Know-Evolve: Deep Temporal Reasoning for Dynamic Knowledge Graphs

By Rakshit S. Trivedi, Hanjun Dai, Yichen Wang, Le Song. 2017

583 citations recorded by the source corpus.

The availability of large scale event data with time stamps has given rise to dynamically evolving knowledge graphs that contain temporal information for each edge. Reasoning over time in such dynamic knowledge graphs is not yet well understood. To this end, we present Know-Evolve, a novel deep evolutionary knowledge network that learns non-linearly evolving entity representations over time. The occurrence of a fact (edge) is modeled as a multivariate point process whose intensity function is modulated by the score for that fact computed based on the learned entity embeddings. We demonstrate significantly improved performance over various relational learning approaches on two large scale real-world datasets. Further, our method effectively predicts occurrence or recurrence time of a fact which is novel compared to prior reasoning approaches in multi-relational setting.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Learning to Simulate Complex Physics with Graph Networks (2020, background): “…et al., 2019; Sanchez-Gonzalez et al., 2019), as well as non-physical systems, such as multi-agent dynamics (Tacchetti et al., 2018; Sun et al., 2019), algorithm execution (Velikovi et al., 2020), and other dynamic graph settings (Trivedi et al., 2019; 2017; Yan et al., 2018; Manessi et al., 2020).”
Recurrent Event Network for Reasoning over Temporal Knowledge Graphs (2019, background): “The intensity function (equation 3 in [22]) is deﬁned as λ s,rr ( t | ¯ t ) = f ( g s,rr (¯ t ))( t − ¯ t ) , where g ( · ) is a score function, t is current time, and ¯ t is the most recent time point when either subject or object entity was involved in an event.”
Learning from History: Modeling Temporal Knowledge Graphs with Sequential Copy-Generation Networks (2020, background): “2016), HyTE (Dasgupta, Ray, and Talukdar 2018), TADistMult (Garcı́a-Durán, Dumancic, and Niepert 2018), Know-Evolve+MLP (Trivedi et al. 2017), DyRep+MLP (Trivedi et al.”
Foundations and Modeling of Dynamic Networks Using Dynamic Graph Neural Networks: A Survey (2020, methodology): “Similarly to Know-Evolve [89], it targets temporal knowledge networks.”
Knowledge Graph Embedding: A Survey of Approaches and Applications (2017, background): “[117] recently proposed to learn non-linearly evolving entity representations over time so as to perform temporal reasoning over dynamic KGs.”
Relational Representation Learning for Dynamic (Knowledge) Graphs: A Survey (2019, methodology): “Recently there has been growing interest towards time predicting decoders [217, 216, 261].”
Incremental Training of Graph Neural Networks on Temporal Graphs under Distribution Shift (2020, background): “Know-Evolve [26] relies on vertex embeddings and multivariate point processes.”
A Comprehensive Survey on Automatic Knowledge Graph Construction (2023, background): “For example, Know-evolve [340] involves a multivariate temporal point process with an enhanced RNN structure that learns a temporal evolutionary representation function.”
Temporal Knowledge Graph Reasoning Based on Evolutional Representation Learning (2021, methodology): “For Know-evolve and DyRep, RE-NET extends them to the temporal reasoning task but does not release their codes.”
Temporal Graph Networks for Deep Learning on Dynamic Graphs (2020, methodology): “Sequence-based approaches for CTDGs (Kumar et al., 2019; Trivedi et al., 2017; 2019; Ma et al., 2018) use RNNs to update representations of the source and destination node each time a new edge appears.”
A Survey on Knowledge Graphs: Representation, Acquisition, and Applications (2020, background): “Know-evolve [141], a deep evolutionary knowledge network, investigates the knowledge evolution phenomenon of entities and their evolved relations.”
DySAT: Deep Neural Representation Learning on Dynamic Graphs via Self-Attention Networks (2020, background): “Dynamic graphs are defined in two common ways: snapshot sequence [16], which is a collection of evolving graph snapshots at multiple discrete time steps; and timestamped graph [31], which is a single graph with continuousvalued timestamped links.”
Learning Steady-States of Iterative Algorithms over Graphs (2018, background): “, 2017) and knowledge base reasoning (Trivedi et al., 2017).”
Recurrent Event Network: Autoregressive Structure Inference over Temporal Knowledge Graphs (2019, methodology): “GDELT ICEWS14 WIKI YAGO MRR H@3 H@10 MRR H@3 H@10 MRR H@3 H@10 MRR H@3 H@10 MRR H@3 H@10”
Answering Visual-Relational Queries in Web-Extracted Knowledge Graphs (2017, background): “, 2017], temporal graph dynamics [Trivedi et al., 2017], and multiple modalities [Pezeshkpour et al.”
Time2Vec: Learning a Vector Representation of Time (2019, background): “There is also a set of techniques that consider time as yet another feature and concatenate time (or some hand designed features of time such as log and/or inverse of delta time) with the input [10, 33, 14, 3, 29, 55, 28, 38].”
Adversarial Attack on Graph Structured Data (2018, background): “Representation learning on the structured data with deep learning methods has shown promising results in various applications, including drug screening (Duvenaud et al., 2015), protein analysis (Hamilton et al., 2017), knowledge graph completion (Trivedi et al., 2017), etc. .”
Diachronic Embedding for Temporal Knowledge Graph Completion (2019, methodology): “For GDELT, we use the subset extracted by (Trivedi et al. 2017) corresponding to the facts from April 1, 2015 to March 31, 2016.”