Search

Large-scale pre-training methodologies for chemical language models represent a breakthrough in cheminformatics. These methods excel in tasks such as property prediction and molecule generation by learning contextualized representations of input tokens through self-supervised learning on large unlabeled corpora. Typically, this involves pre-training on unlabeled data followed by fine-tuning on specific tasks, reducing dependence on annotated datasets and broadening chemical language representation understanding. This paper introduces a large encoder-decoder chemical foundation models pre-trained on a curated dataset of 91 million SMILES samples sourced from PubChem, which is equivalent to 4 billion of molecular tokens. The proposed foundation model supports different complex tasks, including quantum property prediction, and offer flexibility with two main variants (289M and $8\times289M$). Our experiments across multiple benchmark datasets validate the capacity of the proposed model in providing state-of-the-art results for different tasks. We also provide a preliminary assessment of the compositionality of the embedding space as a prerequisite for the reasoning tasks. We demonstrate that the produced latent space is separable compared to the state-of-the-art with few-shot learning capabilities., Comment: 14 pages, 3 figures, 14 tables

It is shown that if the initial condition of the Cauchy problem for the diffusion equation on a general infinite countable ultrametric space is spherically symmetric with respect to some point, then this problem has an exact analytical solution. A general solution of this problem is presented for pure ultrametric diffusion, as well as for ultrametric diffusion with a reaction sink concentrated at the center of spherical symmetry. Conditions on the ultrametric and the distribution of the number of states in ultrametric spheres are found that lead at large times to the asymptotic behavior of the solutions obtained in the form of a power law modulated by a bounded function that is log-periodic under some additional conditions., Comment: 27 pages

The Spin Physics Detector collaboration proposes to install a universal detector in the second interaction point of the NICA collider under construction (JINR, Dubna) to study the spin structure of the proton and deuteron and other spin-related phenomena using a unique possibility to operate with polarized proton and deuteron beams at a collision energy up to 27 GeV and a luminosity up to $10^{32}$ cm$^{-2}$ s$^{-1}$. As the main goal, the experiment aims to provide access to the gluon TMD PDFs in the proton and deuteron, as well as the gluon transversity distribution and tensor PDFs in the deuteron, via the measurement of specific single and double spin asymmetries using different complementary probes such as charmonia, open charm, and prompt photon production processes. Other polarized and unpolarized physics is possible, especially at the first stage of NICA operation with reduced luminosity and collision energy of the proton and ion beams. This document is dedicated exclusively to technical issues of the SPD setup construction.

While several articles have been written on Electrohydrodynamics (EHD) flows or flows with constant vorticity separately, little is known about the extent to which the combined effects of EHD and constant vorticity affect the flow. This study aims to fill this gap by investigating how a horizontal electric field and constant vorticity jointly influence the free surface and the emergence of stagnation points. Using the Euler equations framework, we employ conformal mapping and pseudo-spectral numerical methods. Our findings reveal that increasing the electric field intensity eliminates stagnation points and smoothen the wave profile. This implies that a horizontal electric field acts as a mechanism for the elimination of stagnation points within the fluid body.

BM@N (Baryonic Matter at Nuclotron) is the first experiment operating and taking data at the Nuclotron/NICA ion-accelerating complex.The aim of the BM@N experiment is to study interactions of relativistic heavy-ion beams with fixed targets. We present a technical description of the BM@N spectrometer including all its subsystems., Comment: 34 pages, 47 figures, 6 tables

Computational models are an essential tool for the design, characterization, and discovery of novel materials. Hard computational tasks in materials science stretch the limits of existing high-performance supercomputing centers, consuming much of their simulation, analysis, and data resources. Quantum computing, on the other hand, is an emerging technology with the potential to accelerate many of the computational tasks needed for materials science. In order to do that, the quantum technology must interact with conventional high-performance computing in several ways: approximate results validation, identification of hard problems, and synergies in quantum-centric supercomputing. In this paper, we provide a perspective on how quantum-centric supercomputing can help address critical computational problems in materials science, the challenges to face in order to solve representative use cases, and new suggested directions., Comment: 65 pages, 15 figures; comments welcome

Advanced computational methods are being actively sought for addressing the challenges associated with discovery and development of new combinatorial material such as formulations. A widely adopted approach involves domain informed high-throughput screening of individual components that can be combined into a formulation. This manages to accelerate the discovery of new compounds for a target application but still leave the process of identifying the right 'formulation' from the shortlisted chemical space largely a laboratory experiment-driven process. We report a deep learning model, Formulation Graph Convolution Network (F-GCN), that can map structure-composition relationship of the individual components to the property of liquid formulation as whole. Multiple GCNs are assembled in parallel that featurize formulation constituents domain-intuitively on the fly. The resulting molecular descriptors are scaled based on respective constituent's molar percentage in the formulation, followed by formalizing into a combined descriptor that represents a complete formulation to an external learning architecture. The use case of proposed formulation learning model is demonstrated for battery electrolytes by training and testing it on two exemplary datasets representing electrolyte formulations vs battery performance -- one dataset is sourced from literature about Li/Cu half-cells, while the other is obtained by lab-experiments related to lithium-iodide full-cell chemistry. The model is shown to predict the performance metrics like Coulombic Efficiency (CE) and specific capacity of new electrolyte formulations with lowest reported errors. The best performing F-GCN model uses molecular descriptors derived from molecular graphs that are informed with HOMO-LUMO and electric moment properties of the molecules using a knowledge transfer technique., Comment: 35 pages, 10 figures

