Search

Variational inequalities represent a broad class of problems, including minimization and min-max problems, commonly found in machine learning. Existing second-order and high-order methods for variational inequalities require precise computation of derivatives, often resulting in prohibitively high iteration costs. In this work, we study the impact of Jacobian inaccuracy on second-order methods. For the smooth and monotone case, we establish a lower bound with explicit dependence on the level of Jacobian inaccuracy and propose an optimal algorithm for this key setting. When derivatives are exact, our method converges at the same rate as exact optimal second-order methods. To reduce the cost of solving the auxiliary problem, which arises in all high-order methods with global convergence, we introduce several Quasi-Newton approximations. Our method with Quasi-Newton updates achieves a global sublinear convergence rate. We extend our approach with a tensor generalization for inexact high-order derivatives and support the theory with experiments.

Multi-agent pathfinding (MAPF) is the problem of finding a set of conflict-free paths for a set of agents. Typically, the agents' moves are limited to a pre-defined graph of possible locations and allowed transitions between them, e.g. a 4-neighborhood grid. We explore how to solve MAPF problems when each agent can move between any pair of possible locations as long as traversing the line segment connecting them does not lead to a collision with the obstacles. This is known as any-angle pathfinding. We present the first optimal any-angle multi-agent pathfinding algorithm. Our planner is based on the Continuous Conflict-based Search (CCBS) algorithm and an optimal any-angle variant of the Safe Interval Path Planning (TO-AA-SIPP). The straightforward combination of those, however, scales poorly since any-angle path finding induces search trees with a very large branching factor. To mitigate this, we adapt two techniques from classical MAPF to the any-angle setting, namely Disjoint Splitting and Multi-Constraints. Experimental results on different combinations of these techniques show they enable solving over 30% more problems than the vanilla combination of CCBS and TO-AA-SIPP. In addition, we present a bounded-suboptimal variant of our algorithm, that enables trading runtime for solution cost in a controlled manner., Comment: This is a pre-print version of the paper accepted to IROS 2024. Its main body is similar to the camera-ready version of the conference paper. In addition this pre-print contains Appendix

CW magnetrons, initially developed for industrial RF heaters, were suggested to power RF cavities of superconducting accelerators due to their higher efficiency and lower cost than traditionally used klystrons, IOTs or solid-state amplifiers. RF amplifiers driven by a master oscillator serve as coherent RF sources. CW magnetrons are regenerative RF generators with a huge regenerative gain. This causes regenerative instability with a large noise when a magnetron operates with the anode voltage above the threshold of self-excitation. Traditionally for stabilization of magnetrons is used injection locking by a quite small signal. Then the magnetron except the injection locked oscillations may generate noise. This may preclude use of standard CW magnetrons in some SRF accelerators. Recently we developed briefly described below a mode for forced RF generation of CW magnetrons when the magnetron startup is provided by the injected forcing signal and the regenerative noise is suppressed. The mode is most suitable for powering high Q-factor SRF cavities., Comment: 7 pages, 12 figures, one table

We present a catalog of 150 variable stars, including 13 stars with exoplanet candidates. 37 stars were identified as variables for the first time. As a result of a 2.5-year photometric survey of exoplanets, we have obtained and analyzed light curves for almost 50 thousand stars in fields around white dwarfs WD 0009+501 and GRW +708247. Here we describe observations and data processing, the search for variable stars, their cross-identification with other catalogs and classification. The catalog is published in open access and contains the primary classification of variability, light curves and their parameters., Comment: 10 pages, 3 figures, 2 tables

We investigate the two-pulse photon echo response of excitons in the mixed lead halide perovskite crystal \sample in dependence on the excitation intensity and polarization of the incident laser pulses. Using spectrally narrow picosecond laser pulses, we address localized excitons with long coherence times $T_2 \approx 100\,$ps. This approach offers high sensitivity for the observation of excitation-induced changes in the homogeneous linewidth $\Gamma_2=2\hbar/T_2$ on the $\mu$eV scale. Through intensity-dependent measurements, we evaluate the increase of $\Gamma_2$ by 10~$\mu$eV at an exciton density of 10$^{17}$~cm$^{-3}$ being comparable with the intrinsic linewidth of 14$\,\mu$eV. We observe that the decay of the photon echo and its power dependence are sensitive to the polarization configuration of the excitation pulses, which indicates that spin-dependent exciton-exciton interactions contribute to excitation-induced dephasing. In cross-linear polarization, the decay is faster and its dependence on exciton density is stronger as compared to the co-polarized configuration. Using a two-exciton model accounting for different spin configurations we are able to reproduce the experimental results.

Diluted magnetic semiconductor (DMS) colloidal nanocrystals demonstrate remarkable magneto-optical properties. However, the behavior of circular polarization of their emission in high magnetic fields remains unclear. We measure magneto-optical properties of colloidal CdSe/CdMnS nanoplatelets in high magnetic fields up to 30 T and at cryogenic temperatures. The degree of circular polarization of photoluminescence demonstrates non-monotonous behavior in a magnetic field. In low magnetic fields, the polarization degree is positive, due to an exchange interaction of excitons with localized spins of magnetic Mn ions. The exchange interaction is strong enough to overcome the intrinsic Zeeman splitting, which provides negatively polarized emission in nonmagnetic CdSe/CdS nanoplatelets. After reaching a maximum the polarization degree starts to decrease and reverses the sign to negative in high magnetic fields. The critical magnetic field, in which the sign is reversed, increases when temperature is elevated. We develop a model, which explains this behavior by an interplay of bright and dark exciton recombination.

Optical orientation of carrier spins by circularly polarized light is the basic concept and tool of spin physics in semiconductors. We study the optical orientation of electrons and holes in a crystal of the FA$_{0.9}$Cs$_{0.1}$PbI$_{2.8}$Br$_{0.2}$ lead halide perovskite by means of polarized photoluminescence, time-resolved differential reflectivity, and time-resolved Kerr rotation. At the cryogenic temperature of 1.6 K the optical orientation degree measured for continuous-wave excitaton reaches 6% for localized electrons and 2\% for localized holes. Their contributions are distinguished from each other and from exciton optical orientation through the pronounced Hanle effect in transverse magnetic fields and the polarization recovery effect in longitudinal magnetic fields. The optical orientation degree is highly stable against detuning of the laser photon energy from the band gap by up to 0.25 eV, showing then a gradual decrease for detunings up to 0.9 eV. This evidences the inefficiency of spin relaxation mechanisms for free carriers during their energy relaxation. Spin relaxation for localized electrons and holes is provided by the hyperfine interaction with the nuclear spins. Dynamic polarization of nuclear spins is demonstrated by the Overhauser field reaching 4 mT acting on the electrons and $-76$ mT acting on the holes. This confirms the specifics of lead halide perovskite semiconductors, where the hole hyperfine interaction with the nuclei considerably exceeds that of the electron.

Mapping is one of the crucial tasks enabling autonomous navigation of a mobile robot. Conventional mapping methods output dense geometric map representation, e.g. an occupancy grid, which is not trivial to keep consistent for the prolonged runs covering large environments. Meanwhile, capturing the topological structure of the workspace enables fast path planning, is less prone to odometry error accumulation and does not consume much memory. Following this idea, this paper introduces PRISM-TopoMap -- a topological mapping method that maintains a graph of locally aligned locations not relying on global metric coordinates. The proposed method involves learnable multimodal place recognition paired with the scan matching pipeline for localization and loop closure in the graph of locations. The latter is updated online and the robot is localized in a proper node at each time step. We conduct a broad experimental evaluation of the suggested approach in a range of photo-realistic environments and on a real robot (wheeled differential driven Husky robot), and compare it to state of the art. The results of the empirical evaluation confirm that PRISM-Topomap consistently outperforms competitors across several measures of mapping and navigation efficiency and performs well on a real robot. The code of PRISM-Topomap is open-sourced and available at https://github.com/kirillMouraviev/prism-topomap., Comment: This is a pre-print of the paper submitted to the IROS 2024 conference

System-on-Chip Field-Programmable Gate Arrays (SoC-FPGAs) offer significant throughput gains for machine learning (ML) edge inference applications via the design of co-processor accelerator systems. However, the design effort for training and translating ML models into SoC-FPGA solutions can be substantial and requires specialist knowledge aware trade-offs between model performance, power consumption, latency and resource utilization. Contrary to other ML algorithms, Tsetlin Machine (TM) performs classification by forming logic proposition between boolean actions from the Tsetlin Automata (the learning elements) and boolean input features. A trained TM model, usually, exhibits high sparsity and considerable overlapping of these logic propositions both within and among the classes. The model, thus, can be translated to RTL-level design using a miniscule number of AND and NOT gates. This paper presents MATADOR, an automated boolean-to-silicon tool with GUI interface capable of implementing optimized accelerator design of the TM model onto SoC-FPGA for inference at the edge. It offers automation of the full development pipeline: model training, system level design generation, design verification and deployment. It makes use of the logic sharing that ensues from propositional overlap and creates a compact design by effectively utilizing the TM model's sparsity. MATADOR accelerator designs are shown to be up to 13.4x faster, up to 7x more resource frugal and up to 2x more power efficient when compared to the state-of-the-art Quantized and Binary Deep Neural Network implementations.

The nuclear spin systems in CdTe/(Cd,Zn)Te and CdTe/(Cd,Mg)Te quantum wells (QW) are studied using a multistage technique combining optical pumping and Hanle effect-based detection. The samples demonstrate drastically different nuclear spin dynamics in zero and weak magnetic fields. In CdTe/(Cd,Zn)Te, the nuclear spin relaxation time is found to strongly increase with the magnetic field, growing from 3 s in zero field to tens of seconds in a field of 25 G. In CdTe/(Cd,Mg)Te the relaxation is an order of magnitude slower, and it is field-independent up to at least 70 G. The differences are attributed to the nuclear spin relaxation being mediated by different kinds of resident electrons in these QWs. In CdTe/(Cd,Mg)Te, a residual electron gas trapped in the QW largely determines the relaxation dynamics. In CdTe/(Cd,Zn)Te, the fast relaxation in zero field is due to interaction with localized donor-bound electrons. Nuclear spin diffusion barriers form around neutral donors when the external magnetic field exceeds the local nuclear field, which is about $B_L\approx$0.4 G in CdTe. This inhibits nuclear spin diffusion towards the donors, slowing down relaxation. These findings are supported by theoretical modeling. In particular, we show that the formation of the diffusion barrier is made possible by several features specific to CdTe: (i) the large donor binding energy (about 10 meV), (ii) the low abundance of magnetic isotopes (only $\approx$30 % of nuclei have nonzero spin), and (iii) the absence of nuclear quadrupole interactions between nuclei. The two latter properties are also favorable to nuclear spin cooling via optical pumping followed by adiabatic demagnetization. Under non-optimized conditions we have reached sub-microkelvin nuclear spin temperatures in both samples, lower than all previous results obtained in GaAs., Comment: 12 pages, 9 figures

Clocks are a central part of many computing paradigms, and are mainly used to synchronise the delicate operation of switching, necessary to drive modern computational processes. Unfortunately, this synchronisation process is reaching a natural ``apocalypse''. No longer can clock scaling be used as a blunt tool to accelerate computation, we are up against the natural limits of switching and synchronisation across large processors. Therefore, we need to rethink how time is utilised in computation, using it more naturally in the role of representing data. This can be achieved by using a time interval delineated by discrete start and end events, and by re-casting computational operations into the time domain. With this, computer systems can be developed that are naturally scaleable in time and space, and can use ambient time references built to the best effort of the available technology. Our ambition is to better manage the energy/computation time trade-off, and to explicitly embed the resolution of the data in the time domain. We aim to recast calculations into the ``for free'' format that time offers, and in addition, perform these calculations at the highest clock or oscillator resolution possible.

