arXiv subjects

Artem Petrov

Publications and source records attributed to Artem Petrov.

11 recordsLinked to original sources

arXiv2026

LCST and Closed-Loop Phase Behavior in Non-Associating Fully Symmetric Multicomponent Polymer Systems

Multicomponent liquids can phase separate upon heating, exhibiting a lower critical solution temperature (LCST). Moreover, a narrow class of materials can undergo disordering transition upon further heating, yielding closed-loop phase diagrams. Previously, it was shown that LCST or closed-loop phase behavior can appear in the models of liquids in which components are asymmetric or interaction potentials have a specifically designed attraction. Here, we show theoretically that LCST and closed-loop phase behavior can occur in a significantly wider set of models. In particular, we found that these phenomena can be exhibited by the simplest and widely used coarse-grained (CG) models of any multicomponent liquid in which species are fully symmetric and where all particles interact via an arbitrary repulsive potential. In addition, we discovered that LCST and closed-loop phase behavior in these models emerges merely due to the basic property of CG liquid models, namely, the appearance of strong monomer-monomer positional correlations at low monomer number density $ρ$. We simulated the models of fully symmetric binary blends and diblock copolymer melts where nonbonded monomers interacted via a generic T-independent purely repulsive harmonic potential. As predicted, LCST and closed-loop phase behavior emerged at $ρ$ sufficiently low to cause strong monomer-monomer correlations leading to a strong T-dependence of the effective coordination number, which, in turn, induced a nonmonotonic T-dependence of the Flory-Huggins parameter. To summarize, we discovered that the simplest fully symmetric non-associating CG models of multicomponent liquids can exhibit complex temperature response due to a mechanism stemming from the basic nature of any CG liquid model. This mechanism might contribute to the emergence of LCST and closed-loop phase behavior in many existing polymer materials.

cond-mat.soft↗

arXiv2026

Surface-Adsorbed Nanodroplets of Symmetric Diblock Copolymers Form Versatile and Stimuli-Responsive Nanostructures

Block copolymers often create droplets when placed on a substrate. Such nanostructured droplets can be arranged into regular microstructured arrays, thereby forming hierarchically organized materials that can be used in microelectronics, plasmonics, sensing, photonics, metamaterials production, and even cryptography. However, it is unclear if such materials can be stimuli-responsive, i.e., be able to change their nanostructure on a single droplet level upon applying external stimuli. In this work, we discovered that small (10-100 nm) surface-adsorbed droplets of symmetric diblock copolymers can form a multitude of different externally switchable nanostructures. We obtained a near-equilibrium, comprehensive 4D diagram of droplet morphologies by performing large-scale self-consistent field theory (SCFT) calculations under various wetting and phase separation conditions. The SCFT modeling was augmented with a computational algorithm that established an equilibrium droplet morphology in a given system without assuming potentially equilibrium structures prior to simulation. The discovered droplet nanostructures agreed excellently with previously published experimental data. Crucially, we showed that direct and reversible transitions between different droplet morphologies are possible upon changing the interaction strength between components, which can be tuned externally in experiments by adding surfactants or controlling temperature. We confirmed experimental realizability of such stimuli-responsiveness by modeling surfactant addition that led to a switch between droplet nanostructures. This work demonstrates that even the simplest symmetric diblock copolymers are able to produce versatile and stimuli-responsive structures on a surface when confined to a small nanodroplet. This opens the possibility to produce smart coatings with externally switchable hierarchical micro- and nanostructures.

cond-mat.soft↗

arXiv2025

GPT-5 at CTFs: Case Studies From Top-Tier Cybersecurity Events

OpenAI and DeepMind's AIs recently got gold at the IMO math olympiad and ICPC programming competition. We show frontier AI is similarly good at hacking by letting GPT-5 compete in elite CTF cybersecurity competitions. In one of this year's hardest events, it outperformed 93% of humans finishing 25th: between the world's #3-ranked team (24th place) and #7-ranked team (26th place). This report walks through our methodology, results, and their implications, and dives deep into 3 problems and solutions we found particularly interesting.

