Deep Relational Machines

Lodhi, Huma

doi:10.1007/978-3-642-42042-9_27

Cited by 24 publications

(28 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Incorporating domain-knowledge into deep neural networks has shown considerable success over the years. Possibly the earliest approach is propositional-isation [31,32] leading to construction of deep relational machines [33,34] where the deep network is a multi-layered perceptron. Recent studies on domainknowledge inclusion includes: vertex-enrichment [35], bottom-graph construction via [4].…”

Section: Related Workmentioning

confidence: 99%

“…It is a technique to transform a relational representation into a propositional single-table representation where each column in the table corresponds a feature that represents a relation constructed from data and domain-knowledge. Propositionalisation is the core technique in construction of deep relational machines [34,35,36]: these are multi-layered perceptrons constructed from propositionalised representation of relational data and domain-knowledge. Recent studies on domain-knowledge inclusion include construction of graph neural networks (GNNs) that can learn not only from relational (graph-structured) data but also symbolic domain-knowledge.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Using Domain-Knowledge to Assist Lead Discovery in Early-Stage Drug Design

Dash

Srinivasan

Vig³

et al. 2021

Preprint

View full text Add to dashboard Cite

We are interested in generating new small molecules which could act as inhibitors of a biological target, when there is limited prior information on target-specific inhibitors. This form of drug-design is assuming increasing importance with the advent of new disease threats for which known chemicals only provide limited information about target inhibition. In this paper, we propose the combined use of deep neural networks and Inductive Logic Programming (ILP) that allows the use of symbolic domain-knowledge (B) to explore the large space of possible molecules. Assuming molecules and their activities to be instances of random variables X and Y, the problem is to draw instances from the conditional distribution of X, given Y, B (DX|Y,B). We decompose this into the constituent parts of obtaining the distributions DX|B and DY |X,B, and describe the design and implementation of models to approximate the distributions. The design consists of generators (to approximate DX|B and DX|Y,B) and a discriminator (to approximate DY |X,B). We investigate our approach using the well-studied problem of inhibitors for the Janus kinase (JAK) class of proteins. We assume first that if no data on inhibitors are available for a target protein (JAK2), but a small numbers of inhibitors are known for homologous proteins (JAK1, JAK3 and TYK2). We show that the inclusion of relational domainknowledge results in a potentially more effective generator of inhibitors than simple random sampling from the space of molecules or a generator without access to symbolic relations. The results suggest a way of combining symbolic domain-knowledge and deep generative models to constrain the exploration of the chemical space of molecules, when there is limited information on target-inhibitors. We also show how samples from the conditional generator can be used to identify potentially novel target inhibitors.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Using Domain-Knowledge to Assist Lead Discovery in Early-Stage Drug Design

Dash

Srinivasan

Vig³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Deep neural networks are effective learners in numeric space, capable of constructing intermediate knowledge constructs and thereby improve semantics of baseline input representation. Training deep neural networks on propositionalized relational data were explored by Srinivasan et al (2019), following the work of Lodhi (2013), where Deep Relational Machines (DRMs) were first introduced. In Lodhi's work, the DRMs used bodies of first order Horn clauses as input to restricted Boltzmann machines, where conjuncts of bonds and other molecular structure information compose individual complex features; when all structural properties are present in a given instance, the target's value is true, and false otherwise.…”

Section: Deep Relational Machinesmentioning

confidence: 99%

“…Note that the propositionalized data set P is usually a sparse matrix, which can represent additional challenge for neural networks. The DRMs proposed by Lodhi (2013) were used for prediction of protein folding properties, as well as mutagenicity assessment of small molecules. This approach used feature selection with information theoretic measures such as information gain as the sparse matrix resulting from the propositionalization was not suitable as an input to the neural network.…”

Section: Deep Relational Machinesmentioning

confidence: 99%

“…Section 5 presents a unifying framework for propositionalization and embeddings, allowing for the analysis of characteristic properties of these data transformation approaches. Section 6 proposes a unifying methodology that combines propositionalization and embeddings, which benefits from the advantages of both, and presents two implementations of the proposed unifying framework: an instance-based embedding approach PropDRM based on the existing Deep Relational Machines (DRM) (Srinivasan et al 2019;Lodhi 2013), followed by a novel feature-based embedding approach PropStar proposed in this paper, using the StarSpace entity embedding approach (Wu et al 2018). Experimental evaluation of the proposed implementations is presented in Sect.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Propositionalization and embeddings: two sides of the same coin

2020

View full text Add to dashboard Cite

Data preprocessing is an important component of machine learning pipelines, which requires ample time and resources. An integral part of preprocessing is data transformation into the format required by a given learning algorithm. This paper outlines some of the modern data processing techniques used in relational learning that enable data fusion from different input data types and formats into a single table data representation, focusing on the propositionalization and embedding data transformation approaches. While both approaches aim at transforming data into tabular data format, they use different terminology and task definitions, are perceived to address different goals, and are used in different contexts. This paper contributes a unifying framework that allows for improved understanding of these two data transformation techniques by presenting their unified definitions, and by explaining the similarities and differences between the two approaches as variants of a unified complex data transformation task. In addition to the unifying framework, the novelty of this paper is a unifying methodology combining propositionalization and embeddings, which benefits from the advantages of both in solving complex data transformation and learning tasks. We present two efficient implementations of the unifying methodology: an instance-based PropDRM approach, and a feature-based PropStar approach to data transformation and learning, together with their empirical evaluation on several relational problems. The results show that the new algorithms can outperform existing relational learners and can solve much larger problems.

show abstract

Deep node ranking for neuro‐symbolic structural node embedding and classification

et al. 2021

View full text Add to dashboard Cite

Network node embedding is an active research subfield of complex network analysis. This paper contributes a novel approach to learning network node embeddings and direct node classification using a node ranking scheme, coupled with an autoencoder-based neural network architecture. The main advantages of the proposed Deep Node Ranking (DNR) algorithm are competitive or better classification performance, significantly higher learning speed and lower space requirements when compared to state-of-the-art approaches on 15 real-life structural node classification benchmarks. It also enables exploration of the relationship between symbolic and the derived sub-symbolic node representations, offering insights into the learned node space structure. To avoid the space complexity bottleneck in a direct node classification setting, DNR, if needed, computes stationary distributions of personalized random walks from given nodes in mini-batches, scaling seamlessly to larger networks. The scaling laws associated with DNR were also investigated by considering 1,488 synthetic Erdős-Rényi networks, demonstrating its scalability to tens of millions of links.

show abstract

Deep Relational Machines

Cited by 24 publications

References 11 publications

Using Domain-Knowledge to Assist Lead Discovery in Early-Stage Drug Design

Using Domain-Knowledge to Assist Lead Discovery in Early-Stage Drug Design

Propositionalization and embeddings: two sides of the same coin

Deep node ranking for neuro‐symbolic structural node embedding and classification

Contact Info

Product

Resources

About