MOTIVATION: Knowledge Graph Completion has been increasingly adopted as a useful method for helping address several tasks in biomedical research, such as drug repurposing or drug-target identification. To that end, a variety of datasets and Knowledge Graph Embedding models have been proposed over the years. However, little is known about the properties that render a dataset, and associated modelling choices, useful for a given task. Moreover, even though theoretical properties of Knowledge Graph Embedding models are well understood, their practical utility in this field remains controversial. RESULTS: In this work, we conduct a comprehensive investigation into the topological properties of publicly available biomedical Knowledge Graphs and establish links to the accuracy observed in real-world tasks. By releasing all model predictions and a new suite of analysis tools we invite the community to build upon our work and continue improving the understanding of these crucial applications. AVAILABILITY AND IMPLEMENTATION: The code used to perform experiments and analyze results in this article as well as all experimental data is available at https://github.com/graphcore-research/kg-topology-toolbox/tree/main/the_role_of_graph_topology_paper and archived on Zenodo, at https://doi.org/10.5281/zenodo.12097376.

In recent years, numerous machine learning models which attempt to solve polypharmacy side effect identification, drug-drug interaction prediction, and combination therapy design tasks have been proposed. Here, we present a unified theoretical view of relational machine learning models which can address these tasks. We provide fundamental definitions, compare existing model architectures and discuss performance metrics, datasets, and evaluation protocols. In addition, we emphasize possible high-impact applications and important future research directions in this domain.

Drug discovery and development is a complex and costly process. Machine learning approaches are being investigated to help improve the effectiveness and speed of multiple stages of the drug discovery pipeline. Of these, those that use Knowledge Graphs (KG) have promise in many tasks, including drug repurposing, drug toxicity prediction and target gene-disease prioritization. In a drug discovery KG, crucial elements including genes, diseases and drugs are represented as entities, while relationships between them indicate an interaction. However, to construct high-quality KGs, suitable data are required. In this review, we detail publicly available sources suitable for use in constructing drug discovery focused KGs. We aim to help guide machine learning and KG practitioners who are interested in applying new techniques to the drug discovery field, but who may be unfamiliar with the relevant data sources. The datasets are selected via strict criteria, categorized according to the primary type of information contained within and are considered based upon what information could be extracted to build a KG. We then present a comparative analysis of existing public drug discovery KGs and an evaluation of selected motivating case studies from the literature. Additionally, we raise numerous and unique challenges and issues associated with the domain and its datasets, while also highlighting key future research directions. We hope this review will motivate KGs use in solving key and emerging questions in the drug discovery domain.

Adoption of recently developed methods from machine learning has given rise to creation of drug-discovery knowledge graphs (KG) that utilize the interconnected nature of the domain. Graph-based modelling of the data, combined with KG embedding (KGE) methods, are promising as they provide a more intuitive representation and are suitable for inference tasks such as predicting missing links. One common application is to produce ranked lists of genes for a given disease, where the rank is based on the perceived likelihood of association between the gene and the disease. It is thus critical that these predictions are not only pertinent but also biologically meaningful. However, KGs can be biased either directly due to the underlying data sources that are integrated or due to modeling choices in the construction of the graph, one consequence of which is that certain entities can get topologically overrepresented. We demonstrate the effect of these inherent structural imbalances, resulting in densely-connected entities being highly ranked no matter the context. We provide support for this observation across different datasets, models as well as predictive tasks. Further, we present various graph perturbation experiments which yield more support to the observation that KGE models can be more influenced by the frequency of entities rather than any biological information encoded within the relations. Our results highlight the importance of data modeling choices, and emphasizes the need for practitioners to be mindful of these issues when interpreting model outputs and during KG composition.

