《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (9): 2809-2819.DOI: 10.11772/j.issn.1001-9081.2025081025
• 人工智能 • 上一篇
收稿日期:2025-09-08
修回日期:2025-11-13
接受日期:2025-11-14
发布日期:2025-11-21
出版日期:2026-09-10
通讯作者:
李贯峰
作者简介:尹仁蕊(1999—),女,宁夏中卫人,硕士研究生,CCF会员,主要研究方向:不确定知识图谱嵌入和推理、知识图谱不确定性检测基金资助:
Renrui YIN, Guanfeng LI(
), Shizhuo WANG, Yaya JIANG
Received:2025-09-08
Revised:2025-11-13
Accepted:2025-11-14
Online:2025-11-21
Published:2026-09-10
Contact:
Guanfeng LI
About author:YIN Renrui, born in 1999, M. S. candidate. Her research interests include uncertain knowledge graph embedding and reasoning, uncertainty detection in knowledge graphs.Supported by:摘要:
现有方法虽然在不确定知识图谱嵌入(UKGE)上取得了初步进展,但仍存在诸多不足,尤其在对复杂关系的建模方面表现有限。UKGE框架仅能刻画对称等简单关系模式。尽管复数平面上的旋转模型RotatE能借助复数旋转刻画多关系事实,简洁而高效,但它的表达能力受限于二维复数空间,难以刻画高维旋转,且易出现“万向锁”现象,限制了对复杂关系的进一步刻画。针对上述局限,提出一种基于四元数的不确定知识图谱嵌入模型QUKGE,该模型将实体表示为四维四元数向量,并将关系建模为四元数空间中的旋转算子。在评分函数层面,利用实体与关系的汉密尔顿积实现四元数旋转,从而系统地表征对称、反对称、逆关系、可交换及不可交换等复杂关系模式。四元数的三维旋转特性有效缓解了高维旋转中的“万向锁”,并显著提升了对复杂关系的刻画能力。为验证QUKGE的有效性,在CN15K、NL27K和PPI5K这3个公开不确定知识图谱数据集上进行5项对比实验。实验结果表明,QUKGE在NL27K数据集上与MUKGE(Multiplex Uncertain Knowledge Graph Embedding)模型相比,在关系事实分类任务中的F1分数与准确率分别提升了2.9和1.0个百分点,为UKGE的进一步研究提供了支撑。
中图分类号:
尹仁蕊, 李贯峰, 王世卓, 蒋娅娅. 基于四元数的不确定知识图谱嵌入[J]. 计算机应用, 2026, 46(9): 2809-2819.
Renrui YIN, Guanfeng LI, Shizhuo WANG, Yaya JIANG. Quaternion-based uncertain knowledge graph embedding[J]. Journal of Computer Applications, 2026, 46(9): 2809-2819.
| 数据集 | 实体 数 | 关系 数 | 三元组数 | 置信度 | |||
|---|---|---|---|---|---|---|---|
| 训练 | 验证 | 测试 | 平均值 | 标准值 | |||
| CN15K | 15 000 | 36 | 204 984 | 16 881 | 19 293 | 0.629 | 0.232 |
| NL27K | 27 221 | 404 | 149 100 | 12 278 | 14 034 | 0.797 | 0.242 |
| PPI5K | 4 999 | 7 | 230 929 | 19 017 | 21 720 | 0.415 | 0.213 |
表1 数据集的基本信息
Tab. 1 Basic information of datasets
| 数据集 | 实体 数 | 关系 数 | 三元组数 | 置信度 | |||
|---|---|---|---|---|---|---|---|
| 训练 | 验证 | 测试 | 平均值 | 标准值 | |||
| CN15K | 15 000 | 36 | 204 984 | 16 881 | 19 293 | 0.629 | 0.232 |
| NL27K | 27 221 | 404 | 149 100 | 12 278 | 14 034 | 0.797 | 0.242 |
| PPI5K | 4 999 | 7 | 230 929 | 19 017 | 21 720 | 0.415 | 0.213 |
| 模型 | 得分函数 | 参数 |
|---|---|---|
| QUKGETransE | ||
| QUKGEHolE | ||
| QUKGEComplEx | ||
| QUKGERotatE |
表2 QUKGE变体模型的相关信息
Tab. 2 Relevant information about QUKGE variant models
| 模型 | 得分函数 | 参数 |
|---|---|---|
| QUKGETransE | ||
| QUKGEHolE | ||
| QUKGEComplEx | ||
| QUKGERotatE |
| 模型 | CN15K | NL27K | PPI5K | |||
|---|---|---|---|---|---|---|
| MSE | MAE | MSE | MAE | MSE | MAE | |
| URGE | 0.103 2 | 0.227 2 | 0.074 8 | 0.113 5 | 0.014 4 | 0.060 0 |
| UKGErect | 0.086 1 | 0.199 0 | 0.023 6 | 0.069 0 | 0.009 5 | 0.037 9 |
| UKGElogi | 0.098 6 | 0.207 4 | 0.034 3 | 0.079 3 | 0.009 6 | 0.040 7 |
| BEUrRE | 0.074 9 | 0.020 1 | 0.068 9 | — | — | |
| MUKGErect | 0.070 5 | 0.197 3 | 0.016 3 | 0.067 6 | 0.030 2 | |
| MUKGElogi | 0.099 9 | 0.204 2 | 0.021 1 | 0.003 6 | ||
| QUKGETransE | 0.127 3 | 0.252 6 | 0.073 3 | 0.107 2 | 0.008 1 | 0.048 4 |
| QUKGEHolE | 0.110 3 | 0.224 4 | 0.051 2 | 0.095 9 | 0.009 0 | 0.047 8 |
| QUKGEComplEx | 0.109 0 | 0.209 4 | 0.026 7 | 0.075 5 | 0.005 9 | 0.040 6 |
| QUKGERotatE | 0.100 4 | 0.206 7 | 0.023 3 | 0.072 3 | 0.007 8 | 0.045 4 |
| QUKGE | 0.199 5 | 0.060 8 | 0.002 8 | 0.023 4 | ||
表3 不同模型的置信度预测结果
Tab. 3 Confidence prediction results of different models
| 模型 | CN15K | NL27K | PPI5K | |||
|---|---|---|---|---|---|---|
| MSE | MAE | MSE | MAE | MSE | MAE | |
| URGE | 0.103 2 | 0.227 2 | 0.074 8 | 0.113 5 | 0.014 4 | 0.060 0 |
| UKGErect | 0.086 1 | 0.199 0 | 0.023 6 | 0.069 0 | 0.009 5 | 0.037 9 |
| UKGElogi | 0.098 6 | 0.207 4 | 0.034 3 | 0.079 3 | 0.009 6 | 0.040 7 |
| BEUrRE | 0.074 9 | 0.020 1 | 0.068 9 | — | — | |
| MUKGErect | 0.070 5 | 0.197 3 | 0.016 3 | 0.067 6 | 0.030 2 | |
| MUKGElogi | 0.099 9 | 0.204 2 | 0.021 1 | 0.003 6 | ||
| QUKGETransE | 0.127 3 | 0.252 6 | 0.073 3 | 0.107 2 | 0.008 1 | 0.048 4 |
| QUKGEHolE | 0.110 3 | 0.224 4 | 0.051 2 | 0.095 9 | 0.009 0 | 0.047 8 |
| QUKGEComplEx | 0.109 0 | 0.209 4 | 0.026 7 | 0.075 5 | 0.005 9 | 0.040 6 |
| QUKGERotatE | 0.100 4 | 0.206 7 | 0.023 3 | 0.072 3 | 0.007 8 | 0.045 4 |
| QUKGE | 0.199 5 | 0.060 8 | 0.002 8 | 0.023 4 | ||
| 模型 | CN15K | NL27K | PPI5K | |||
|---|---|---|---|---|---|---|
| Linear | Exp | Linear | Exp | Linear | Exp | |
| TransE | 0.601 | 0.591 | 0.730 | 0.722 | 0.710 | 0.700 |
| DistMult | 0.689 | 0.677 | 0.911 | 0.897 | 0.894 | 0.880 |
| ComplEx | 0.723 | 0.712 | 0.921 | 0.913 | 0.896 | 0.881 |
| URGE | 0.572 | 0.570 | 0.593 | 0.593 | 0.726 | 0.723 |
| UKGErect | 0.773 | 0.775 | 0.939 | 0.942 | 0.946 | 0.946 |
| UKGElogi | 0.789 | 0.788 | 0.970 | 0.969 | ||
| BEUrRE | 0.796 | 0.795 | 0.942 | 0.942 | 0.951 | 0.952 |
| MUKGErect | 0.832 | 0.835 | 0.877 | 0.882 | 0.959 | 0.958 |
| MUKGElogi | 0.849 | 0.850 | 0.945 | 0.947 | ||
| QUKGETransE | 0.804 | 0.796 | 0.893 | 0.896 | 0.986 | 0.993 |
| QUKGEHolE | 0.811 | 0.808 | 0.882 | 0.925 | 0.980 | 0.992 |
| QUKGEComplEx | 0.819 | 0.806 | 0.947 | 0.953 | 0.983 | 0.992 |
| QUKGERotatE | 0.834 | 0.820 | 0.942 | 0.942 | 0.987 | 0.994 |
| QUKGE | 0.967 | 0.968 | 0.997 | 0.995 | ||
表4 不同模型的关系事实排序结果
Tab. 4 Ranking results of relational facts by different models
| 模型 | CN15K | NL27K | PPI5K | |||
|---|---|---|---|---|---|---|
| Linear | Exp | Linear | Exp | Linear | Exp | |
| TransE | 0.601 | 0.591 | 0.730 | 0.722 | 0.710 | 0.700 |
| DistMult | 0.689 | 0.677 | 0.911 | 0.897 | 0.894 | 0.880 |
| ComplEx | 0.723 | 0.712 | 0.921 | 0.913 | 0.896 | 0.881 |
| URGE | 0.572 | 0.570 | 0.593 | 0.593 | 0.726 | 0.723 |
| UKGErect | 0.773 | 0.775 | 0.939 | 0.942 | 0.946 | 0.946 |
| UKGElogi | 0.789 | 0.788 | 0.970 | 0.969 | ||
| BEUrRE | 0.796 | 0.795 | 0.942 | 0.942 | 0.951 | 0.952 |
| MUKGErect | 0.832 | 0.835 | 0.877 | 0.882 | 0.959 | 0.958 |
| MUKGElogi | 0.849 | 0.850 | 0.945 | 0.947 | ||
| QUKGETransE | 0.804 | 0.796 | 0.893 | 0.896 | 0.986 | 0.993 |
| QUKGEHolE | 0.811 | 0.808 | 0.882 | 0.925 | 0.980 | 0.992 |
| QUKGEComplEx | 0.819 | 0.806 | 0.947 | 0.953 | 0.983 | 0.992 |
| QUKGERotatE | 0.834 | 0.820 | 0.942 | 0.942 | 0.987 | 0.994 |
| QUKGE | 0.967 | 0.968 | 0.997 | 0.995 | ||
| 模型 | CN15K | NL27K | PPI5K | |||
|---|---|---|---|---|---|---|
| F1 | Acc | F1 | Acc | F1 | Acc | |
| TransE | 23.1 | 67.9 | 65.1 | 53.4 | 83.2 | 98.5 |
| DistMult | 27.9 | 71.1 | 72.1 | 70.1 | 86.9 | 97.1 |
| ComplEx | 18.9 | 73.2 | 63.3 | 53.4 | 83.2 | 98.9 |
| URGE | 21.2 | 86.0 | 83.6 | 88.7 | 85.2 | 98.6 |
| UKGErect | 28.8 | 92.3 | 95.2 | 95.1 | ||
| UKGElogi | 25.9 | 90.1 | 88.4 | 93.0 | 94.5 | 98.7 |
| BEUrRE | 28.7 | 89.9 | 95.6 | 95.0 | ||
| MUKGErect | 32.0 | 87.5 | 92.5 | 94.7 | 99.3 | |
| MUKGElogi | 28.3 | 90.2 | 91.5 | 94.4 | 99.7 | |
| QUKGETransE | 22.6 | 85.8 | 89.1 | 90.4 | 88.4 | 96.9 |
| QUKGEHolE | 23.4 | 85.7 | 90.3 | 91.9 | 83.3 | 94.2 |
| QUKGEComplEx | 25.6 | 87.6 | 84.8 | 86.5 | 92.7 | 98.0 |
| QUKGERotatE | 29.2 | 89.4 | 86.0 | 86.9 | 82.3 | 95.7 |
| QUKGE | 90.8 | 94.4 | 97.8 | |||
表5 不同模型的关系事实分类结果 (%)
Tab. 5 Classification results of relational facts by different models
| 模型 | CN15K | NL27K | PPI5K | |||
|---|---|---|---|---|---|---|
| F1 | Acc | F1 | Acc | F1 | Acc | |
| TransE | 23.1 | 67.9 | 65.1 | 53.4 | 83.2 | 98.5 |
| DistMult | 27.9 | 71.1 | 72.1 | 70.1 | 86.9 | 97.1 |
| ComplEx | 18.9 | 73.2 | 63.3 | 53.4 | 83.2 | 98.9 |
| URGE | 21.2 | 86.0 | 83.6 | 88.7 | 85.2 | 98.6 |
| UKGErect | 28.8 | 92.3 | 95.2 | 95.1 | ||
| UKGElogi | 25.9 | 90.1 | 88.4 | 93.0 | 94.5 | 98.7 |
| BEUrRE | 28.7 | 89.9 | 95.6 | 95.0 | ||
| MUKGErect | 32.0 | 87.5 | 92.5 | 94.7 | 99.3 | |
| MUKGElogi | 28.3 | 90.2 | 91.5 | 94.4 | 99.7 | |
| QUKGETransE | 22.6 | 85.8 | 89.1 | 90.4 | 88.4 | 96.9 |
| QUKGEHolE | 23.4 | 85.7 | 90.3 | 91.9 | 83.3 | 94.2 |
| QUKGEComplEx | 25.6 | 87.6 | 84.8 | 86.5 | 92.7 | 98.0 |
| QUKGERotatE | 29.2 | 89.4 | 86.0 | 86.9 | 82.3 | 95.7 |
| QUKGE | 90.8 | 94.4 | 97.8 | |||
| 关系 | MRR | 三元组数 | 关系类型 | |
|---|---|---|---|---|
| QUKGE | QUKGERotatE | |||
| city located in geopolitical location | 1.000 | 0.767 | 5 | 1-1 |
| headquartered in | 0.938 | 0.938 | 8 | 1-1 |
| organization lead by person | 0.831 | 0.805 | 44 | 1-1 |
| proxy for | 0.753 | 0.510 | 29 | 1-1 |
| inverse of arthropod and other arthropods | 0.453 | 0.352 | 6 | 1-n |
| companies headquartered here | 0.264 | 0.301 | 11 | 1-n |
| agent collaborates with agent | 0.443 | 0.390 | 187 | 1-n |
| city television station | 0.407 | 0.365 | 14 | 1-n |
| animal that feeds on insects | 0.045 | 0.046 | 7 | 1-n |
| language of university | 0.096 | 0.080 | 2 | n-1 |
表6 QUKGE和QUKGERotatE在NL27K数据集上的MRR对比
Tab. 6 Comparison of MRR between QUKGE and QUKGERotatE on NL27K dataset
| 关系 | MRR | 三元组数 | 关系类型 | |
|---|---|---|---|---|
| QUKGE | QUKGERotatE | |||
| city located in geopolitical location | 1.000 | 0.767 | 5 | 1-1 |
| headquartered in | 0.938 | 0.938 | 8 | 1-1 |
| organization lead by person | 0.831 | 0.805 | 44 | 1-1 |
| proxy for | 0.753 | 0.510 | 29 | 1-1 |
| inverse of arthropod and other arthropods | 0.453 | 0.352 | 6 | 1-n |
| companies headquartered here | 0.264 | 0.301 | 11 | 1-n |
| agent collaborates with agent | 0.443 | 0.390 | 187 | 1-n |
| city television station | 0.407 | 0.365 | 14 | 1-n |
| animal that feeds on insects | 0.045 | 0.046 | 7 | 1-n |
| language of university | 0.096 | 0.080 | 2 | n-1 |
| 对比模型 | 单次训练 耗时/s | 全程平均 每轮耗时/s | 收敛迭代次数 | 实际训练轮数 |
|---|---|---|---|---|
| QUKGETransE | 0.318 | 0.281 | 2 180 | 2 380 |
| QUKGEHolE | 0.329 | 0.290 | 1 500 | 1 700 |
| QUKGEComplEx | 0.467 | 0.431 | 1 260 | 1 460 |
| QUKGERotatE | 0.459 | 0.424 | 2 000 | 2 200 |
| QUKGETF | 0.634 | 0.455 | 1 420 | 1 620 |
| QUKGE | 0.420 | 0.392 | 1 320 | 1 520 |
表7 不同模型的效率对比
Tab. 7 Comparison of efficiencies of different models
| 对比模型 | 单次训练 耗时/s | 全程平均 每轮耗时/s | 收敛迭代次数 | 实际训练轮数 |
|---|---|---|---|---|
| QUKGETransE | 0.318 | 0.281 | 2 180 | 2 380 |
| QUKGEHolE | 0.329 | 0.290 | 1 500 | 1 700 |
| QUKGEComplEx | 0.467 | 0.431 | 1 260 | 1 460 |
| QUKGERotatE | 0.459 | 0.424 | 2 000 | 2 200 |
| QUKGETF | 0.634 | 0.455 | 1 420 | 1 620 |
| QUKGE | 0.420 | 0.392 | 1 320 | 1 520 |
| lr | MSE | MAE | Linear | Exp | F1/% | Acc/% |
|---|---|---|---|---|---|---|
| 0.000 2 | 0.79 | 6.18 | 0.981 | 0.991 | 77.4 | 94.5 |
| 0.000 5 | 0.30 | 2.92 | 0.996 | 0.989 | 93.6 | 98.2 |
| 0.001 0 | 0.28 | 2.34 | 0.998 | 0.997 | 97.8 | 99.4 |
| 0.005 0 | 0.30 | 2.53 | 0.993 | 0.992 | 96.4 | 99.0 |
| 0.010 0 | 0.32 | 2.60 | 0.998 | 0.997 | 96.8 | 99.1 |
表8 在PPI5K数据集上不同学习率对性能的影响
Tab. 8 Impact of different learning rates on performance on PPI5K dataset
| lr | MSE | MAE | Linear | Exp | F1/% | Acc/% |
|---|---|---|---|---|---|---|
| 0.000 2 | 0.79 | 6.18 | 0.981 | 0.991 | 77.4 | 94.5 |
| 0.000 5 | 0.30 | 2.92 | 0.996 | 0.989 | 93.6 | 98.2 |
| 0.001 0 | 0.28 | 2.34 | 0.998 | 0.997 | 97.8 | 99.4 |
| 0.005 0 | 0.30 | 2.53 | 0.993 | 0.992 | 96.4 | 99.0 |
| 0.010 0 | 0.32 | 2.60 | 0.998 | 0.997 | 96.8 | 99.1 |
| 嵌入维度 | MSE | MAE | Linear | Exp | F1/% | Acc/% |
|---|---|---|---|---|---|---|
| 8 | 0.27 | 2.65 | 0.997 | 0.996 | 96.4 | 99.0 |
| 16 | 0.28 | 2.34 | 0.998 | 0.997 | 97.8 | 99.4 |
| 32 | 0.29 | 2.35 | 0.998 | 0.997 | 97.7 | 99.4 |
| 64 | 0.34 | 2.48 | 0.997 | 0.996 | 96.4 | 99.3 |
| 128 | 0.36 | 2.53 | 0.996 | 0.995 | 96.5 | 99.2 |
| 256 | 0.38 | 2.57 | 0.996 | 0.995 | 96.3 | 99.2 |
表9 在PPI5K数据集上不同嵌入维度对性能的影响
Tab. 9 Impact of different embedding dimensions on performance on PPI5K dataset
| 嵌入维度 | MSE | MAE | Linear | Exp | F1/% | Acc/% |
|---|---|---|---|---|---|---|
| 8 | 0.27 | 2.65 | 0.997 | 0.996 | 96.4 | 99.0 |
| 16 | 0.28 | 2.34 | 0.998 | 0.997 | 97.8 | 99.4 |
| 32 | 0.29 | 2.35 | 0.998 | 0.997 | 97.7 | 99.4 |
| 64 | 0.34 | 2.48 | 0.997 | 0.996 | 96.4 | 99.3 |
| 128 | 0.36 | 2.53 | 0.996 | 0.995 | 96.5 | 99.2 |
| 256 | 0.38 | 2.57 | 0.996 | 0.995 | 96.3 | 99.2 |
| [1] | McGuinness D L, van Harmelen F. OWL web ontology language overview [EB/OL]. (2004-02-10) [2025-05-28].. |
| [2] | Bordes A, Usunier N, Garcia-Duran A, et al. Translating embeddings for modeling multi-relational data [C]// NeurIPS 2013 — Volume 2. Red Hook: Curran Associates Inc., 2013: 2787-2795. |
| [3] | 赖舒婷,温煜. 医学知识图谱在医疗问答系统中的应用研究[J]. 信息与电脑, 2025, 37(17): 24-26. |
| Lai Shuting, Wen Yu. Research on the application of medical knowledge graph in medical question answering system [J]. Information and Computer, 2025, 37(17): 24-26. | |
| [4] | Speer R, Chin J, Havasi C. ConceptNet 5.5: an open multilingual graph of general knowledge [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2017, 31(1): 4444-4451. |
| [5] | Mitchell T, Cohen W, Hruschka E, et al. Never-ending learning[J]. Communications of the ACM, 2018, 61(5): 103-115. |
| [6] | Wu W, Li H, Wang H, et al. Probase: a probabilistic taxonomy for text understanding [C]// SIGMOD 2012. New York: ACM, 2012: 481-492. |
