Academic literature on the topic 'Embedding methi'

Introduction: one of the ways to protect multimedia information from unauthorized distribution is the introduction of digital watermarks in the video sequence. In this task, the most important parameter is the stability of the hidden message, which should remain retrievable after various impacts and operations with video data. At the same time, the method of implementing a hidden message should provide as little visual distortion of video data as possible. Purpose: to increase the robustness and secrecy of the method of introducing digital watermarks into the video sequence. Methods: a modification of the method of introducing digital watermarks into video data is proposed, which consists in supplementing the already known methods with new functions based on the fundamental differences between a video stream and a static photo. To solve this problem, it is proposed to apply a terminative approach, the essence of which is to determine the techniques and methods for ensuring the robustness of the marker in the form of threat terminators. Results: an overview of the use of various methods of introducing digital watermarks into graphic images is carried out. A typical information system is described, which uses meths to protect multimedia content from unauthorized distribution. The model of threats to the introduction of digital watermarks is considered. The general statement of the scientific task of developing a method for introducing a robust invisible digital watermark into video data and a description of the stages of the proposed method are presented. The analysis of existing methods and methods of increasing the robust" ness of digital watermarks is carried out in order to identify the necessary terminator techniques that allow leveling the described threats. The developed method is based on the integration of methods and techniques of threat terminators determined by the object of research. Discussion: determining the optimal values of the frequency domain range and the depth of embedding of the marker require further experimental studies.

Adsorbates on metal surfaces have attracted recently the scientific interest both from a fundamental point of view (quantum confinement) and in perspective of application in technology. In particular the electronic properties of such reduced symmetry systems are strictly related to their low dimensionality. An accurate theoretical description of their spectral properties has necessarily to deal with the absence of periodicity that characterizes these systems in one (thin films) or more (adatoms) directions. The embedding method allows to overcome this drawback considering a really infinite system both on vacuum and bulk side. In this thesis the ab initio spectral properties of single adatoms (alkali-metal atom and Ba) on Cu(111) and thin overlayers (K/Cu(111), Bi/Cu(100), O/Fe(100)) are analyzed, also in comparison with experimental results. The capabilities of the theoretical method adopted allow to evidence the role of the substrate band structure on the adsorbates induced electronic states. In particular the aspects related to the resonant charge transfer from the adsorbate's states to the bulk continuum are analyzed. This process represents an elastic decay channel for the surface electronic states and contributes to the elastic lifetime, that we can estimate according to the description of a continuous substrate band structure. The results presented for single adatoms evidence the blockade effect that a surface projected energy gap of the substrate produces on the resonant charge transfer. On the other hand the results relative to the overlayer allows to analyze the complex hybridization mechanism between surface features and substrate states in different points of the surface Brillouin zone. In addition the theoretical description of the electronic properties of overlayers on metal surfaces is devoted in this thesis also to the simulation of experimental findings, namely STM and STS images and photoemission spectra, showing the predictive character of the theoretical approach used.

Nowadays, item-item recommendation plays an important role in modern recommender systems. Traditionally, this is either solved by behavior-based collaborative filtering or content-based meth- ods. However, both kinds of methods often suffer from cold-start problems, or poor performance due to few behavior supervision; and hybrid methods which can leverage the strength of both kinds of methods are needed. In this paper, we propose a semi-parametric embedding framework for this problem. Specifically, the embedding of an item is composed of two parts, i.e., the parametric part from content information and the non-parametric part designed to encode behavior information; meanwhile, a deep learning algorithm is proposed to learn two parts simultaneously. Extensive experiments on real-world datasets demonstrate the effectiveness and robustness of the proposed method.

Recommendations are commonly used to modify user’s natural behavior, for example, increasing product sales or the time spent on a website. This results in a gap between the ultimate business ob- jective and the classical setup where recommenda- tions are optimized to be coherent with past user be- havior. To bridge this gap, we propose a new learn- ing setup for recommendation that optimizes for the Incremental Treatment Effect (ITE) of the policy. We show this is equivalent to learning to predict recommendation outcomes under a fully random recommendation policy and propose a new domain adaptation algorithm that learns from logged data containing outcomes from a biased recommenda- tion policy and predicts recommendation outcomes according to random exposure. We compare our method against state-of-the-art factorization meth- ods, in addition to new approaches of causal rec- ommendation and show significant improvements.