Abstract:Few-shot knowledge graph completion (FKGC) aims to query the unseen facts of a relation given its few-shot reference entity pairs. The side effect of noises due to the uncertainty of entities and triples may limit the few-shot learning, but existing FKGC works neglect such uncertainty, which leads them more susceptible to limited reference samples with noises. In this paper, we propose a novel uncertainty-aware few-shot KG completion framework (UFKGC) to model uncertainty for a better understanding of the limited data by learning representations under Gaussian distribution. Uncertainty representation is first designed for estimating the uncertainty scope of the entity pairs after transferring feature representations into a Gaussian distribution. Further, to better integrate the neighbors with uncertainty characteristics for entity features, we design an uncertainty-aware relational graph neural network (UR-GNN) to conduct convolution operations between the Gaussian distributions. Then, multiple random samplings are conducted for reference triples within the Gaussian distribution to generate smooth reference representations during the optimization. The final completion score for each query instance is measured by the designed uncertainty optimization to make our approach more robust to the noises in few-shot scenarios. Experimental results show that our approach achieves excellent performance on two benchmark datasets compared to its competitors.

What problem does this paper attempt to address?

### Problems Addressed by the Paper This paper primarily aims to address the issue of uncertainty in **Few-Shot Knowledge Graph Completion (FKGC)**. Specifically: 1. **Problem Background**: - Knowledge Graphs (KGs) typically contain a large number of entities and their relationships, but they still suffer from incompleteness, meaning many entities and relationships are missing. - Traditional knowledge graph completion methods require a large amount of training data to learn representations for each relationship. However, in real-world scenarios, existing knowledge graphs often face the long-tail problem, where many relationships appear infrequently and lack sufficient training samples. - The task of Few-Shot Knowledge Graph Completion (FKGC) aims to predict unknown facts by given a small number of reference entity pairs. 2. **Main Challenges**: - Impact of noise: Due to the uncertainty of entities and triples, noise may limit the effectiveness of few-shot learning. Existing FKGC methods ignore this uncertainty, making them more vulnerable in the face of limited and noisy reference samples. 3. **Proposed Method**: - This paper proposes a novel uncertainty-aware few-shot knowledge graph completion framework (UFKGC), which uses Gaussian distribution to model uncertainty and better understand limited data. - By designing uncertainty representations to estimate the uncertainty range of entity pairs, and further designing an uncertainty-aware relational graph neural network (UR-GNN) to perform convolution operations. - Multiple random samplings are conducted during the optimization process to generate smooth reference representations, and uncertainty optimization is designed to improve the robustness of the method to noise. 4. **Experimental Results**: - Experiments show that this method achieves excellent performance on two benchmark datasets, outperforming existing competitors. In summary, this paper aims to improve the task of few-shot knowledge graph completion by introducing uncertainty modeling, thereby enhancing the robustness and effectiveness of the model in limited sample scenarios.

Uncertainty-Aware Relational Graph Neural Network for Few-Shot Knowledge Graph Completion

Gaussian Metric Learning for Few-Shot Uncertain Knowledge Graph Completion

Relational Learning with Gated and Attentive Neighbor Aggregator for Few-Shot Knowledge Graph Completion

Relation-Aware Network with Attention-Based Loss for Few-Shot Knowledge Graph Completion

Graph Stochastic Neural Process for Inductive Few-shot Knowledge Graph Completion

A Method Integrating Enhanced Hinge Loss Function for Few-Shot Knowledge Graph Completion

Implicit relational attention network for few-shot knowledge graph completion

MULTIFORM: Few-Shot Knowledge Graph Completion via Multi-modal Contexts.

Relational learning with hierarchical attention encoder and recoding validator for few-shot knowledge graph completion.

Learning Discriminative Representation for Few-Shot Knowledge Graph Completion.

Fine-grained relational learning for few-shot knowledge graph completion

Learning Relation-Specific Representations for Few-shot Knowledge Graph Completion

Towards Trustworthy Knowledge Graph Reasoning: An Uncertainty Aware Perspective

Capsule Neural Tensor Networks with Multi-Aspect Information for Few-shot Knowledge Graph Completion.

Relational multi-scale metric learning for few-shot knowledge graph completion

LLM-based Multi-Level Knowledge Generation for Few-shot Knowledge Graph Completion

Multi-view semantic enhancement model for few-shot knowledge graph completion

Accurate and Scalable Estimation of Epistemic Uncertainty for Graph Neural Networks

Enhancing Path Information with Reinforcement Learning for Few-shot Knowledge Graph Completion.

GFedKG: GNN-based Federated Embedding Model for Knowledge Graph Completion

Adaptive Prototype Interaction Network for Few-Shot Knowledge Graph Completion.