Siamese Neural Network for Keystroke Dynamics-Based Authentication on Partial Passwords.

The paper addresses issues concerning secure authentication in computer systems. We focus on multi-factor authentication methods using two or more independent mechanisms to identify a user. User-specific behavioral biometrics is widely used to increase login security. The usage of behavioral biometrics can support verification without bothering the user with a requirement of an additional interaction. Our research aimed to check whether using information about how partial passwords are typed is possible to strengthen user authentication security. The partial password is a query of a subset of characters from a full password. The use of partial passwords makes it difficult for attackers who can observe password entry to acquire sensitive information. In this paper, we use a Siamese neural network and n-shot classification using past recent logins to verify user identity based on keystroke dynamics obtained from the static text. The experimental results on real data demonstrate that keystroke dynamics authentication can be successfully used for partial password typing patterns. Our method can support the basic authentication process and increase users' confidence.

Hybrid Polymer-Inorganic Materials with Hyaluronic Acid as Controlled Antibiotic Release Systems.

In recent years, significant developments have taken place in scientific fields such as tissue and materials engineering, which allow for the development of new, intelligent biomaterials. An example of such biomaterials is drug delivery systems that release the active substance directly at the site where the therapeutic effect is required. In this research, polymeric materials and ceramic-polymer composites were developed as carriers for the antibiotic clindamycin. The preparation and characterization of biomaterials based on hyaluronic acid, collagen, and nano brushite obtained using the photocrosslinking technique under UV (ultraviolet) light are described. Physical and chemical analyses of the materials obtained were carried out using Fourier transform infrared spectroscopy (FT-IR) and optical microscopy. The sorption capacities were determined and subjected to in vitro incubation in simulated biological environments such as Ringer's solution, simulated body fluid (SBF), phosphate-buffered saline (PBS), and distilled water. The antibiotic release rate was also measured. The study confirmed higher swelling capacity for materials with no addition of a ceramic phase, thus it can be concluded that brushite inhibits the penetration of the liquid medium into the interior of the samples, leading to faster absorption of the liquid medium. In addition, incubation tests confirmed preliminary biocompatibility. No drastic changes in pH values were observed, which suggests that the materials are stable under these conditions. The release rate of the antibiotic from the biomaterial into the incubation medium was determined using high-pressure liquid chromatography (HPLC). The concentration of the antibiotic in the incubation fluid increased steadily following a 14-day incubation in PBS, indicating continuous antibiotic release. Based on the results, it can be concluded that the developed polymeric material demonstrates potential for use as a carrier for the active substance.

Classification of masked image data.

Data classification is one of the most commonly used applications of machine learning. The are many developed algorithms that can work in various environments and for different data distributions that perform this task with excellence. Classification algorithms, just like other machine learning algorithms have one thing in common: in order to operate on data, they must see the data. In the present world, where concerns about privacy, GDPR (General Data Protection Regulation), business confidentiality and security are growing bigger and bigger; this requirement to work directly on the original data might become, in some situations, a burden. In this paper, an approach to the classification of images that cannot be directly accessed during training has been made. It has been shown that one can train a deep neural network to create such a representation of the original data that i) without additional information, the original data cannot be restored, and ii) that this representation-called a masked form-can still be used for classification purposes. Moreover, it has been shown that classification of the masked data can be done using both classical and neural network-based classifiers.