New neuro-symbolic model LiFTER offers inspectable graph forecasting

Researchers have introduced LiFTER, a novel neuro-symbolic forecasting model designed for continuous-time dynamic graphs. This model preserves observed interactions as temporal facts and uses executable temporal rules to pre-query these facts, allowing for inspectable and verifiable predictions. LiFTER achieves competitive forecasting accuracy while also serving as a "microscope" to analyze the contributions of recurrence, history, and transitions across datasets, tracing them to individual facts. The system demonstrated high accuracy in reconstructing predictions across four benchmarks, turning future-link forecasting into a grounded computation. AI

IMPACT Introduces a new method for inspectable and verifiable predictions in continuous-time dynamic graph forecasting.

RANK_REASON The item is a research paper detailing a new model and methodology. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Influence Flower
• LiFTER
• Link-Fact Temporal Rule Inducer
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →