A large number of fibroblast foci (FF) predict mortality in idiopathic pulmonary fibrosis (IPF). Other prognostic histological markers have not been identified. Artificial intelligence (AI) offers a possibility to quantitate possible prognostic histological features in IPF. We aimed to test the use of AI in IPF lung tissue samples by quantitating FF, interstitial mononuclear inflammation, and intra-alveolar macrophages with a deep convolutional neural network (CNN). Lung tissue samples of 71 patients with IPF from the FinnishIPF registry were analyzed by an AI model developed in the Aiforia® platform. The model was trained to detect tissue, air spaces, FF, interstitial mononuclear inflammation, and intra-alveolar macrophages with 20 samples. For survival analysis, cut-point values for high and low values of histological parameters were determined with maximally selected rank statistics. Survival was analyzed using the Kaplan-Meier method. A large area of FF predicted poor prognosis in IPF (p = 0.01). High numbers of interstitial mononuclear inflammatory cells and intra-alveolar macrophages were associated with prolonged survival (p = 0.01 and p = 0.01, respectively). Of lung function values, low diffusing capacity for carbon monoxide was connected to a high density of FF (p = 0.03) and a high forced vital capacity of predicted was associated with a high intra-alveolar macrophage density (p = 0.03). The deep CNN detected histological features that are difficult to quantitate manually. Interstitial mononuclear inflammation and intra-alveolar macrophages were novel prognostic histological biomarkers in IPF. Evaluating histological features with AI provides novel information on the prognostic estimation of IPF.

Background The objective was to build a novel method for automated image analysis to locate and quantify the number of cytokeratin 7 (K7)-positive hepatocytes reflecting cholestasis by applying deep learning neural networks (AI model) in a cohort of 210 liver specimens. We aimed to study the correlation between the AI model's results and disease progression. The cohort of liver biopsies which served as a model of chronic cholestatic liver disease comprised of patients diagnosed with primary sclerosing cholangitis (PSC). Methods In a cohort of patients with PSC identified from the PSC registry of the University Hospital of Helsinki, their K7-stained liver biopsy specimens were scored by a pathologist (human K7 score) and then digitally analyzed for K7-positive hepatocytes (K7%area). The digital analysis was by a K7-AI model created in an Aiforia Technologies cloud platform. For validation, values were human K7 score, stage of disease (Metavir and Nakunuma fibrosis score), and plasma liver enzymes indicating clinical cholestasis, all subjected to correlation analysis. Results The K7-AI model results (K7%area) correlated with the human K7 score (0.896; p < 2.2e(- 16)). In addition, K7%area correlated with stage of PSC (Metavir 0.446; p < 1.849e(- 10) and Nakanuma 0.424; p < 4.23e(- 10)) and with plasma alkaline phosphatase (P-ALP) levels (0.369, p < 5.749e(- 5)). Conclusions The accuracy of the AI-based analysis was comparable to that of the human K7 score. Automated quantitative image analysis correlated with stage of PSC and with P-ALP. Based on the results of the K7-AI model, we recommend K7 staining in the assessment of cholestasis by means of automated methods that provide fast (9.75 s/specimen) quantitative analysis.

Background: The proper estimate of the risk of recurrences in early-stage oral tongue squamous cell carcinoma (OTSCC) is mandatory for individual treatment-decision making. However, this remains a challenge even for experienced multidisciplinary centers. Objectives: We compared the performance of four machine learning (ML) algorithms for predicting the risk of locoregional recurrences in patients with OTSCC. These algorithms were Support Vector Machine (SVM), Naive Bayes (NB), Boosted Decision Tree (BDT), and Decision Forest (DF). Materials and methods: The study cohort comprised 311 cases from the five University Hospitals in Finland and A.C. Camargo Cancer Center, Sao Paulo, Brazil. For comparison of the algorithms, we used the harmonic mean of precision and recall called F1 score, specificity, and accuracy values. These algorithms and their corresponding permutation feature importance (PFI) with the input parameters were externally tested on 59 new cases. Furthermore, we compared the performance of the algorithm that showed the highest prediction accuracy with the prognostic significance of depth of invasion (DOI). Results: The results showed that the average specificity of all the algorithms was 71% The SVM showed an accuracy of 68% and F1 score of 0.63, NB an accuracy of 70% and F1 score of 0.64, BDT an accuracy of 81% and F1 score of 0.78, and DF an accuracy of 78% and F1 score of 0.70. Additionally, these algorithms outperformed the DOI-based approach, which gave an accuracy of 63%. With PFI-analysis, there was no significant difference in the overall accuracies of three of the algorithms; PFI-BDT accuracy increased to 83.1%, PFI-DF increased to 80%, PFI-SVM decreased to 64.4%, while PFI-NB accuracy increased significantly to 81.4%. Conclusions: Our findings show that the best classification accuracy was achieved with the boosted decision tree algorithm. Additionally, these algorithms outperformed the DOI-based approach. Furthermore, with few parameters identified in the PFI analysis, ML technique still showed the ability to predict locoregional recurrence. The application of boosted decision tree machine learning algorithm can stratify OTSCC patients and thus aid in their individual treatment planning.

The paper describes an approach that combines work from three fields with previously separate research commu-nities: social robotics, conversational AI, and graph databases. The aim is to develop a generic framework in which a variety of social robots can provide high-quality information to users by accessing semantically-rich knowledge graphs about multiple different domains. An example implementation uses a Furhat robot with Rasa open source conversational AI and knowledge graphs in Neo4j graph databases.

