Temporal knowledge graphs (TKGs) are crucial for modeling evolving real-world facts and are widely applied in event forecasting and risk analysis. However, current TKG reasoning models struggle to separate causal signals from noisy observations, align temporal dynamics with semantic structures, and integrate long-term and short-term knowledge effectively. To address these challenges, we propose the Temporal Causal Contrast Graph Network (TCCGN), a unified framework that disentangles causal features from noise via orthogonal decomposition and adversarial learning; applies dual-domain contrastive learning to enhance both temporal and semantic consistency; and introduces a gated fusion module for adaptive integration of static and dynamic features across time scales. Extensive experiments on five benchmarks (ICEWS14/05-15/18, YAGO, GDELT) show that TCCGN consistently outperforms prior models. On ICEWS14, it achieves 42.46% MRR and 31.63% Hits@1, surpassing RE-GCN by 1.21 points. On the high-noise GDELT dataset, it improves MRR by 1.0%. These results highlight TCCGN’s robustness and its promise for real-world temporal reasoning tasks involving fine-grained causal inference under noisy conditions.

Temporal knowledge graph (TKG) completion is essential for predicting missing links by modeling the evolution of entities and relations over time, with critical applications in areas like event forecasting and trend analysis. However, existing methods often prioritize entity and relationship dynamics, treating temporal information as a secondary feature, which hampers their ability to effectively capture changes across diverse time scales. This limitation restricts their capacity to model complex temporal patterns, such as periodicity and long-term trends, vital for representing time-sensitive data accurately. To overcome these challenges, we propose TETFD (time embedding and time-frequency decoder), a novel method that incorporates time embeddings to explicitly capture temporal dynamics and integrates them into a specialized decoder. This decoder not only processes entity and relationship data but also incorporates frequency domain features extracted from temporal sequences, enhancing the model’s ability to reflect both short-term variations and prolonged trends. Experimental results across five benchmark datasets demonstrate that TETFD significantly outperforms existing approaches, establishing new standards in TKG completion tasks.