<p>The rapid analysis of complex samples using mass spectrometry (MS) provides valuable information in both point-of-care (e.g. drug testing) and laboratory-based applications, including the generation of spectral libraries for classification of biosamples, the identification of biomarkers through large-scale studies, as well as the synthesis and bioactivity assessments of large compound sets necessary for drug discovery. In all these cases, the inherent speed of MS is attractive, but rarely fully utilized due to the widespread use of sample purification techniques prior to analysis. Ambient ionization methodologies can help circumvent this drawback by facilitating high-throughput qualitative and quantitative analysis directly from the complex samples without any need for work-up. For instance, the use of swabs or paper substrates allows for rapid identification, quantification, and confirmation, of drugs of abuse from biofluids or surfaces of forensic interest in a matter of minutes, as described in the first two chapters of this dissertation. Faster analysis can be achieved using an automated desorption electrospray ionization (DESI) platform which allows for the rapid and direct screening of complex-sample microarrays with throughputs better than 1 sample per second, giving access to rich spectral information from tens of thousands of samples per day. The development of the bioanalytical capabilities of this platform, particularly within the context of drug discovery (e.g. bioactivity assays, biosample analysis), is described across most other chapters of this dissertation. The use of DESI, a contactless ambient ionization method developed in our laboratory and whose 20 years of history are overviewed in the introduction of this document, provides an additional advantage as the secondary microdroplets generated through the DESI process act as reaction vessels that can accelerate organic reactions by up to six orders of magnitude, facilitating on-the-fly synthesis of new compounds from arrays of starting materials. Unique implications of this microdroplet chemistry in the prebiotic synthesis of peptides and spontaneous redox chemistry at air-solution interfaces, together with its practical applications to the synthesis of new drug molecules, are also overviewed. The success obtained with the first automated DESI-MS system, developed within the DARPA Make It program, led to increased interest in a new-generation platform which was designed over the past year, as overviewed in the last section of this dissertation, and which is currently being installed for validation prior to the transfer of the technology to NCATS, where we anticipate it will make a significant impact through the consolidation and acceleration of the early drug discovery workflow.</p>

The widespread availability of image editing tools and improvements in image processing techniques make image manipulation feasible for the general population. Oftentimes, easy-to-use yet sophisticated image editing tools produce results that contain modifications imperceptible to the human observer. Distribution of forged images can have drastic ramifications, especially when coupled with the speed and vastness of the Internet. Therefore, verifying image integrity poses an immense and important challenge to the digital forensic community. Satellite images specifically can be modified in a number of ways, such as inserting objects into an image to hide existing scenes and structures. In this thesis, we describe the use of a Conditional Generative Adversarial Network (cGAN) to identify the presence of such spliced forgeries within satellite images. Additionally, we identify their locations and shapes. Trained on pristine and falsified images, our method achieves high success on these detection and localization objectives.

Virtuální realita je v současné době využívána nejen pro zábavu, ale i pro práci a sociální interakci, kde má soukromí a důvěrnost informací vysokou prioritu. Avšak bohužel, bezpečnostní opatření uplatňovaná dodavateli softwaru často nejsou dostačující. Tato práce přináší rozsáhlou bezpečnostní analýzu populární aplikace Bigscreen pro virtuální realitu, která má více než 500 000 uživatelů. Byly využity techniky analýzy síťového provozu, penetračního testování, reverzního inženýrství a dokonce i metody pro application crippling. Výzkum vedl k odhalení kritických zranitelností, které přímo narušovaly soukromí uživatelů a umožnily útočníkovi plně převzít kontrolu nad počítačem oběti. Nalezené bezpečnostní chyby umožnily distribuci škodlivého softwaru a vytvoření botnetu pomocí počítačového červa šířícího se ve virtuálních prostředích. Byl vytvořen nový kybernetický útok ve virtální realitě nazvaný Man-in-the-Room. Dále byla objevena bezpečnostní chyba v Unity engine. Zodpovědné nahlášení objevených chyb pomohlo zmírnit rizika pro více než půl milionu uživatelů aplikace Bigscreen a uživatele všech dotčených aplikací v Unity po celém světě.