Multi-target prediction for dummies using two-branch neural networks

Abstract

Multi-target prediction (MTP) serves as an umbrella term for machine learning tasks that concern the simultaneous prediction of multiple target variables. Classical instantiations are multi-label classification, multivariate regression, multi-task learning, dyadic prediction, zero-shot learning, network inference, and matrix completion. Despite the significant similarities, all these domains have evolved separately into distinct research areas over the last two decades. This led to the development of a plethora of highly-engineered methods, and created a substantially-high entrance barrier for machine learning practitioners that are not experts in the field. In this work we present a generic deep learning methodology that can be used for a wide range of multi-target prediction problems. We introduce a flexible multi-branch neural network architecture, partially configured via a questionnaire that helps end users to select a suitable MTP problem setting for their needs. Experimental results for a wide range of domains illustrate that the proposed methodology manifests a competitive performance compared to methods from specific MTP domains.

Prediction of pipe failures in water supply networks for longer time periods through multi-label classification

“…Actually, specialised libraries have recently emerged with new functionalities as multi-label stratification or data set management (Szymanski & Kajdanowicz, 2019). The applicability and usefulness of these methods strongly depends on the form and type of the targets to be predicted, as well as the processing of the data (Iliadis, De Baets, & Waegeman, 2022). For this reason, we have explored different data processing strategies to improve the results.…”

An antimicrobial drug recommender system using MALDI-TOF MS and dual-branch neural networks

“…It ensures the dot products to be of manageable magnitudes, even for large values of h . This score can be used together with the sigmoid function and the cross-entropy loss to optimize the two-branch neural network to map a spectrum-drug pair to a resistance label (Iliadis et al, 2022). An overview of the model is visualized in Figure 1.…”

“…The idea of processing and combining two separate streams of information with two neural networks is applied in many fields of machine learning, collectively referred to as deep multi-target prediction (Waegeman et al, 2019; Iliadis et al, 2022).…”

Hyperparameter optimisation in differential evolution using Summed Local Difference Strings, a rugged but easily calculated landscape for combinatorial search problems

“…In [44], a multi-label classification and complex regression problem has been addressed for auto-parameter tuning in deep learning models. In [45], ANNs have been used to predict the parameters for DE using 24 test problems from a Black-Box Optimisation Benchmarking dataset.…”