| [7] | Sun Z, Deng Z H, Nie J Y, et al. RotatE: knowledge graph embedding by relational rotation in complex space [PP/OL]. arXiv (2019-02-26) [2025-05-26].. |
| [8] | Trouillon T, Welbl J, Riedel S, et al. Complex embeddings for simple link prediction [J]. Proceedings of Machine Learning Research, 2016, 48: 2071-2080. |
| [9] | Nickel M, Rosasco L, Poggio T. Holographic embeddings of knowledge graphs [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2016, 30(1): 1955-1961. |
| [10] | Lin Y, Liu Z, Sun M, et al. Learning entity and relation embeddings for knowledge graph completion [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2015, 29(1): 2181-2187. |
| [11] | Wang Z, Zhang J, Feng J, et al. Knowledge graph embedding by translating on hyperplanes [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2014, 28(1): 1112-1119. |
| [12] | Yang B, Yih W T, He X, et al. Embedding entities and relations for learning and inference in knowledge bases[PP/OL]. V4. arXiv (2021-08-29) [2025-05-11].. |
| [13] | Bordes A, Glorot X, Weston J, et al. A semantic matching energy function for learning with multi-relational data: application to word-sense disambiguation [J]. Machine Learning, 2014, 94(2): 233-259. |
| [14] | Ebisu T, Ichise R. TorusE: knowledge graph embedding on a lie group [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2018, 32(1): 1819-1826. |
| [15] | Zhang S, Tay Y, Yao L, et al. Quaternion knowledge graph embeddings [C]// NeurIPS 2019. Red Hook: Curran Associates Inc., 2019: 2735-2745. |
| [16] | Nickel M, Kiela D. Poincaré embeddings for learning hierarchical representations [C]// NeurIPS 2017. Red Hook: Curran Associates Inc., 2017: 6341-6350. |
| [17] | Nguyen D Q, Nguyen T D, Nguyen D Q, et al. A novel embedding model for knowledge base completion based on convolutional neural network [C]// NAACL-HLT 2018, Volume 2 (Short Papers). Stroudsburg: ACL, 2018: 327-333. |
| [18] | Dettmers T, Minervini P, Stenetorp P, et al. Convolutional 2D knowledge graph embeddings [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2018, 32(1): 1811-1818. |
| [19] | Shang C, Tang Y, Huang J, et al. End-to-end structure-aware convolutional networks for knowledge base completion [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 3060-3067. |
| [20] | Hu J, Cheng R, Huang Z, et al. On embedding uncertain graphs[C]// CIKM 2017. New York: ACM, 2017: 157-166. |
| [21] | Xie R, Liu Z, Lin F, et al. Does William Shakespeare REALLY write Hamlet? knowledge representation learning with confidence[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2018, 32(1): 4954-4961. |
