Researchers have developed a new model called PAGSL (Path-Augmenting Graph Structure Learning) designed to improve explainable scholar recommendations. This model addresses limitations in existing frameworks by mitigating optimization-induced noise and "hub node bias," which often result in overly simplistic explanations. PAGSL integrates representation learning with semantic path knowledge, activating latent academic paths and using a path-based neural network for encoding. Experiments on three academic datasets demonstrated that PAGSL outperforms current state-of-the-art methods, showing significant improvements in F1-score, Accuracy, and AUC, while also providing enhanced interpretability through quantitative results and visualizations. AI
IMPACT Enhances explainability in academic recommendation systems, potentially improving research discovery.
RANK_REASON The item is an academic paper detailing a new model and its experimental evaluation. [lever_c_demoted from research: ic=1 ai=1.0]
Read on arXiv cs.IR (Information Retrieval) →
- alphaXiv
- CatalyzeX Code Finder for Papers
- Connected Papers
- CORE Recommender
- DagsHub
- Gotit.pub
- Hugging Face
- Influence Flower
- Litmaps
- PAGSL
- ResearchGate
- Scholat
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
