У оквиру семинара Лабораторије за спектроскопију плазме и ласере Института за физику у Београду, у понедељак, 07. новембра 2022. године у 12 часова у читаоници библиотеке „др Драган Поповић“, др Благоје Ђорђевић (Национална лабораторија Лоренц Ливермор, САД ), одржаће предавање:

Transfer learning and multi-fidelity modeling of laser-driven particle acceleration

САЖЕТАК:

Computer models of intense, laser-driven ion acceleration require expensive particle-in-cell (PIC) simulations that may struggle to capture all the multi-scale,multi-dimensional physics involved at reasonable costs. Explored is an approach to ameliorate this deficiency using a multi-fidelity model that can incorporate physical trends and phenomena at different levels. As the base framework for this study, an ensemble of approximately 10,000 1D PIC simulations was generated to buttress separate ensembles of hundreds of higher fidelity 1D and 2D simulations. Using transfer learning with deep neural networks, one can reproduce the results of more complex physics at a much smaller cost. The networks trained in this fashion can in turn act as surrogate models for the simulations themselves, allowing for quick and efficient exploration of the parameter space of interest. Standard figures-of-merit were used as benchmarks such as the hot electron temperature, peak ion energy, conversion efficiency, etc. These surrogate models are also useful for incorporating more complex schemes, such as pulse shaping. We can rapidly identify and explore under what conditions dimensionality becomes an important effect and search for outliers in feature space.

У оквиру семинара Центра за изучавање комплексних система Института за физику у Београду, у четвртак, 03. новембра 2022. године у 14 часова путем Zoom платформе, др Наташа Аџић (Центар за изучавање комплексних система, Институт за физику у Београду ), одржаће предавање:

DNA-based dendrimers: from a single molecule to the new state of matter

САЖЕТАК:

In this talk we will present a joint theoretical-experimental study of a novel class of macromolecules composed of short DNA segments joined together to form tree-like structures [1]. To model various generations of these DNA-based dendrimers in ionic solution at room temperature, we will employ a bead-spring model and show the obtained equilibrium characteristics of a single molecule [2]. These results classify DNA-dendrimers as good candidates for nanocarriers due to their robustness, regular voids, and high resistance to stimuli – responsiveness. The second part of the talk will be dedicated to the self-assembling of the second generation of DNA-based dendrimers. We will demonstrate how the method of effective interactions can be used to study organization of the molecules in dense solutions and test our results with fully atomistic molecular dynamics simulations and experiments. Both computational and experimental results show an anomalous liquid-like behavior for wide range of DNA-dendrimers concentrations. Finally, we will analyze the system composed of two dendritic DNA segments connected by a polymer chain and show that solutions of these molecules undergo crystallization to the new state of matter. They form soft cluster crystals, periodic structure with sites occupied by several overlapping molecules [3].

[1] Y. Li, Y. Tseng, and D. Luo, Nat. Mater. 3, 38 (2004).
[2] C. Jochum, N. Adžić, E. Stiakakis et al., Nanoscale 11, 1604 (2019).
[3] E. Stiakakis, N. Jung, N. Adžić et al.,  Nat. Commun. 12, 7167 (2021).

Приступите предавању
Meeting ID: 858 1982 4510
Passcode: 868079

У оквиру семинара Центра за изучавање комплексних система Института за физику у Београду, у четвртак, 06. октобра 2022. године у 14 часова путем Zoom платформе, др Дарија Обрадовић (Лабораторија за биомиметику, Институт за физику у Београду ), одржаће предавање:

Modeling the retention and molecular mechanisms in chromatography: fundamentals and applicability in drug development and discovery

САЖЕТАК:

The discovery and development of new drugs is a complex time-consuming, multi-step process. The first phases of drug discovery are increasingly dependent on advanced computer simulations, and laboratory technologies. Technological progress has made it possible to reliably select the most effective drugs-candidates in the first stages, which will further enter animal and human testing. In this way, production costs and ethical risks are reduced, while efficiency is increased. The liquid chromatography techniques can be used as laboratory models for preliminary screening in drug discovery. The types of interactions that can occur in a chromatographic system often correspond to drug interactions in the living environment. The successful application of chromatography is ensured by understanding the physical aspects of the leading mechanisms that affect the drug behavior in the system (e.g., entropy-controlled behavior), by mathematical modeling – predicting the drug behavior in relation to the system parameters, derivation of chromatographic parameters and their physico-chemical meaning, and by defining the main molecular mechanisms (e.g., van der Waals interactions). So far, there are no general patterns by which the drug behavior in the chromatographic system can be predicted and defined. Depending on the type of system and group of compounds, it is necessary to determine reliable conditions and models regarding the main fundamental processes.

In this talk, we will discuss methods for modeling and understanding the chromatographic behavior of drugs that have central neural activity for the system of dual hydrophilic/reversed-phase interactions, and supercritical liquid chromatography [1-4].

[1] D. Obradović et al., J. Chromatogr. A 1678, 463340 (2022).
[2] D. Obradović et al., J. Chromatogr. A 1619, 460951 (2020).
[3] D. Obradović et al., J. Chromatogr. A 1603, 371 (2019).
[4] D. Obradović et al., J. Chromatogr. A 1585, 92 (2019).

Приступите предавању
Meeting ID: 831 5700 4145
Passcode: 151410