Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Uncertain Knowledge Graph Completion via Semi-Supervised Confidence Distribution Learning

Tianxing Wu, Shutong Zhu, Jingting Wang, Ning Xu, Guilin Qi, Haofen Wang

TL;DR

The paper tackles uncertain knowledge graph completion under highly imbalanced triple confidences by introducing Confidence Distribution Learning, which converts confidences into discrete distributions to enrich supervision. It presents ssCDL, a semi-supervised framework combining a CDL-based relational learner (CDL-RL) with a meta-optimized pseudo confidence distribution generator (PCDG) in a meta self-training loop. Empirical results on NL27k and CN15k show ssCDL consistently surpasses state-of-the-art baselines in confidence prediction and, to a lesser extent, link prediction, with ablations confirming the value of confidence distributions and pseudo-labeling. The approach enhances UKG embeddings by leveraging unlabeled data and distributing supervision across neighboring confidence levels, offering a practical advance for UKG completion tasks and potential extensions to reasoning and retrieval-augmented applications.

Abstract

Uncertain knowledge graphs (UKGs) associate each triple with a confidence score to provide more precise knowledge representations. Recently, since real-world UKGs suffer from the incompleteness, uncertain knowledge graph (UKG) completion attracts more attention, aiming to complete missing triples and confidences. Current studies attempt to learn UKG embeddings to solve this problem, but they neglect the extremely imbalanced distributions of triple confidences. This causes that the learnt embeddings are insufficient to high-quality UKG completion. Thus, in this paper, to address the above issue, we propose a new semi-supervised Confidence Distribution Learning (ssCDL) method for UKG completion, where each triple confidence is transformed into a confidence distribution to introduce more supervision information of different confidences to reinforce the embedding learning process. ssCDL iteratively learns UKG embedding by relational learning on labeled data (i.e., existing triples with confidences) and unlabeled data with pseudo labels (i.e., unseen triples with the generated confidences), which are predicted by meta-learning to augment the training data and rebalance the distribution of triple confidences. Experiments on two UKG datasets demonstrate that ssCDL consistently outperforms state-of-the-art baselines in different evaluation metrics.

Uncertain Knowledge Graph Completion via Semi-Supervised Confidence Distribution Learning

TL;DR

The paper tackles uncertain knowledge graph completion under highly imbalanced triple confidences by introducing Confidence Distribution Learning, which converts confidences into discrete distributions to enrich supervision. It presents ssCDL, a semi-supervised framework combining a CDL-based relational learner (CDL-RL) with a meta-optimized pseudo confidence distribution generator (PCDG) in a meta self-training loop. Empirical results on NL27k and CN15k show ssCDL consistently surpasses state-of-the-art baselines in confidence prediction and, to a lesser extent, link prediction, with ablations confirming the value of confidence distributions and pseudo-labeling. The approach enhances UKG embeddings by leveraging unlabeled data and distributing supervision across neighboring confidence levels, offering a practical advance for UKG completion tasks and potential extensions to reasoning and retrieval-augmented applications.

Abstract

Uncertain knowledge graphs (UKGs) associate each triple with a confidence score to provide more precise knowledge representations. Recently, since real-world UKGs suffer from the incompleteness, uncertain knowledge graph (UKG) completion attracts more attention, aiming to complete missing triples and confidences. Current studies attempt to learn UKG embeddings to solve this problem, but they neglect the extremely imbalanced distributions of triple confidences. This causes that the learnt embeddings are insufficient to high-quality UKG completion. Thus, in this paper, to address the above issue, we propose a new semi-supervised Confidence Distribution Learning (ssCDL) method for UKG completion, where each triple confidence is transformed into a confidence distribution to introduce more supervision information of different confidences to reinforce the embedding learning process. ssCDL iteratively learns UKG embedding by relational learning on labeled data (i.e., existing triples with confidences) and unlabeled data with pseudo labels (i.e., unseen triples with the generated confidences), which are predicted by meta-learning to augment the training data and rebalance the distribution of triple confidences. Experiments on two UKG datasets demonstrate that ssCDL consistently outperforms state-of-the-art baselines in different evaluation metrics.
Paper Structure (22 sections, 8 equations, 4 figures, 3 tables, 1 algorithm)

This paper contains 22 sections, 8 equations, 4 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Problem Definition
  5. Confidence Distribution Learning
  6. Methodology
  7. Overview
  8. CDL-based Relational Learner
  9. Pseudo Confidence Distribution Generator
  10. Meta Self-training
  11. Experiments
  12. Experimental Setup
  13. Confidence Prediction
  14. Link Prediction
  15. Ablation Study
  16. ...and 7 more sections

Figures (4)

  • Figure 1: (a) The histogram of triple confidence in NELL; (b) An example of a triple confidence equal to $0.78$, which is transformed to a confidence distribution.
  • Figure 2: (a) The overview of ssCDL; (b) The framework of CDL-RL and PCDG.
  • Figure 3: The comparison results of confidence prediction on low-confidence triples in NL27k and CN15k.
  • Figure : The impacts of hyper-parameters on confidence prediction. (a) (b) (c) show the impacts of $\sigma$, $w_p$, and the threshold for pseudo labeled data selection on NL27k and CN15k respectively.

Theorems & Definitions (2)

  • Definition 1
  • Definition 2