We present a novel multimodal language model approach for predicting molecular properties by combining chemical language representation with physicochemical features. Our approach, MULTIMODAL-MOLFORMER, utilizes a causal multistage feature selection method that identifies physicochemical features based on their direct causal effect on a specific target property. These causal features are then integrated with the vector space generated by molecular embeddings from MOLFORMER. In particular, we employ Mordred descriptors as physicochemical features and identify the Markov blanket of the target property, which theoretically contains the most relevant features for accurate prediction. Our results demonstrate a superior performance of our proposed approach compared to existing state-of-the-art algorithms, including the chemical language-based MOLFORMER and graph neural networks, in predicting complex tasks such as biodegradability and PFAS toxicity estimation. Moreover, we demonstrate the effectiveness of our feature selection method in reducing the dimensionality of the Mordred feature space while maintaining or improving the model's performance. Our approach opens up promising avenues for future research in molecular property prediction by harnessing the synergistic potential of both chemical language and physicochemical features, leading to enhanced performance and advancements in the field., Comment: 14 pages, 6 Figures, 5 tables. Submited to NEURIPS 2023, Under review

Logarithmic oscillations superimposed on a power-law trend appear in the behavior of various complex hierarchical systems. In this paper, we study the logarithmic oscillations of relaxation curves in p-adic diffusion models that are used to describe the conformational dynamics of protein. We consider the case of a purely p-adic diffusion, as well as the case of p-adic diffusion with a reaction sink. We show that, relaxation curves for large times in these two cases are described by a power law on which logarithmic oscillations are superimposed whose period and amplitude are determined by the parameters of the model. We also provide a physical explanation of the emergence of oscillations in relaxation curves and discuss the relation of the results to the experiments on relaxation dynamics of protein., Comment: 23 pages, 5 figures

Generative models are a powerful tool in AI for material discovery. We are designing a software framework that supports a human-AI co-creation process to accelerate finding replacements for the ``forever chemicals''-- chemicals that enable our modern lives, but are harmful to the environment and the human health. Our approach combines AI capabilities with the domain-specific tacit knowledge of subject matter experts to accelerate the material discovery. Our co-creation process starts with the interaction between the subject matter experts and a generative model that can generate new molecule designs. In this position paper, we discuss our hypothesis that these subject matter experts can benefit from a more iterative interaction with the generative model, asking for smaller samples and ``guiding'' the exploration of the discovery space with their knowledge., Comment: 5 pages, Generative AI and HCI (GenAICHI) Workshop at CHI 23 (ACM CHI Conference on Human Factors in Computing Systems)

First physics results of the BM@N experiment at the Nuclotron/NICA complex are presented on {\pi}+ and K+ meson production in interactions of an argon beam with fixed targets of C, Al, Cu, Sn and Pb at 3.2 AGeV. Transverse momentum distributions, rapidity spectra and multiplicities of $\pi^+$ and $K^+$ mesons are measured. The results are compared with predictions of theoretical models and with other measurements at lower energies., Comment: 29 pages, 20 figures

Direct numerical simulation of threedimensional chaotic motion of a dielectric liquid with a free surface under the action of external horizontal electric field is carried out. The numerical model takes into account the effects of surface tension, viscosity, and external isotropic random forcing acting at large scales. A transition from dispersive capillary wave turbulence to quasi-isotropic nondispersive EHD surface turbulence with increase of the external electric field strength is numerically observed for the first time. At the regime of developed EHD wave turbulence, the total electrical energy is found to be much greater than the energy of capillary waves, i.e., electrohydrodynamic effects play a dominant role. At the same time, anisotropic effects are detected that lead to the generation of capillary wave packets traveling perpendicular to the external field direction. Despite the revealed anisotropy, the calculated spectrum of EHD wave turbulence is in very good agreement with the analytical spectrum obtained on the basis of dimensional analysis of weak turbulence spectra., Comment: 7 pages,7 figures

While the capabilities of generative models heavily improved in different domains (images, text, graphs, molecules, etc.), their evaluation metrics largely remain based on simplified quantities or manual inspection with limited practicality. To this end, we propose a framework for Multi-level Performance Evaluation of Generative mOdels (MPEGO), which could be employed across different domains. MPEGO aims to quantify generation performance hierarchically, starting from a sub-feature-based low-level evaluation to a global features-based high-level evaluation. MPEGO offers great customizability as the employed features are entirely user-driven and can thus be highly domain/problem-specific while being arbitrarily complex (e.g., outcomes of experimental procedures). We validate MPEGO using multiple generative models across several datasets from the material discovery domain. An ablation study is conducted to study the plausibility of intermediate steps in MPEGO. Results demonstrate that MPEGO provides a flexible, user-driven, and multi-level evaluation framework, with practical insights on the generation quality. The framework, source code, and experiments will be available at https://github.com/GT4SD/mpego.

