Leveraging Implicit Contexts via LLM–Graph Fusion for Temporal Knowledge Graph Reasoning

Structure-aware generative framework for temporal knowledge graph reasoning with historical evidence

Temporal knowledge graph reasoning via multi-granularity knowledge refinement

FITTER: Vocabulary-Agnostic Cross-Domain Inference on Temporal Knowledge Graphs

Temporal knowledge graphs are central to many uses of the Semantic Web, but existing completion methods assume the entities, relation names, and timestamps to be reasoned about are already known at training time, restricting each model to a single graph and vocabulary. We propose FITTER, the first fully-inductive structural model for temporal knowledge graph link prediction that supports cross-domain transfer: the inference graph may contain entirely unseen entities, relation names, and timestamps drawn from a different domain. FITTER represents each predicate by its interaction patterns with others and time through encodings of relative rather than absolute ordering; message-passing fuses local and global temporal context to produce vocabulary-agnostic embeddings. We prove the temporal encoding is time-shift invariant and evaluate FITTER on cross-domain, cross-graph transfer over six temporal knowledge graph benchmarks of diverse domains, granularities, and time spans. FITTER consistently outperforms inductive baselines without retraining, indicating that vocabulary-agnostic structural learning is a viable foundation for inference over the heterogeneous knowledge graphs of the Semantic Web.

MGIT: Multi-Granularity implicit temporal framework for knowledge graph question answering

Temporal-Weighted Transfer Network for Knowledge Graph Extrapolation

Temporal Knowledge Graph (TKG) reasoning aims to predict future events by analyzing historical snapshots across different timestamps, making it an important research direction. However, most existing methods rely either on probabilistic statistical models or on historical propagation mechanisms in isolation, without the ability to simultaneously capture both. This limitation makes it difficult to fully model dependencies between related and unrelated events, while also overlooking time-sensitive features and structural relationships across timestamps. To address these challenges, we propose the Temporal-Weighted Transfer Network (TWTNET), a novel reasoning model that jointly leverages statistical modeling and historical information transfer. TWTNET is composed of two modules: the Temporal Memory Module, which applies a copy-based scoring mechanism to capture event dependencies and emphasize critical entities, and the Dynamic Temporal Information Exchange Module, which learns entity evolution patterns and dynamically integrates temporal information into embeddings. By simultaneously capturing probabilistic regularities and historical dependencies, TWTNET achieves more robust and interpretable extrapolative reasoning. Experiments on five benchmark datasets demonstrate that TWTNET consistently outperforms state-of-the-art models, confirming its effectiveness in TKG reasoning tasks.

ODL-TempLLM: Ontology-Guided and Description Logic-Reasoned Temporal Reasoning with LLMs

Temporal reasoning is crucial for large language models (LLMs) to understand event concurrency and complex temporal interactions in natural language. Recent approaches rely on the LLM to infer temporal relations between events and largely overlook the inherent structural nature of temporal relationships. In this work, we propose ODL-TempLLM ( O ntology-Guided and D escription L ogic–Constrained Temp oral Reasoning with LLM s), a novel paradigm for temporal reasoning with LLMs that shifts focus from internal inference to the explicit modeling of temporal structure. ODL-TempLLM leverages ontology learning to explicitly construct structured temporal knowledge, employs a symbolic reasoner to deductively reason about temporal relations and uses logic-constrained retrieval augmentation to obtain relevant facts. Experiments re-sults evaluated across three datasets via various LLM backbones show that our method outperforms state-of-the-art methods by 2.07–31.83 F1 points and 1.00–30.73 EM points, exhibiting strong generalization and highlighting the potential of explicit temporal reasoning.