The drug discovery and development process is a long and expensive one, costing over 1 billion USD on average per drug and taking 10-15 years. To reduce the high levels of attrition throughout the process, there has been a growing interest in applying machine learning methodologies to various stages of drug discovery and development in the recent decade, especially at the earliest stage - identification of druggable disease genes. In this paper, we have developed a new tensor factorisation model to predict potential drug targets (genes or proteins) for treating diseases. We created a three-dimensional data tensor consisting of 1,048 gene targets, 860 diseases and 230,011 evidence attributes and clinical outcomes connecting them, using data extracted from the Open Targets and PharmaProjects databases. We enriched the data with gene target representations learned from a drug discovery-oriented knowledge graph and applied our proposed method to predict the clinical outcomes for unseen gene target and disease pairs. We designed three evaluation strategies to measure the prediction performance and benchmarked several commonly used machine learning classifiers together with Bayesian matrix and tensor factorisation methods. The result shows that incorporating knowledge graph embeddings significantly improves the prediction accuracy and that training tensor factorisation alongside a dense neural network outperforms all other baselines. In summary, our framework combines two actively studied machine learning approaches to disease target identification, namely tensor factorisation and knowledge graph representation learning, which could be a promising avenue for further exploration in data-driven drug discovery.

Text classification has long been a staple within Natural Language Processing (NLP) with applications spanning across diverse areas such as sentiment analysis, recommender systems and spam detection. With such a powerful solution, it is often tempting to use it as the go-to tool for all NLP problems since when you are holding a hammer, everything looks like a nail. However, we argue here that many tasks which are currently addressed using classification are in fact being shoehorned into a classification mould and that if we instead address them as a ranking problem, we not only improve the model, but we achieve better performance. We propose a novel end-to-end ranking approach consisting of a Transformer network responsible for producing representations for a pair of text sequences, which are in turn passed into a context aggregating network outputting ranking scores used to determine an ordering to the sequences based on some notion of relevance. We perform numerous experiments on publicly-available datasets and investigate the applications of ranking in problems often solved using classification. In an experiment on a heavily- skewed sentiment analysis dataset, converting ranking results to classification labels yields an approximately 22% improvement over state-of-the-art text classification, demonstrating the efficacy of text ranking over text classification in certain scenarios.

We show that correlations between the camera used to acquire an image and the class label of that image can be exploited by convolutional neural networks (CNN), resulting in a model that “cheats” at an image classification task by recognizing which camera took the image and inferring the class label from the camera. We show that models trained on a dataset with camera / label correlations do not generalize well to images in which those correlations are absent, nor to images from unencountered cameras. Furthermore, we investigate which visual features they are exploiting for camera recognition. Our experiments present evidence against the importance of global color statistics, lens deformation and chromatic aberration, and in favor of high frequency features, which may be introduced by image processing algorithms built into the cameras.

With the growing significance of graphs as an effective representation of data in numerous applications, efficient graph analysis using modern machine learning is receiving a growing level of attention. Deep learning approaches often operate over the entire adjacency matrix - as the input and intermediate network layers are all designed in proportion to the size of the adjacency matrix - leading to intensive computation and large memory requirements as the graph size increases. It is therefore desirable to identify efficient measures to reduce both run-time and memory requirements allowing for the analysis of the largest graphs possible. The use of reduced precision operations within the forward and backward passes of a deep neural network along with novel specialised hardware in modern GPUs could offer promising avenues towards efficiency. In this paper, we provide an in-depth exploration of the use of reduced-precision operations, easily integrable into the highly popular PyTorch framework, and an analysis of the effects of Tensor Cores on graph convolutional neural networks. We perform an extensive experimental evaluation of three GPU architectures and two widely-used graph analysis tasks (vertex classification and link prediction) using well-known benchmark and synthetically generated datasets. Thus allowing us to make important observations on the effects of reduced-precision operations and Tensor Cores on computational and memory usage of graph convolutional neural networks - often neglected in the literature.