cs.CR↗

arXiv2025

Paired-Sampling Contrastive Framework for Joint Physical-Digital Face Attack Detection

Modern face recognition systems remain vulnerable to spoofing attempts, including both physical presentation attacks and digital forgeries. Traditionally, these two attack vectors have been handled by separate models, each targeting its own artifacts and modalities. However, maintaining distinct detectors increases system complexity and inference latency and leaves systems exposed to combined attack vectors. We propose the Paired-Sampling Contrastive Framework, a unified training approach that leverages automatically matched pairs of genuine and attack selfies to learn modality-agnostic liveness cues. Evaluated on the 6th Face Anti-Spoofing Challenge Unified Physical-Digital Attack Detection benchmark, our method achieves an average classification error rate (ACER) of 2.10 percent, outperforming prior solutions. The framework is lightweight (4.46 GFLOPs) and trains in under one hour, making it practical for real-world deployment. Code and pretrained models are available at https://github.com/xPONYx/iccv2025_deepfake_challenge.

cs.CV↗

arXiv2025

Evaluating AI cyber capabilities with crowdsourced elicitation

As AI systems become increasingly capable, understanding their offensive cyber potential is critical for informed governance and responsible deployment. However, it's hard to accurately bound their capabilities, and some prior evaluations dramatically underestimated them. The art of extracting maximum task-specific performance from AIs is called "AI elicitation", and today's safety organizations typically conduct it in-house. In this paper, we explore crowdsourcing elicitation efforts as an alternative to in-house elicitation work. We host open-access AI tracks at two Capture The Flag (CTF) competitions: AI vs. Humans (400 teams) and Cyber Apocalypse (8000 teams). The AI teams achieve outstanding performance at both events, ranking top-5% and top-10% respectively for a total of \$7500 in bounties. This impressive performance suggests that open-market elicitation may offer an effective complement to in-house elicitation. We propose elicitation bounties as a practical mechanism for maintaining timely, cost-effective situational awareness of emerging AI capabilities. Another advantage of open elicitations is the option to collect human performance data at scale. Applying METR's methodology, we found that AI agents can reliably solve cyber challenges requiring one hour or less of effort from a median human CTF participant.

cs.CR↗

arXiv2024

Hacking CTFs with Plain Agents

We saturate a high-school-level hacking benchmark with plain LLM agent design. Concretely, we obtain 95% performance on InterCode-CTF, a popular offensive security benchmark, using prompting, tool use, and multiple attempts. This beats prior work by Phuong et al. 2024 (29%) and Abramovich et al. 2024 (72%). Our results suggest that current LLMs have surpassed the high school level in offensive cybersecurity. Their hacking capabilities remain underelicited: our ReAct&Plan prompting strategy solves many challenges in 1-2 turns without complex engineering or advanced harnessing.

cs.CR↗

arXiv2024

Simple Calibration of Block Copolymer Melt Models

According to the universality hypothesis, the phase behavior of different block copolymer melt models having fixed composition depends solely on two parameters: the invariant chain length $\bar{N}$ and the effective interaction parameter $χN$. If models behave universally, they can be compared to each other and can predict experiment quantitatively. Here, we present a simple way to achieve this universality for coarse-grained models. Our method relies on the properties of the monomer interaction potential energy $z$ distribution. In particular, models having near-symmetric $z$-distributions exhibit universal phase behavior using the standard linear definition of the Flory-Huggins parameter $χ\proptoα$, where $α= ε_{AB}-(ε_{AA}+ε_{BB})/2$, and $ε_{xy}$ is the interaction energy between monomers of type $x$ and $y$. Previously, universality had been achieved using a nonlinear $χ(α)$ function which is difficult to obtain and interpret physically. The main parameter controlling the symmetry of the $z$-distribution is the monomer density $ρ$. Above certain $ρ$, models have symmetric $z$-distributions, and their order-disorder transition points follow the universal curve predicted by Fredrickson-Helfand theory in the experimentally relevant $\bar{N} > 10^2$ range. On the other hand, low-$ρ$ models exhibit skewed $z$-distributions, and the simple $χ\proptoα$ formula is no longer universally applicable to them. Our results can be used for correct block copolymer model building leading to a simple and direct comparison of simulations to experiments, which will facilitate the screening of new block copolymer morphologies and support materials design.

