Radiologists make the diagnoses of bone fractures through examining X-ray radiographs and document them in radiology reports. Applying information extraction techniques on such radiology reports to retrieve the information of bone fracture diagnosis could yield a source of structured data for medical cohort studies, image labelling and decision support concerning bone fractures. In this study, we proposed an information extraction system of Bone X-ray radiology reports to retrieve the details of bone fracture detection and diagnosis, based on a bio-medically pre-trained Bidirectional Encoder Representations from Transformers (BERT) natural language processing (NLP) model by Google. The model, named as BoneBert, was first trained on annotations automatically generated by a handcrafted rule-based labelling system using a dataset of 6,048 X-ray radiology reports and then finetuned on a small set of 4,890 expert annotations. Thus, the model was trained in a "semi-supervised" fashion. We evaluated the performance of the proposed model and compared it with the conventional rule-based labelling system on two typical tasks: Assertion Classification (AC) for bone fracture status detection (positive, negative or uncertainty) and Named Entity Recognition (NER) related to the fracture type, the bone type and location of a fracture occurs. BoneBert outperformed the rulebased system in both tasks, showing great potential for automated information extraction of the detection and diagnosis of bone fracture from radiology reports, such as, the clinical status, type and location of bone fracture, and more related observations.

Component-level heterogeneous redundancy is gaining popularity as an approach for preventing single-point security breaches in Industrial Control Systems (ICSs), especially with regard to core components such as Programmable Logic Controllers (PLCs). To take control of a system with componentlevel heterogeneous redundancy, an adversary must uncover and concurrently exploit vulnerabilities across multiple versions of hardened components. As such, attackers incur increased costs and delays when seeking to launch a successful attack. Existing approaches advocate attack resilience via pairwise comparison among outputs from multiple PLCs. These approaches incur increased resource costs due to them having a high degree of redundancy and do not address concurrent attacks. In this paper we address both issues, demonstrating a data-driven component selection approach that achieves a trade-off between resources cost and security. In particular, we propose (i) a novel dual-PLC ICS architecture with native pairwise comparison which can offer limited yet comparable defence against single-point breaches, (ii) a machine-learning based selection mechanisms which can deliver resilience against non-concurrent attacks under resource constraints, (iii) a scaled up variant of the proposed architecture to counteract concurrent attacks with modest resource implications.