KnowledgeFace AI Platform Prototype
Summary: Documentation for the “KnowledgeFace” (Scholar AI) web application prototype, designed as a comprehensive learning, contribution, and community platform for AI/ML.
| Resource: KnowledgeFace Live Prototype | GCP Console Overview |
1. Prototype Overview
KnowledgeFace is structured around three core pillars: Learning, Contribution, and Community. The interface is designed for intuitive navigation and seamless access to AI research resources, inspired by platforms like Kaggle.
2. Platform Structure & Navigation
A. Public View
- Registration: Allows signing up via email/password, Google OAuth, or GitHub OAuth.
- Login: Directs authenticated users to the personalized “Scholar” dashboard.
B. User Dashboard (“Scholar” View)
A prominent Create button provides immediate access to contribution workflows:
- Notebook: Create a new interactive code environment.
- Import Notebook: Upload an existing Jupyter/Colab notebook.
- Dataset: Upload a new data resource.
- Model: Link or upload a pre-trained ML model.
- Benchmark: Define a new task for model evaluation.
- Benchmark Task: Submit a model’s performance against an existing benchmark.
C. Persistent Navigation Panel
- Home: Personalized feed and suggested content.
- Datasets: Browse and search public/private data.
- Models: Explore pre-trained models.
- Benchmarks: Leaderboards and evaluation metrics.
- Code: Access user-created notebooks.
- Discussions: Community Q&A forums.
- Learn - Research: Dedicated educational hub.
- More: Access sub-menus for Progression, Documentation, Blog, and Educator Resources.
3. Learn - Research Filters
The “Learn - Research” page uses a horizontal navigation bar to filter content formats:
- Math Concept: Theoretical explanations of core mathematical principles (e.g., Linear Algebra).
- Notes: Study notes and reading summaries.
- Paper: Direct access to external research papers (e.g., Attention Is All You Need).
- Code: Associated implementation code or notebooks.
- Contribute (GitHub): Guidelines on submitting pull requests.
- Research: Summaries of current trends (e.g., Quantum ML).
- Applications: Real-world case studies (e.g., Geospatial Reasoning).
4. Platform Contribution Guidelines
- Datasets: Must include a description, dataset license, and a detailed data dictionary.
- Notebooks: Must be fully executable and document step-by-step with clear markdown.
- Models: Must provide validation metrics, an inference script, and an appropriate license.
- Benchmarks: Must define measurable target metrics and provide a public validation dataset.
5. Computing Quota (Free Tier)
- CPU: 10 hours of computation per week.
- GPU: 5 hours of compute per month.
- Monitoring: Real-time quota metrics are available via the profile menu.
6. Development Roadmap
| Phase | Description | Status / Target |
|---|---|---|
| Phase 1: MVP | Registration, Login, and basic Notebook execution | Done |
| Phase 2: Data & Models | Complete CRUD operations for Datasets and Models | Planned |
| Phase 3: Community & Learning | Launch Discussions forum and populate initial learning entries | Planned |
| Phase 4: Benchmarks | Deploy leaderboards and first benchmark evaluation task | Planned |