Although knee measurements yield high classification rates in metric sex estimation, there is a paucity of studies exploring the knee in artificial intelligence-based sexing. This proof-of-concept study aimed to develop deep learning algorithms for sex estimation from radiographs of reconstructed cadaver knee joints belonging to the Terry Anatomical Collection. A total of 199 knee radiographs were obtained from 100 skeletons (46 male and 54 female cadavers; mean age at death 64.2 years, range 50-102 years) whose tibiofemoral joints were recon-structed in standard anatomical position. The AIDeveloper software was used to train, validate, and test neural network architectures in sex estimation based on image classification. Of the explored algorithms, an MhNet-based model reached the highest overall testing accuracy of 90.3%. The model was able to classify all females (100.0%) and most males (78.6%) correctly. These preliminary findings encourage further research on artificial intelligence-based methods in sex estimation from the knee joint. Combining radiographic data with automated and externally validated algorithms may establish valuable tools to be utilized in forensic anthropology.

The usage of different types of wearable mHealth solutions for consumers has exploded especially since the start of the COVID-19 pandemic. A big question regarding these devices is the quality and accuracy of the data produced by them. When the consumer can use these devices to measure their heartbeat, blood sugar levels, sleep quality, blood oxygen levels etc. the quality and accuracy of this data is getting more important by the day. Not only for the consumer but also for the development of Artificial Intelligence the quality of data is of utmost importance. The importance of the data produced by these devices which the consumers wear voluntarily for long periods of time for the development of Artificial Intelligence in the medical sector cannot be overstated. Many of these mHealth devices also use Artificial Intelligence in one way or another already. In this Thesis the research question is how EU regulation affects the obligations of the producers of mHealth devices in regards the data quality of these devices. The starting point for the research is the define Artificial Intelligence in general and data quality by the EU standards. The method for this research is a legal dogmatic approach to present and future EU regulation surrounding this topic with the viewpoint of ensuring high quality data for Artificial Intelligence development. In the scope of this research there are the Medical Device Regulations for current regulation and the regulations based on the EU Data Strategy, Data Governance Act, the proposal for the Data Act and finally the proposal for the Artificial Intelligence Act. I note that there are many other important aspects to this topic that do not fit into the scope of this Thesis, namely access to data, movement of data, data protection, unfair commercial activities and “soft law” -type of regulation especially standards. The result of the research is that the situation is unclear in the light of the regulations inside the scope of this Thesis. For medical devices, the many obligations for medical devices do ensure that the devices need work as intended and as such ensure the data quality too. Many of the mHealth solutions, however, do not fit into the scope of either of the Medical Device Regulations, because their intended purpose is not ‘medical’. As these devices produce more and more intricate health data, the question left to be answered is when does the intended purpose become medical. EU has tried to tackle this problem mainly by soft law -instruments with the latest being the ISO/TS 82304-2 standard in regards the quality of health and wellness apps released in 2021. For the upcoming regulations the duo of Data related Acts do not bring any light to the problem. They mainly focus on access to data and movement of data with the data quality parts focusing on interoperability of data. The proposal for Artificial Intelligence Act has obligations mainly for the AI systems classified as ‘high-risk’. The interesting part for this paper is how medical devices and security systems for them would be classified as high-risk. This however leads the research back to the Medical Device Regulations and the issue with devices whose intended purpose is not medical.

Fuzzy Cognitive Maps (FCMs) are widely applied for describing the major components of complex systems and their interconnections. The popularity of FCMs is mostly based on their simple system representation, easy model creation and usage, and its decision support capabilities. The preferable way of model construction is based on historical, measured data of the investigated system and a suitable learning technique. Such data are not always available, however. In these cases experts have to define the strength and direction of causal connections among the components of the system, and their decisions are unavoidably affected by more or less subjective elements. Unfortunately, even a small change in the estimated strength may lead to significantly different simulation outcome, which could pose significant decision risks. Therefore, the preliminary exploration of model ‘sensitivity’ to subtle weight modifications is very important to decision makers. This way their attention can be attracted to possible problems. This paper deals with the advanced version of a behavioral analysis. Based on the experiences of the authors, their method is further improved to generate more life-like, slightly modified model versions based on the original one suggested by experts. The details of the method is described, its application and the results are presented by an example of a banking application. The combination of Pareto-fronts and Bacterial Evolutionary Algorithm is a novelty of the approach. © Springer International Publishing AG, part of Springer Nature 2018.

The potential of artificial intelligence (AI) has grown exponentially in recent years, which not only generates value but also creates risks. AI systems are characterised by their complexity, opacity and autonomy in operation. Now and in the foreseeable future, AI systems will be operating in a manner that is not fully autonomous. This signifies that providing appropriate incentives to the human par- ties involved is still of great importance in reducing AI-related harm. Therefore, liability rules should be adapted in such a way to provide the relevant parties with incentives to efficiently reduce the social costs of potential accidents. Relying on a law and economics approach, we address the theo- retical question of what kind of liability rules should be applied to different parties along the value chain related to AI. In addition, we critically analyse the ongoing policy debates in the European Union, discussing the risk that European policymakers will fail to determine efficient liability rules with regard to different stakeholders.