🛢 Information credibility measuring pipeline §

A pipeline used to measure ️⚙️ System objective information credibility

🐾 Steps §

1. 📚 For data-set: Document level relevance (Pyserini BM25) -> ranking R0
2. 📗 Per document: Passage level relevance (monot5) -> ranking R1
3. 📝 Most relevant documents: Passage level reliability (super/unsupervised) -> is (not) reliable
4. 🖇 Score fusion = R0 + R1 + is(not)reliable -> R2

➡ Conclusion §

• 🔓 Unsupervised machine learning > 👮‍♂️ Supervised machine learning
• Score fusion > 🧠 Machine Learning