Extend temporal modeling to complex temporal dependencies

Extend TimelyRAG’s interpretable temporal distance function to model complex temporal logic, causal relations, and multi-event dependencies in time-critical question answering over overlapping-evolving documents.

Background

TimelyRAG ranks evolving documents by combining semantic similarity with an interpretable temporal distance function. The current formulation represents temporal compatibility through timestamp and validity-interval alignments, with fallback handling for missing or ambiguous temporal signals.

The paper explicitly identifies modeling complex temporal logic, causal relations, and dependencies among multiple events as an unresolved direction for extending the framework beyond its current temporal distance formulation.