TempHypE-GNN: Hyperbolic Graph Neural ODEs for Hierarchical Temporal Knowledge Graphs
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Pervasive systems such as IoT networks and dynamic knowledge-driven platforms demand models that can capture both temporal evolution and hierarchical structure with high fidelity. While TempHypE introduced a powerful combination of hyperbolic geometry and Neural Ordinary Differential Equations (ODEs) for continuous-time temporal reasoning, it lacks explicit mechanisms to model local graph connectivity and relational propagation. We address this limitation with TempHypE-GNN, a novel extension that integrates hyperbolic graph neural networks (GNNs) into the TempHypE framework. The GNN component enables message passing in hyperbolic space, allowing entities to aggregate neighborhood information across multi-hop relations and improving local relational reasoning under hierarchical constraints. This enhancement complements the global temporal modeling of Neural ODEs, leading to richer representations. TempHypE-GNN jointly embeds entities in the Poincaré ball to capture hierarchical dependencies, applies Neural ODEs to model continuous-time dynamics, and incorporates hyperbolic GNN layers for relational feature propagation. Experimental results on ICEWS14, ICEWS18, and GDELT demonstrate that TempHypE-GNN consistently outperforms both Euclidean models and static hyperbolic baselines. Specifically, it achieves a 17.6% relative improvement in MRR over DySAT on the GDELT dataset and an 8.5% relative increase in Hits@10 over HyperKG on ICEWS18. Furthermore, ablation studies averaged across all benchmarks show that incorporating hyperbolic GNN layers into the TempHypE architecture leads to performance gains of up to 4.3% in MRR and 5.9% in Hits@10, highlighting the added value of relational message passing in hyperbolic space. This demonstrates not only theoretical advances but also the potential for deployment in real-world domains such as IoT, smart-city analytics, and large-scale temporal knowledge management.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.006 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.005 | 0.002 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it