We survey SoTA open-domain conversational AI models with the purpose of presenting the prevailing challenges that still exist to spur future research. In addition, we provide statistics on the gender of conversational AI in order to guide the ethics discussion surrounding the issue. Open-domain conversational AI are known to have several challenges, including bland responses and performance degradation when prompted with figurative language, among others. First, we provide some background by discussing some topics of interest in conversational AI. We then discuss the method applied to the two investigations carried out that make up this study. The first investigation involves a search for recent SoTA open-domain conversational AI models while the second involves the search for 100 conversational AI to assess their gender. Results of the survey show that progress has been made with recent SoTA conversational AI, but there are still persistent challenges that need to be solved, and the female gender is more common than the male for conversational AI. One main take-away is that hybrid models of conversational AI offer more advantages than any single architecture. The key contributions of this survey are 1) the identification of prevailing challenges in SoTA open-domain conversational AI, 2) the unusual discussion about open-domain conversational AI for low-resource languages, and 3) the discussion about the ethics surrounding the gender of conversational AI.

One of the critical problems in multiclass classification tasks is the imbalance of the dataset. This is especially true when using contemporary pre-trained neural networks, where, in fact, the last layers of the neural network are retrained. Therefore, the large datasets with highly unbalanced classes are not good for models’ training since the use of such a dataset leads to overfitting and, accordingly, poor metrics on test and validation datasets. In this paper the sensitivity to a dataset imbalance of Xception, ViT-384, ViT-224, VGG19, ResNet34, ResNet50, ResNet101, Inception_v3, DenseNet201, DenseNet161, DeIT was studied using a highly imbalanced dataset of 20,971 images sorted into 7 classes. It is shown that the best metrics were obtained when using a cropped dataset with augmentation of missing images in classes up to 15% of the initial number. So, the metrics can be increased by 2-6% compared to the metrics of the models on the initial unbalanced data set. Moreover, the metrics of the rare classes' classification also improved significantly – the TruePositive value can be increased by 0.3 and more. As result, the best approach to train considered networks on an initially unbalanced dataset was formulated.

In this paper, we present the results of the research concerning extraction of informative gene expression profiles from high-dimensional array of gene expressions considering the state of patients' health using clustering method, ensemble of binary classifiers and fuzzy inference system. Applying of the proposed stepwise procedure can allow us to extract the most informative genes taking into account both the subtypes of disease or state of the patient's health for further reconstruction of gene regulatory networks based on the allocated genes and following simulation of the reconstructed models. We used the publicly available gene expressions data as the experimental ones which were obtained using DNA microarray experiments and contained two types of patients' gene expression profiles: the patients with lung cancer tumor and healthy patients. The stepwise procedure of the data processing assumes the following steps: in beginning, we reduce the number of genes by removing non-informative genes in terms of statistical criteria and Shannon entropy; then, we perform the stepwise hierarchical clustering of gene expression profiles at hierarchical levels from 1 to 10 using SOTA clustering algorithm with correlation distance metric. The quality of the obtained clustering was evaluated using complex clustering quality criterion which is considered both the gene expression profiles distribution relative to center of the clusters were these gene expression profiles are allocated and the centers of the clusters distribution. The result of this stage execution was selection of the optimal cluster at each of the hierarchical levels which corresponded to minimum value of the quality criterion. At the next step, we have implemented classification procedure of the examined objects using four well known binary classifiers: logistic regression, support-vector machine, decision trees and random forest classifier. The effectiveness of the appropriate technique was evaluated based on the use of ROC analysis using criteria included as the components the errors of both the first and the second kinds. The final decision concerning extraction of the most informative subset of gene expression profiles was taken based on the use of fuzzy inference system, the inputs of which are the results of the appropriate single classifiers operation and output is the final solution concerning state of the patient's health. To our mind, the implementation of the proposed stepwise procedure of the informative gene expression profiles extraction create the conditions for increasing effectiveness of the further procedure of gene regulatory networks reconstruction and the following simulation of the reconstructed models considering the subtypes of the disease and/or state of the patient’s health.

ABSTRACT. The immunotherapies, as a modern therapeutic approach, get an attention because of theirs’ promise to treat a large number of different medical disorders. Immunomodulation effects of low titres (10 HA/ml) of NDV (Newcastle Disease Virus) ZG1999HDS or La Sota were tested on TLT (Human macrophage cell line) bound to PBMC (Peripheral Blood Mononuclear Cells). During the immunomodulation, the amount of NO, H2O2, lysozym and induced antibacterial activity against Gram - positive bacteria (Staphylococcus aureus, MRSA, Streptococcus pyogenes, Streptococcus agalactiae and Streptococcus mutants) and against Gram - negative bacteria (Escherichia coli, Pseudomonas aeruginosa, Proteus mirabilis and Acinetobacter baumanii) were analysed. In addition, the cytokine secretion, IL-1α, IL-2, IL-4, GM-CSF, TNF-α, IFN-α and IFN-α were evaluated. Firstly, the TLT cells are activated through the NDV ZG1999HDS or La Sota binding, followed by the NO “burst” and H2O2 and lysozyme level increase. Secondly, after the binding to the TLT cells and interaction with the PBMCs, the decrease of GM-CSF, and an increase of TNF – α and IFN – γ were found. Simultaneously, the decrease of pro – inflammatory cytokine IFN-α and the differentially increase of IL-1α, IL-2 and IL-4 were recorded. During the induction of the antibacterial response, against Gram - positive bacteria (Staphylococcus aureus, MRSA, Streptococcus pyogenes, Streptococcus agalactiae and Streptococcus mutants) the effect was one third higher with NDV ZG1999HDS compared to La Sota. Antibacterial response against Gram - negative bacteria (Escherichia coli, Pseudomonas aeruginosa, Proteus mirabilis and Acinetobacter baumanii) was not so clear. In general, NDV ZG1999HDS or La Sota activated TLT cells, further bound to PBMC; the ZG1999HDS is stronger immunomodulator than La Sota.

The proliferation of deep learning in natural language processing (NLP) has led to the development and release of innovative technologies capable of understanding and generating human language with remarkable proficiency. Atinuke, a Transformer-based neural network, optimises performance across various language tasks by utilising a unique configuration. The architecture interweaves layers for processing sequential data with attention mechanisms to draw meaningful affinities between inputs and outputs. Due to the configuration of its topology and hyperparameter tuning, it can emulate human-like language by extracting features and learning complex mappings. Atinuke is modular, extensible, and integrates seamlessly with existing machine learning pipelines. Advanced matrix operations like softmax, embeddings, and multi-head attention enable nuanced handling of textual, acoustic, and visual signals. By unifying modern deep learning techniques with software design principles and mathematical theory, the system achieves state-of-the-art results on natural language tasks whilst remaining interpretable and robust.