Recently, substantial progress has been made in the area of Brain-Computer Interface (BCI) using modern machine learning techniques to decode and interpret brain signals. While Electroencephalography (EEG) has provided a non-invasive method of interfacing with a human brain, the acquired data is often heavily subject and session dependent. This makes the seamless incorporation of such data into realworld applications intractable as the subject and session data variance can lead to long and tedious calibration requirements and cross-subject generalisation issues. Focusing on a Steady State Visual Evoked Potential (SSVEP) classification systems, we propose a novel means of generating highly-realistic synthetic EEG data invariant to any subject, session or other environmental conditions. Our approach, entitled the Subject Invariant SSVEP Generative Adversarial Network (SIS-GAN), produces synthetic EEG data from multiple SSVEP classes using a single network. Additionally, by taking advantage of a fixed-weight pre-trained subject classification network, we ensure that our generative model remains agnostic to subject-specific features and thus produces subject-invariant data that can be applied to new previously unseen subjects. Our extensive experimental evaluation demonstrates the efficacy of our synthetic data, leading to superior performance, with improvements of up to 16 percentage points in zero-calibration classification tasks when trained using our subject-invariant synthetic EEG signals.

A common task for recommender systems is to build a profile of the interests of a user from items in their browsing history and later to recommend items to the user from the same catalog. The users' behavior consists of two parts: the sequence of items that they viewed without intervention (the organic part) and the sequences of items recommended to them and their outcome (the bandit part).

Newly emerging variants of ransomware pose an ever-growing threat to computer systems governing every aspect of modern life through the handling and analysis of big data. While various recent security-based approaches have focused on ransomware detection at the network or system level, easy-to-use post-infection ransomware classification for the lay user has not been attempted before. In this paper, we investigate the possibility of classifying the ransomware a system is infected with simply based on a screenshot of the splash screen or the ransom note captured using a consumer camera commonly found in any modern mobile device. To train and evaluate our system, we create a sample dataset of the splash screens of 50 well-known ransomware variants. In our dataset, only a single training image is available per ransomware. Instead of creating a large training dataset of ransomware screenshots, we simulate screenshot capture conditions via carefully-designed data augmentation techniques, enabling simple and efficient one-shot learning. Moreover, using model uncertainty obtained via Bayesian approximation, we ensure special input cases such as unrelated non-ransomware images and previously-unseen ransomware variants are correctly identified for special handling and not mis-classified. Extensive experimental evaluation demonstrates the efficacy of our work, with accuracy levels of up to 93.6% for ransomware classification.

With the recent growth in the number of malicious activities on the internet, cybersecurity research has seen a boost in the past few years. However, as certain variants of malware can provide highly lucrative opportunities for bad actors, significant resources are dedicated to innovations and improvements by vast criminal organisations. Among these forms of malware, ransomware has experienced a significant recent rise as it offers the perpetrators great financial incentive. Ransomware variants operate by removing system access from the user by either locking the system or encrypting some or all of the data, and subsequently demanding payment or ransom in exchange for returning system access or providing a decryption key to the victim. Due to the ubiquity of sensitive data in many aspects of modern life, many victims of such attacks, be they an individual home user or operators of a business, are forced to pay the ransom to regain access to their data, which in many cases does not happen as renormalisation of system operations is never guaranteed. As the problem of ransomware does not seem to be subsiding, it is very important to investigate the underlying forces driving and facilitating such attacks in order to create preventative measures. As such, in this paper, we discuss and provide further insight into variants of ransomware and their victims in order to understand how and why they have been targeted and what can be done to prevent or mitigate the effects of such attacks.

Graphs have become a crucial way to represent large, complex and often temporal datasets across a wide range of scientific disciplines. However, when graphs are used as input to machine learning models, this rich temporal information is frequently disregarded during the learning process, resulting in suboptimal performance on certain temporal inference tasks. To combat this, we introduce Temporal Neighbourhood Aggregation (TNA), a novel vertex representation model architecture designed to capture both topological and temporal information to directly predict future graph states. Our model exploits hierarchical recurrence at different depths within the graph to enable exploration of changes in temporal neighbourhoods, whilst requiring no additional features or labels to be present. The final vertex representations are created using variational sampling and are optimised to directly predict the next graph in the sequence. Our claims are supported by experimental evaluation on both real and synthetic benchmark datasets, where our approach demonstrates superior performance compared to competing methods, outperforming them at predicting new temporal edges by as much as 23% on real-world datasets, whilst also requiring fewer overall model parameters.

