, volunteer service, and sports activities support communication, collaboration, organization, and sustained participation.
- Boundary
- Public evidence explains broader development, not a substitute for formal student-work records.
- Evidence count
- 5 public evidence item(s)