cond-mat.soft↗

arXiv2022

Optimal Entanglement of Polymers Promotes Formation of Highly Oriented Fibers

Polymer fibers consist of macromolecules oriented along the fiber axis. Better alignment of chains leads to an increased strength of the fiber. It is believed that the key factor preventing formation of a perfectly oriented fiber is entanglement of polymers. We performed large-scale computer simulations of uniaxial stretching of semicrystalline ultrahigh molecular weight polyethylene. We discovered that there is an optimal number of entanglements per macromolecule necessary to maximize chain orientation in a fiber. Polymers that were entangled too strongly formed less oriented fibers. On the other hand, when polymers had too few entanglements per chain, they disentangled during stretching, and the strong fiber was not formed. We constructed a microscopic analytical theory describing both the fiber formation and disentanglement processes. Our work presents a novel view on the role of entanglements during fiber production and predicts the existence of a single universal optimal number of entanglements per chain maximizing the fiber quality: approximately $10^2$ entanglements.

cond-mat.soft↗

arXiv2020

A novel heterogeneous structure formed by a single multiblock copolymer chain

We studied structures formed by a single $(AB)_k$ multiblock copolymer chain, in which interaction between A-type beads was purely repulsive, and B-type beads tended to aggregate. We studied how attraction between A-type beads and B-type beads affected the structure of the chain. We discovered formation of an equilibrium globular structure, which had unique heterogeneous checkerboard-like distribution of contact density. Unlike the structures usually formed by a single $(AB)_k$ multiblock copolymer chain, this structure had contact enrichment at the boundaries of A and B blocks. This structure was formed by a multiblock copolymer chain, in which B-type beads could form maximum two reversible bonds either with A-type or B-type beads, A-type beads could form maximum one reversible bond with a B-type bead, and interactions between A-type beads were purely repulsive. Multiblock copolymer chains with this type of intrachain interactions can model structure of chromatin in various organisms.

cond-mat.soft↗

arXiv2020

Low-entangled UHMWPE melt: analytical model and computer simulations

In this work we developed a theoretical model to describe the polymerization of very long macromolecules with simultaneous precipitation. As a reference system for our model we consider UHMWPE polymerization process, via homogeneous catalysis and with molecular weight above $10^6$ monomers. We derived time dependency of the entanglement length in the system, and assessed how polymerization rate and fraction of initiators affects this dependency. We show that decrease of the fraction of initiators leads to increase of the average entanglement length in the melt after complete polymerization. The results are supported by computer simulations and in principle could be applied to the wide set of polymerized materials, both crystallized and amorphous.

cond-mat.soft↗

arXiv2019

Kinetic mechanisms of crumpled globule formation

Homopolymer chain with beads forming pairwise reversible bonds is a well-known model in polymer physics. We studied kinetics of homopolymer chain collapse, which was induced by pairwise reversible bonds formation. We compared kinetic mechanism of this coil-globule transition with the mechanism of collapse in a poor solvent. We discovered, that coil-globule transition occurs sufficiently more homogeneously on different scales, if collapse is induced by pairwise reversible bonds formation. This effect leads to formation of transient structures, which are not similar to the classical pearl-necklace conformations formed during collapse in a poor solvent. However, both types of collapse lead to formation of a metastable state of crumpled globule, which is one of the well-known models of interphase chromatin structure in different organisms. Moreover, we found out that stability and dynamics of this state can be controlled by fraction of reversible bonds and bond lifetime.

cond-mat.soft↗