Deep Convolutional Neural Networks (CNN) have been extensively applied in various computer vision tasks. Although such approaches have demonstrated exceptionally high performance in various open challenges, adapting them to more specialised tasks can be non-trivial. In this paper we discuss our design and implementation of a batchcode detection system capable of accurate segmentation of batchcode regions within images of consumer products. A batchcode is a unique identifier printed on the packaging of many products that encodes useful information such as date and location of manufacture. Detection of batchcodes in images of products is a useful step in many processes, including quality control, supply chain tracking and counterfeit detection. Beginning with a unique dataset of product images and a set of crowdsourced coarse annotations that roughly correspond to the locations of batchcodes, we demonstrate that such annotations are insufficient for training a reliable model, and subsequently describe a novel label refinement process, which we call the Maximally Stable Global Region (MSGR) method, that we use to generate accurate ground-truth data suitable for training a robust neural network. We also show that detection accuracy can be further improved by applying MSGR to the output of the neural network. We evaluate our approach using a manually labelled test dataset of images of shampoo bottles, and demonstrate the efficacy of the proposed method for accurate real-time batchcode detection.

Recommendations are commonly used to modify user’s natural behavior, for example, increasing product sales or the time spent on a website. This results in a gap between the ultimate business ob- jective and the classical setup where recommenda- tions are optimized to be coherent with past user be- havior. To bridge this gap, we propose a new learn- ing setup for recommendation that optimizes for the Incremental Treatment Effect (ITE) of the policy. We show this is equivalent to learning to predict recommendation outcomes under a fully random recommendation policy and propose a new domain adaptation algorithm that learns from logged data containing outcomes from a biased recommenda- tion policy and predicts recommendation outcomes according to random exposure. We compare our method against state-of-the-art factorization meth- ods, in addition to new approaches of causal rec- ommendation and show significant improvements.

Despite significant recent progress in the area of Brain-Computer Interface (BCI), there are numerous shortcomings associated with collecting Electroencephalography (EEG) signals in real-world environments. These include, but are not limited to, subject and session data variance, long and arduous calibration processes and predictive generalisation issues across different subjects or sessions. This implies that many downstream applications, including Steady State Visual Evoked Potential (SSVEP) based classification systems, can suffer from a shortage of reliable data. Generating meaningful and realistic synthetic data can therefore be of significant value in circumventing this problem. We explore the use of modern neural-based generative models trained on a limited quantity of EEG data collected from different subjects to generate supplementary synthetic EEG signal vectors, subsequently utilised to train an SSVEP classifier. Extensive experimental analysis demonstrates the efficacy of our generated data, leading to improvements across a variety of evaluations, with the crucial task of cross-subject generalisation improving by over 35% with the use of such synthetic data.

Graph embeddings have become a key and widely used technique within the field of graph mining, proving to be successful across a broad range of domains including social, citation, transportation and biological. Unsupervised graph embedding techniques aim to automatically create a low-dimensional representation of a given graph, which captures key structural elements in the resulting embedding space. However, to date, there has been little work exploring exactly which topological structures are being learned in the embeddings, which could be a possible way to bring interpretability to the process. In this paper, we investigate if graph embeddings are approximating something analogous to traditional vertex-level graph features. If such a relationship can be found, it could be used to provide a theoretical insight into how graph embedding approaches function. We perform this investigation by predicting known topological features, using supervised and unsupervised methods, directly from the embedding space. If a mapping between the embeddings and topological features can be found, then we argue that the structural information encapsulated by the features is represented in the embedding space. To explore this, we present extensive experimental evaluation with five state-of-the-art unsupervised graph embedding techniques, across a range of empirical graph datasets, measuring a selection of topological features. We demonstrate that several topological features are indeed being approximated in the embedding space, allowing key insight into how graph embeddings create good representations.