| [22] | Fan M, Zhou Q, Zheng T F. Learning embedding representations for knowledge inference on imperfect and incomplete repositories[C]// WI 2016. Piscataway: IEEE, 2016: 42-48. |
| [23] | Chen X, Chen M, Shi W, et al. Embedding uncertain knowledge graphs [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 3363-3370. |
| [24] | Nayyeri M, Cil G M, Vahdati S, et al. Link prediction of weighted triples for knowledge graph completion within the scholarly domain[J]. IEEE Access, 2021, 9: 116002-116014. |
| [25] | Chen Z M, Yeh M Y, Kuo T W. PASSLEAF: a pool-based semi-supervised learning framework for uncertain knowledge graph embedding [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(5): 4019-4026. |
| [26] | Yang S, Zhang W, Tang R. Fast confidence prediction of uncertainty based on knowledge graph embedding [C]// ACAI 2020. New York: ACM, 2020: No.54. |
| [27] | Zhang J, Wu T, Qi G. Gaussian metric learning for few-shot uncertain knowledge graph completion [C]// DASFAA 2021, LNCS 12681. Cham: Springer, 2021: 256-271. |
| [28] | Wang J, Wu T, Zhang J. Incorporating uncertainty of entities and relations into few-shot uncertain knowledge graph embedding [C]// CCKS 2022. Singapore: Springer, 2022: 16-28. |
| [29] | 徐遥,何世柱,刘康,等. 基于Beta分布和半监督学习的非确定性知识图谱嵌入模型[J]. 中文信息学报, 2022, 36(10): 54-62. |
| Xu Yao, He Shizhu, Liu Kang, et al. Uncertain knowledge graph embedding by Beta distribution and semi-supervised learning[J]. Journal of Chinese Information Processing, 2022, 36(10): 54-62. | |
| [30] | Ji S, Zhang Z, Ying S, et al. Kullback-Leibler divergence metric learning [J]. IEEE Transactions on Cybernetics, 2022, 52(4): 2047-2058. |
| [31] | Patel D, Dasgupta S S, Boratko M, et al. Representing joint hierarchies with box embeddings[EB/OL]. (2020) [2025-05-10].. |
| [32] | Chen X, Boratko M, Chen M, et al. Probabilistic box embeddings for uncertain knowledge graph reasoning [C]// NAACL-HLT 2021. Stroudsburg: ACL, 2021: 882-893. |
| [33] | Boutouhami K, Zhang J, Qi G, et al. Uncertain ontology-aware knowledge graph embeddings [C]// JIST 2019, CCIS 1157. Singapore: Springer, 2020: 129-136. |
| [34] | Kertkeidkachorn N, Liu X, Ichise R. GTransE: generalizing translation-based model on uncertain knowledge graph embedding[C]// JSAI 2019, AISC 1128. Cham: Springer, 2020: 170-178. |
| [35] | Liu Q, Zhang Q, Zhao F, et al. Uncertain knowledge graph embedding: an effective method combining multi-relation and multi-path [J]. Frontiers of Computer Science, 2024, 18(3): No.183311. |
