Expert Finder
Discovery Telecom Experts and Research Papers Using Hybrid Semantic Retrieval plus Knowledge Graph Reasoning.
This project explores how internal knowledge and expertise inside large organizations can be discovered using a combination of semantic search and graph-based reasoning. Instead of returning keyword matches, the system retrieves relevant research, identifies domain experts, and explains relationships between them.
The goal is to improve knowledge discovery, expert identification, and technical insight navigation across complex information environments.
What this system does
Capabilities
- Finds relevant research papers
- Identifies domain experts
- Explains relationships between concepts and authors
- Surfaces non-obvious connections
Current Challenges
- Ranking tradeoffs between semantic vs graph relevance
- External data inconsistency
- Static ranking weights
- Limited contextual expansion depth
Planned Improvements
- Adaptive ranking
- Feedback learning
- Multi-hop reasoning
- Temporal weighting
How Relevance Works
- Papers: semantic similarity + query alignment + graph authority + centrality.
- Experts: semantic evidence + query alignment + recency + topic coverage + graph centrality.
- Ask: grounded answer from retrieved chunks with citations and recommended experts.
2-Minute Tutorial
-
1
Select a telecom-focused query from the examples.
-
2
Start in Papers to inspect relevance and snippets.
-
3
Move to Experts to find ranked researchers and institutions.
-
4
Open Graph to explain relationships between authors, papers, and topics.
-
5
Use Ask for a grounded answer with citations.
Test Mode
Start with Papers
Run one telecom query and validate relevance, snippet quality, and timing.
Switch to Experts
Show ranked experts with score breakdown and why-ranked reasoning.
Open Graph
Click nodes to inspect Author-Paper-Topic paths and explainability.
Finish with Ask
Generate a concise answer with citations and recommended experts.
Try These Telecom Queries
Click any query to jump directly into the interactive workspace.
Seeded Trending Articles
EExApp: GNN-Based Reinforcement Learning for Radio Unit Energy Optimization in 5G O-RAN
2026-02-09
SAGE-5GC: Security-Aware Guidelines for Evaluating Anomaly Detection in the 5G Core Network
2026-02-03
Efficient Self-Learning and Model Versioning for AI-native O-RAN Edge
2026-01-24
Interoperable rApp/xApp Control over O-RAN for Mobility-aware Dynamic Spectrum Allocation
2026-01-24
Seeded Experts
Tommaso Melodia
unknown
Michele Polese
unknown
Leonardo Bonati
unknown
Anastasios Giannopoulos
National Technical University of Athens
Salvatore D'Oro
unknown
Panagiotis K. Gkonis
National and Kapodistrian University of Athens