Graphs are a commonly used construct for representing relationships between elements in complex high dimensional datasets. Many real-world phenomenon are dynamic in nature, meaning that any graph used to represent them is inherently temporal. However, many of the machine learning models designed to capture knowledge about the structure of these graphs ignore this rich temporal information when creating representations of the graph. This results in models which do not perform well when used to make predictions about the future state of the graph - especially when the delta between time stamps is not small. In this work, we explore a novel training procedure and an associated unsupervised model which creates graph representations optimised to predict the future state of the graph. We make use of graph convo-lutional neural networks to encode the graph into a latent representation, which we then use to train our temporal offset reconstruction method, inspired by auto-encoders, to predict a later time point - multiple time steps into the future. Using our method, we demonstrate superior performance for the task of future link prediction compared with none-temporal state-of-the-art baselines. We show our approach to be capable of outperforming non-temporal baselines by 38% on a real world dataset.

Deep learning is rapidly becoming a go-to tool for many artificial intelligence problems due to its ability to outperform other approaches and even humans at many problems. Despite its popularity we are still unable to accurately predict the time it will take to train a deep learning network to solve a given problem. This training time can be seen as the product of the training time per epoch and the number of epochs which need to be performed to reach the desired level of accuracy. Some work has been carried out to predict the training time for an epoch - most have been based around the assumption that the training time is linearly related to the number of floating point operations required. However, this relationship is not true and becomes exacerbated in cases where other activities start to dominate the execution time. Such as the time to load data from memory or loss of performance due to non-optimal parallel execution. In this work we propose an alternative approach in which we train a deep learning network to predict the execution time for parts of a deep learning network. Timings for these individual parts can then be combined to provide a prediction for the whole execution time. This has advantages over linear approaches as it can model more complex scenarios. But, also, it has the ability to predict execution times for scenarios unseen in the training data. Therefore, our approach can be used not only to infer the execution time for a batch, or entire epoch, but it can also support making a well-informed choice for the appropriate hardware and model.

Many current applications use recommendations in order to modify the natural user behavior, such as to increase the number of sales or the time spent on a website. This results in a gap between the final recommendation objective and the classical setup where recommendation candidates are evaluated by their coherence with past user behavior, by predicting either the missing entries in the user-item matrix, or the most likely next event. To bridge this gap, we optimize a recommendation policy for the task of increasing the desired outcome versus the organic user behavior. We show this is equivalent to learning to predict recommendation outcomes under a fully random recommendation policy. To this end, we propose a new domain adaptation algorithm that learns from logged data containing outcomes from a biased recommendation policy and predicts recommendation outcomes according to random exposure. We compare our method against state-of-the-art factorization methods, in addition to new approaches of causal recommendation and show significant improvements.

Electroencephalography (EEG) is a common signal acquisition approach employed for Brain-Computer Interface (BCI) research. Nevertheless, the majority of EEG acquisition devices rely on the cumbersome application of conductive gel (so-called wet-EEG) to ensure a high quality signal is obtained. However, this process is unpleasant for the experimental participants and thus limits the practical application of BCI. In this work, we explore the use of a commercially available dry-EEG headset to obtain visual cortical ensemble signals. Whilst improving the usability of EEG within the BCI context, dry-EEG suffers from inherently reduced signal quality due to the lack of conduit gel, making the classification of such signals significantly more challenging. In this paper, we propose a novel Convolutional Neural Network (CNN) approach for the classification of raw dry-EEG signals without any data pre-processing. To illustrate the effectiveness of our approach, we utilise the Steady State Visual Evoked Potential (SSVEP) paradigm as our use case. SSVEP can be utilised to allow people with severe physical disabilities such as Complete Locked-In Syndrome or Amyotrophic Lateral Sclerosis to be aided via BCI applications, as it requires only the subject to fixate upon the sensory stimuli of interest. Here we utilise SSVEP flicker frequencies between 10 to 30 Hz, which we record as subject cortical waveforms via the dry-EEG headset. Our proposed end-to-end CNN allows us to automatically and accurately classify SSVEP stimulation directly from the dry-EEG waveforms. Our CNN architecture utilises a common SSVEP Convolutional Unit (SCU), comprising of a 1D convolutional layer, batch normalization and max pooling. Furthermore, We compare several deep learning neural network variants with our primary CNN architecture, in addition to traditional machine learning classification approaches. Experimental evaluation shows our CNN architecture to be significantly better than competing approaches, achieving a classification accuracy of 96% whilst demonstrating superior cross-subject performance and even being able to generalise well to unseen subjects whose data is entirely absent from the training process.

