Time Imprint: Learning Time-Aware Representations in Multi-Modal Knowledge Graphs
Quick Answer
The Time Imprint framework enhances Multi-Modal Knowledge Graphs by treating time as a distinct modality, achieving state-of-the-art link prediction performance with up to 6.07% improvement in Hits@1.
Quick Take
The Time Imprint framework enhances Multi-Modal Knowledge Graphs by treating time as a distinct modality, achieving state-of-the-art link prediction performance with up to 6.07% improvement in Hits@1. By addressing challenges like sparse temporal semantics and multiple timestamps, it effectively disambiguates entities, yielding significant gains on ambiguous samples. The approach adds minimal training overhead while clarifying the benefits of temporal information in representation learning.
Key Points
- Time Imprint treats time as a separate modality in MMKGs.
- Achieves up to 6.07% improvement in Hits@1 on link prediction tasks.
- Addresses challenges of sparse temporal semantics and multiple timestamps.
- Utilizes attention pooling for effective temporal embedding aggregation.
- Demonstrates significant gains on the top-1% ambiguous samples.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Multi-Modal Knowledge Graphs (MMKGs) enrich entities with multiple modalities such as text and images, yet entities with highly similar multi-modal features remain difficult to distinguish. Temporal information of an entity can serve as an additional modality to disambiguate such entities, but existing approaches rarely treat time as a separate modality alongside text and images due to two major challenges: (1) sparse temporal semantics, which hinder alignment with richer modalities, and (2) multiple timestamps, which introduce noise or reduce robustness in representation learning. To address these challenges, we propose Time Imprint, a framework that treats time as an entity-level modality and jointly aligns temporal, textual, and visual representations via a three-view contrastive objective. Additionally, to mitigate multi-timestamp ambiguity, Time Imprint studies a compact timestamp subset selection design space and aggregates the selected timestamps into a discriminative temporal embedding with attention pooling, balancing temporal specificity and robustness. Experiments on three MMKG benchmarks demonstrate that Time Imprint achieves state-of-the-art link prediction performance, improving Hits@1 by up to 6.07\% overall and yielding up to 58\% gains on the subset of the top-1\% ambiguity samples. We further examine different fusion strategies and the sensitivity to timestamp availability and quality, clarifying when and why time-as-modality is most beneficial, while adding only modest training overhead. We release our code at this https URL.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2607.09777 [cs.CV] |
| (or arXiv:2607.09777v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.09777 arXiv-issued DOI via DataCite |
Submission history
From: Pengyu Zhang [view email]
[v1]
Wed, 8 Jul 2026 06:52:44 UTC (682 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.