Problem
Knowledge graph models need careful sampling and ranking evaluation to make research results meaningful.
ResearchResearch engineeringMachine learningData systems
Augmented KGE Research
This work gives the portfolio a research signal: experimental design, model evaluation, ranking metrics, and reproducible machine-learning workflows. The future public-facing version should make the evaluation story easier to scan and understand.
Signal
Link prediction
Signal
Ranking metrics
Signal
Reproducible experiments
Case study
What this project demonstrates
Approach
- Develop model experiments with negative sampling strategies.
- Evaluate link prediction using ranking-focused metrics.
- Optimize data-processing workflows for faster experimentation.
Outcome
- Shows research fluency alongside practical engineering implementation.
- Creates a bridge between AI research, data pipelines, and measurable evaluation.
Stack
Technology used or planned
Tooling belongs here, where it supports the project story instead of crowding the homepage hero.
PyTorchPythonKnowledge GraphsEvaluation