In this paper we study three state-of-the-art, but competing, approaches for generating graph embeddings using unsupervised neural networks. Graph embeddings aim to discover the `best' representation for a graph automatically and have been applied to graphs from numerous domains, including social networks. We evaluate their effectiveness at capturing a good representation of a graph's topological structure by using the embeddings to predict a series of topological features at the vertex level. We hypothesise that an `ideal' high quality graph embedding should be able to capture key parts of the graph's topology, thus we should be able to use it to predict common measures of the topology, for example vertex centrality. This could also be used to better understand which topological structures are truly being captured by the embeddings. We first review these three graph embedding techniques and then evaluate how close they are to being `ideal'. We provide a framework, with extensive experimental evaluation on empirical and synthetic datasets, to assess the effectiveness of several approaches at creating graph embeddings which capture detailed topological structure.

The classification of graphs is a key challenge within many scientific fields using graphs to represent data and is an active area of research. Graph classification can be critical in identifying and labelling unknown graphs within a dataset and has seen application across many scientific fields. Graph classification poses two distinct problems: the classification of elements within a graph and the classification of the entire graph. Whilst there is considerable work on the first problem, the efficient and accurate classification of massive graphs into one or more classes has, thus far, received less attention. In this paper we propose the Deep Topology Classification (DTC) approach for global graph classification. DTC extracts both global and vertex level topological features from a graph to create a highly discriminate representation in feature space. A deep feed-forward neural network is designed and trained to classify these graph feature vectors. This approach is shown to be over 99% accurate at discerning graph classes over two datasets. Additionally, it is shown to be more accurate than current state of the art approaches both in binary and multi-class graph classification tasks.

The problem of how to compare empirical graphs is an area of great interest within the field of network science. The ability to accurately but efficiently compare graphs has a significant impact in such areas as temporal graph evolution, anomaly detection and protein comparison. The comparison problem is compounded when working with massive graphs containing millions of vertices and edges. This paper introduces a parallel feature extraction based approach for the efficient comparison of large unlabelled graph datasets using Apache Spark. The approach acts by producing a `Graph Fingerprint' which represents both vertex level and global level topological features from a graph. By using Spark we are able to efficiently compare graphs considered unmanageably large to other approaches. The runtime of the approach is shown to scale sub-linearly with the size and complexity of the graphs being fingerprinted. Importantly, the approach is shown to not only be comparable to existing approaches, but on when comparing topology and size, more sensitive at detecting variation between graphs.

Recent technological advances in modern healthcare have lead to the ability to collect a vast wealth of patient monitoring data. This data can be utilised for patient diagnosis but it also holds the potential for use within medical research. However, these datasets often contain errors which limit their value to medical research, with one study finding error rates ranging from 2.3%-26.9% in a selection of medical databases. Previous methods for automatically assessing data quality normally rely on threshold rules, which are often unable to correctly identify errors, as further complex domain knowledge is required. To combat this, a semantic web based framework has previously been developed to assess the quality of medical data. However, early work, based solely on traditional semantic web technologies, revealed they are either unable or inefficient at scaling to the vast volumes of medical data. In this paper we present a new method for storing and querying medical RDF datasets using Hadoop Map / Reduce. This approach exploits the inherent parallelism found within RDF datasets and queries, allowing us to scale with both dataset and system size. Unlike previous solutions, this framework uses highly optimised (SPARQL) joining strategies, intelligent data caching and the use of a super-query to enable the completion of eight distinct SPARQL lookups, comprising over eighty distinct joins, in only two Map / Reduce iterations. Results are presented comparing both the Jena and a previous Hadoop implementation demonstrating the superior performance of the new methodology. The new method is shown to be five times faster than Jena and twice as fast as the previous approach.