Data Mining Practical Machine Learning Tools and Techniques

The report reflects an agreement based on the consensus conference of the International Standardization Committee on the Objective Assessment of the Nasal Airway in Riga, 2nd Nov. 2016. The aim of the conference was to address the existing nasal airway function tests and to take into account physical, mathematical and technical correctness as a base of international standardization as well as the requirements of the Council Directive 93/42/EEC of 14 June 1993 concerning medical devices. Rhinomanometry, acoustic rhinometry, peak nasal inspiratory flow, Odiosoft-Rhino, optical rhinometry, 24-h measurements, computational fluid dynamics, nasometry and the mirrow test were evaluated for important diagnostic criteria, which are the precision of the equipment including calibration and the software applied; validity with sensitivity, specificity, positive and negative predictive values, reliability with intra-individual and inter-individual reproducibility and responsiveness in clinical studies. For rhinomanometry, the logarithmic effective resistance was set as the parameter of high diagnostic relevance. In acoustic rhinometry, the area of interest for the minimal cross-sectional area will need further standardization. Peak nasal inspiratory flow is a reproducible and fast test, which showed a high range of mean values in different studies. The state of the art with computational fluid dynamics for the simulation of the airway still depends on high performance computing hardware and will, after standardization of the software and both the software and hardware for imaging protocols, certainly deliver a better understanding of the nasal airway flux.

/pdf/the-new-agreement-of-the-international-riga-consensus-5jdrug43kb.pdf

The new agreement of the international RIGA consensus conference on nasal airway function tests.

Recent studies report that autoregressive language models can successfully solve many NLP tasks via zero- and few-shot learning paradigms, which opens up new possibilities for using the pre-trained language models. This paper introduces two autoregressive GPT-like models with 1.3 billion and 13 billion parameters trained on 60 languages from 25 language families using Wikipedia and Colossal Clean Crawled Corpus. We reproduce the GPT-3 architecture using GPT-2 sources and the sparse attention mechanism; Deepspeed and Megatron frameworks allow us to parallelize the training and inference steps effectively. The resulting models show performance on par with the recently released XGLM models by Facebook, covering more languages and enhancing NLP possibilities for low resource languages of CIS countries and Russian small nations. We detail the motivation for the choices of the architecture design, thoroughly describe the data preparation pipeline, and train five small versions of the model to choose the most optimal multilingual tokenization strategy. We measure the model perplexity in all covered languages and evaluate it on the wide spectre of multilingual tasks, including classification, generative, sequence labeling and knowledge probing. The models were evaluated with the zero-shot and few-shot methods. Furthermore, we compared the classification tasks with the state-of-the-art multilingual model XGLM. source code and the mGPT XL model are publicly released.

mGPT: Few-Shot Learners Go Multilingual

In this paper, the problem of load balancing in network intrusion detection system is considered. Load balancing method based on work of several components of network intrusion detection system and on the analysis of multifractal properties of incoming traffic is proposed. The proposed method takes into account a degree of multifractality for calculation of deep packet inspection time, on the basis of which the time necessary for comparing the packet with the signatures is calculated. Load balancing rules are generated using the estimated average deep packet inspection time and the multifractality parameters of incoming load. Comparative analysis of the proposed load balancing method with the standard one showed that the proposed method improves the quality of service parameters and the percentage of packets that are not analyzed.

Method of self-similar load balancing in network intrusion detection system

Spiral computed tomography (SCT) is the most accessible and informative imaging method. Medical specialists face a number of difficulties with searching of CT-images. The goal of our study was to optimize interpretation of the data of radiological methods used for assessment of the paranasal sinuses (PNSs) with estimation of measurement uncertainty. The results of this study allow us to objectively assess the condition of the bone tissue and increase the efficiency of SCT examination of the PNSs, obtained taking into account both the minimum and maximum bone thickness, giving a possibility to predict the course of pathological processes in the PNSs, depending on their anatomical features. After studies, data were obtained to optimize the assessment of morphological parameters of the bone tissue of the paranasal sinuses (bone thickness and density). Potential fluctuations have been identified that should be considered when interpreting the results of calculating indices of PNS bone thickness and density.

Uncertainty of Measurement Results for Anatomical Structures of Paranasal Sinuses

Developing artificial learning systems that can understand and generate natural language has been one of the long-standing goals of artificial intelligence. Recent decades have witnessed an impressive progress on both of these problems, giving rise to a new family of approaches. Especially, the advances in deep learning over the past couple of years have led to neural approaches to natural language generation (NLG). These methods combine generative language learning techniques with neural-networks based frameworks. With a wide range of applications in natural language processing, neural NLG (NNLG) is a new and fast growing field of research. In this state-of-the-art report, we investigate the recent developments and applications of NNLG in its full extent from a multidimensional view, covering critical perspectives such as multimodality, multilinguality, controllability and learning strategies. We summarize the fundamental building blocks of NNLG approaches from these aspects and provide detailed reviews of commonly used preprocessing steps and basic neural architectures. This report also focuses on the seminal applications of these NNLG models such as machine translation, description generation, automatic speech recognition, abstractive summarization, text simplification, question answering and generation, and dialogue generation. Finally, we conclude with a thorough discussion of the described frameworks by pointing out some open research directions.

Neural Natural Language Generation: A Survey on Multilinguality, Multimodality, Controllability and Learning

The paper proposes a new 3D point cloud filtering approach using F-transform. We achieve this by usage of uniform fuzzy partitioning and applying direct and inverse discrete F-transform on a point cloud data. The point cloud was filtered in a depth domain. The performance of developed approach was compared with several common-used statistical-based point cloud filtering methods.

F-transform 3D Point Cloud Filtering Algorithm

Active anterior rhinomanometry is important method for diagnosis of rhinological disorders. This paper presents the new approach for feature extraction based on chaos theory for tasks of rhinology. It has been demonstrated that rhinomanometric signals have a fractal properties. The usage of phase space diagram for feature extraction for rhinomanometric data was proposed.

Usage of phase space diagram to finding significant features of rhinomanometric signals

Aerodynamic characteristics of nasal airflow are important for obstruction problems detection. In clinical practice, we analyze behavior of hydrodynamic resistance coefficient. This paper proposes a component of real-time automated system for rhinomanometric measurements. The component is a real-time alert to help users define a correct measurement.

A new intelligence-based approach for rhinomanometric data processing

A LARS-based method is proposed for constructing a fuzzy regression model. Peculiarities of the use of fuzzy regression analysis for medical diagnosis are considered. The method allows one to reduce the number of parameters of the model that exert influence on the predictable degree of nasal obstruction and to avoid the risk of "overtraining" of the model.

Oleksii Turuta

Papers

Neural Natural Language Generation: A Survey on Multilinguality, Multimodality, Controllability and Learning

F-transform 3D Point Cloud Filtering Algorithm

Usage of phase space diagram to finding significant features of rhinomanometric signals

A new intelligence-based approach for rhinomanometric data processing

A Lars-Based Method of the Construction of a Fuzzy Regression Model for the Selection of Significant Features