The weakly nonlinear dynamics of the free surface of a dielectric liquid in an electric field directed tangentially to the unperturbed boundary is investigated numerically. Within the framework of the strong field model, when the effects of capillarity and gravity are not taken into account, it is shown that nonlinear surface waves have a tendency to break. In result of the collapse of the surface waves, the curvature of the boundary and the gradient of the local electric field undergo infinite discontinuity on the surface of the liquid. The angles of the boundary inclination remain small. The characteristic collapse time of a surface wave traveling in a given direction is calculated versus the dielectric constant of the liquid. It is shown that the time of the singularity formation increases infinitely at small and high values of the dielectric constant. The first case corresponds to the transition of the system to the neutral stability regime (the jump in electrostatic pressure at the boundary turns into zero). At high values of the dielectric constant of the liquid, the collapse time also increases. This effect is associated with the realization of a special regime of fluid motion, in which the propagation of nonlinear surface waves of an arbitrary configuration occurs without distortions. For the liquids with relative permittivity close to five, the wave breaking time reaches a minimum, i.e., the collapse of the surface waves for such liquids occurs most intensively., Comment: 7 pages, 5 figures

There is an increasing need in our society to achieve faster advances in Science to tackle urgent problems, such as climate changes, environmental hazards, sustainable energy systems, pandemics, among others. In certain domains like chemistry, scientific discovery carries the extra burden of assessing risks of the proposed novel solutions before moving to the experimental stage. Despite several recent advances in Machine Learning and AI to address some of these challenges, there is still a gap in technologies to support end-to-end discovery applications, integrating the myriad of available technologies into a coherent, orchestrated, yet flexible discovery process. Such applications need to handle complex knowledge management at scale, enabling knowledge consumption and production in a timely and efficient way for subject matter experts (SMEs). Furthermore, the discovery of novel functional materials strongly relies on the development of exploration strategies in the chemical space. For instance, generative models have gained attention within the scientific community due to their ability to generate enormous volumes of novel molecules across material domains. These models exhibit extreme creativity that often translates in low viability of the generated candidates. In this work, we propose a workbench framework that aims at enabling the human-AI co-creation to reduce the time until the first discovery and the opportunity costs involved. This framework relies on a knowledge base with domain and process knowledge, and user-interaction components to acquire knowledge and advise the SMEs. Currently,the framework supports four main activities: generative modeling, dataset triage, molecule adjudication, and risk assessment., Comment: 9 pages, 5 figures, NeurIPS 2022 WS: AI4Science

The incoherent Thomson scattering diagnostic with multiple line of sight is installed at the gas dynamic trap (GDT) for measurements of radial profiles of the plasma electron temperature and density. The diagnostic is based on a Nd:YAG laser with the wavelength of $1064 nm$. Its collecting lens uses $90^\circ$ scattering geometry having eleven lines of sight in total. At the current stage, six of them are equipped with interference filter spectrometers. Results of the initial operation within a month-long experimental campaign at GDT show that radial profiles are delivered with the accuracy of few percent for $T_e \geqslant 20 eV$ and $n_e \geqslant 10^{19} m^{-3}$. The next planned boost of the diagnostic performance is upgrade with the pulse burst laser to study a fast dynamics of plasma parameters in every GDT shot.

We introduce an approach to the targeted completion of lacunae in molecular data sets which is driven by topological data analysis, such as Mapper algorithm. Lacunae are filled in using scaffold-constrained generative models trained with different scoring functions. The approach enables addition of links and vertices to the skeletonized representations of the data, such as Mapper graph, and falls in the broad category of network completion methods. We illustrate application of the topology-driven data completion strategy by creating a lacuna in the data set of onium cations extracted from USPTO patents, and repairing it., Comment: 10 pages, talk presented at APS March Meeting 2021

We consider the problem of the distribution of the sojourn time in a compact set $\mathbb{Z}_{p}$ in the case of a $p$-adic random walk. We rely on the results of our previous studies of the distribution of the first return time for a $p$-adic random walk and the results of Takacs on the study of the sojourn time problem for a wide class of random processes. For a $p$-adic random walk we find the mean sojourn time of the trajectory in $\mathbb{Z}_{p}$ and the asymptotics as $t\rightarrow\infty$ of arbitrary moments of the distribution of the sojourn time in $\mathbb{Z}_{p}$. We also discuss some possible applications of our results to the modeling of relaxation processes related to the conformational dynamics of protein., Comment: 15 pages