| [36] | Yang S, Zhang W, Tang R, et al. Approximate inferring with confidence predicting based on uncertain knowledge graph embedding [J]. Information Sciences, 2022, 609: 679-690. |
| [37] | Mikolov T, Chen K, Corrado G, et al. Efficient estimation of word representations in vector space[PP/OL]. V3. arXiv (2013-09-07) [2025-05-24].. |
| [38] | Graves A. Long short-term memory [M]// Supervised sequence labelling with recurrent neural networks, SCI 385. Berlin: Springer, 2012: 37-45. |
| [39] | Yang S, Tang R. Learning knowledge uncertainty from the pretrained language model [C]// ICISE 2021. New York: ACM, 2021: 37-42. |
| [40] | Devlin J, Chang M W, Lee K, et al. BERT: pre-training of deep bidirectional Transformers for language understanding [C]// NAACL-HLT 2019, Volume 1 (Long and Short Papers). Stroudsburg: ACL, 2019: 4171-4186. |
| [41] | Bahaj A, Ghogho M. Uncertain knowledge graph embedding using auxiliary information [J]. IEEE Access, 2024, 12: 138351-138361. |
| [42] | Shen Q, Qu A. CosUKG: a representation learning framework for uncertain knowledge graphs [J]. Mathematics, 2024, 12(10): No.1419. |
| [1] | 刘威, 李维刚, 田志强. 面向点云分类与分割的层次化旋转不变几何结构的表征学习方法[J]. 《计算机应用》唯一官方网站, 2026, 46(8): 2594-2602. |
| [2] | 高健, 李智, 樊缤, 姜传贤. 基于光线投射采样和四元数正交矩的高效三维医学影像鲁棒零水印算法[J]. 《计算机应用》唯一官方网站, 2023, 43(4): 1191-1197. |
| [3] | 赵云涛, 谢万琪, 李维刚, 胡佳明. 基于协方差矩阵自适应进化策略的机器人手眼标定算法[J]. 《计算机应用》唯一官方网站, 2023, 43(10): 3225-3229. |
| [4] | 张剑, 程培源, 邵思羽. 基于改进残差卷积自编码网络的类自适应旋转机械故障诊断[J]. 《计算机应用》唯一官方网站, 2022, 42(8): 2440-2449. |
| [5] | 薛远亮, 金国栋, 谭力宁, 许剑锟. 基于像素分类的多尺度无人机航拍目标旋转跟踪算法[J]. 《计算机应用》唯一官方网站, 2022, 42(7): 2239-2247. |
| [6] | 许盛伟, 康婕. 基于逻辑单粒子的多方量子密钥协商协议[J]. 《计算机应用》唯一官方网站, 2022, 42(1): 157-161. |
| [7] | 宋中山, 梁家锐, 郑禄, 刘振宇, 帖军. 基于双向门控尺度特征融合的遥感场景分类[J]. 计算机应用, 2021, 41(9): 2726-2735. |
| [8] | 陈俊, 何庆. 基于余弦相似度的改进蝴蝶优化算法[J]. 计算机应用, 2021, 41(9): 2668-2677. |
| [9] | 姬张建, 任兴旺. 带旋转与尺度估计的全卷积孪生网络目标跟踪算法[J]. 计算机应用, 2021, 41(9): 2705-2711. |
| [10] | 刘望华, 刘光帅, 陈晓文, 李旭瑞. 结合微分特征和Haar小波分解的鲁棒纹理表达[J]. 计算机应用, 2020, 40(9): 2728-2736. |
| [11] | 曹小威, 曲志坚, 徐玲玲, 刘晓红. 基于胶囊网络的三维模型识别[J]. 计算机应用, 2020, 40(5): 1309-1314. |
| [12] | 周翔, 唐丽玉, 林定. 基于二进制鲁棒不变尺度关键点-加速稳健特征的自然特征虚实注册方法[J]. 计算机应用, 2020, 40(5): 1403-1408. |
| [13] | 石志良, 蔡旺月, 汪国强, 熊林杰. 基于自适应邻域的固有形状特征算法[J]. 计算机应用, 2020, 40(4): 1151-1156. |
| [14] | 蔡芸, 刘朋青, 熊禾根. 基于量子遗传混合算法的泊位联合调度[J]. 计算机应用, 2020, 40(3): 897-901. |
| [15] | 李为, 李为相, 张璠, 揭伟. 基于运动平滑约束项的快速误匹配剔除算法[J]. 计算机应用, 2018, 38(9): 2678-2682. |
| 阅读次数 | ||||||
|
全文 |
|
|||||
|
摘要 |
|
|||||