Joint representation learning over multi-sourced knowledge graphs (KGs) yields transferable and expressive embeddings that improve downstream tasks. Entity alignment (EA) is a critical step in this process. Despite recent considerable research progress in embedding-based EA, how it works remains to be explored. In this paper, we provide a similarity flooding perspective to explain existing translation-based and aggregation-based EA models. We prove that the embedding learning process of these models actually seeks a fixpoint of pairwise similarities between entities. We also provide experimental evidence to support our theoretical analysis. We propose two simple but effective methods inspired by the fixpoint computation in similarity flooding, and demonstrate their effectiveness on benchmark datasets. Our work bridges the gap between recent embedding-based models and the conventional similarity flooding algorithm. It would improve our understanding of and increase our faith in embedding-based EA.

In this paper, we present the "joint pre-training and local re-training'' framework for learning and applying multi-source knowledge graph (KG) embeddings. We are motivated by the fact that different KGs contain complementary information to improve KG embeddings and downstream tasks. We pre-train a large teacher KG embedding model over linked multi-source KGs and distill knowledge to train a student model for a task-specific KG. To enable knowledge transfer across different KGs, we use entity alignment to build a linked subgraph for connecting the pre-trained KGs and the target KG. The linked subgraph is re-trained for three-level knowledge distillation from the teacher to the student, i.e., feature knowledge distillation, network knowledge distillation, and prediction knowledge distillation, to generate more expressive embeddings. The teacher model can be reused for different target KGs and tasks without having to train from scratch. We conduct extensive experiments to demonstrate the effectiveness and efficiency of our framework.

Knowledge graphs (KGs) store rich facts about the real world. In this paper, we study KG alignment, which aims to find alignment between not only entities but also relations and classes in different KGs. Alignment at the entity level can cross-fertilize alignment at the schema level. We propose a new KG alignment approach, called DAAKG, based on deep learning and active learning. With deep learning, it learns the embeddings of entities, relations and classes, and jointly aligns them in a semi-supervised manner. With active learning, it estimates how likely an entity, relation or class pair can be inferred, and selects the best batch for human labeling. We design two approximation algorithms for efficient solution to batch selection. Our experiments on benchmark datasets show the superior accuracy and generalization of DAAKG and validate the effectiveness of all its modules.

Knowledge graphs (KGs) have become a valuable asset for many AI applications. Although some KGs contain plenty of facts, they are widely acknowledged as incomplete. To address this issue, many KG completion methods are proposed. Among them, open KG completion methods leverage the Web to find missing facts. However, noisy data collected from diverse sources may damage the completion accuracy. In this paper, we propose a new trustworthy method that exploits facts for a KG based on multi-sourced noisy data and existing facts in the KG. Specifically, we introduce a graph neural network with a holistic scoring function to judge the plausibility of facts with various value types. We design value alignment networks to resolve the heterogeneity between values and map them to entities even outside the KG. Furthermore, we present a truth inference model that incorporates data source qualities into the fact scoring function, and design a semi-supervised learning way to infer the truths from heterogeneous values. We conduct extensive experiments to compare our method with the state-of-the-arts. The results show that our method achieves superior accuracy not only in completing missing facts but also in discovering new facts.

Railway track is never perfect, as rail distortions, namely, geometric irregularities, exist at all locations along the track. However, these distortions can be regarded as valuable indicators for train localization, since track irregularities present location-dependent characteristics, the measurements of which using onboard sensors are repeatable for the same track. In this research, we study the possibility of determining a train’s position by matching the track irregularity measurements to a predefined map. A train-borne experiment on a real track is used to preliminarily demonstrate the feasibility, evaluate the performance and determine the key parameters for practical implementation. The results show that a submeter longitudinal localization accuracy can be achieved even when using a low-cost cabin-mounted microelectromechanical system (MEMS) inertial measurement unit (IMU), which measures the train’s responses to track irregularities. The proposed method can enhance the positioning accuracy and improve the robustness of multisensory train localization systems.

Nonholonomic constraint (NHC) and odometer speed have been proven to significantly improve the navigation accuracy of a global navigation satellite system (GNSS)-aided inertial navigation system (INS) for land vehicular applications. Exploiting the full potential of the NHC and odometer aids requires the inertial measurement unit (IMU) mounting angles, i.e., angular misalignment with respect to the host vehicle, to be precisely known. We address the accurate estimation of the IMU mounting angles through an aided dead reckoning (DR) approach. In this method, DR using the GNSS/INS integrated attitude and distance traveled is fused with the GNSS/INS integrated position through a straightforward Kalman filter. Simulation and field tests are carried out to validate the proposed algorithm for different grade IMUs, including typical navigation-grade, tactical-grade and low-cost IMUs. The results demonstrate that the pitch and heading mounting angles can be estimated with a comparable accuracy with the GNSS/INS attitude solution, for example, 0.001° accuracy can be achieved for a navigation-grade GNSS/INS integrated system. The roll mounting angle can not be estimated due to lack of observability in this approach, and the heading mounting angle estimation may be influenced by the GNSS/INS heading accuracy drift to some extent for the low-cost IMUs.