A prototype cryomodule containing eight Single Spoke Resonators type-1 (SSR1) operating at 325 MHz and four superconducting focusing lenses was successfully assembled, cold tested, and accelerated beam in the framework of the PIP-II project at Fermilab. The impact of induced residual magnetic fields from the solenoids on performance of cavities is presented in this contribution. In addition, design optimizations for the production cryomodules as a result of this impact are highlighted., Comment: SRF2023 Conference

Lead halide perovskite nanocrystals (NCs) in a glass matrix combine excellent optical properties and stability against environment. The spectral and temporal characteristics of photoluminescence from CsPbBr$_3$ and CsPb(Cl,Br)$_3$ nanocrystals (NCs) in a fluorophosphate glass matrix are measured in a temperature range from 6 to 270 K in order to reveal factors that determine their quantum yield and recombination dynamics. At low temperatures, the recombination dynamics is characterized by three decay components with time scales on the order of 1 ns, 10 ns, and 1 $\mu$s. The relative contributions of the corresponding processes and their characteristic times are strongly temperature dependent. The emission intensity decreases with growing temperature. This effect is stronger in smaller NCs, which highlights the role of surface states. These experimental results are discussed on the basis of a model taking into account the NC energy structure and the presence of electron and hole surface trap states. The photoluminescence dynamics at low temperatures is dominated by charge-carrier radiative recombination and relaxation to shallow traps. At temperatures exceeding 100 K, the dynamics is affected by carrier activation to the excited states., Comment: 12 pages, 3 figures

The Multi-Agent Pathfinding (MAPF) problem involves finding a set of conflict-free paths for a group of agents confined to a graph. In typical MAPF scenarios, the graph and the agents' starting and ending vertices are known beforehand, allowing the use of centralized planning algorithms. However, in this study, we focus on the decentralized MAPF setting, where the agents may observe the other agents only locally and are restricted in communications with each other. Specifically, we investigate the lifelong variant of MAPF, where new goals are continually assigned to the agents upon completion of previous ones. Drawing inspiration from the successful AlphaZero approach, we propose a decentralized multi-agent Monte Carlo Tree Search (MCTS) method for MAPF tasks. Our approach utilizes the agent's observations to recreate the intrinsic Markov decision process, which is then used for planning with a tailored for multi-agent tasks version of neural MCTS. The experimental results show that our approach outperforms state-of-the-art learnable MAPF solvers. The source code is available at https://github.com/AIRI-Institute/mats-lp., Comment: The paper is accepted to AAAI-2024 conference

Photonic computing has recently become an interesting paradigm for high-speed calculation of computing processes using light-matter interactions. Here, we propose and study an electromagnetic wave-based structure with the ability to calculate the solution of partial differential equations in the form of the Helmholtz wave equation. To do this, we make use of a network of interconnected waveguides filled with dielectric inserts. In so doing, it is shown how the proposed network can mimic the response of a network of T-circuit elements formed by two series and a parallel impedance, i.e., the waveguide network effectively behaves as a metatronic network. An in-depth theoretical analysis of the proposed metatronic structure is presented showing how the governing equation for the currents and impedances of the metatronic network resembles that of the finite difference representation of the Helmholtz wave equation. Different studies are then discussed including the solution of partial differential equations for Dirichlet and open boundary value problems, demonstrating how the proposed metatronic-based structure has the ability to calculate their solutions., Comment: 23 pages, 4 figures, 69 references

We consider an Anonymous Multi-Agent Path-Finding (AMAPF) problem where the set of agents is confined to a graph, a set of goal vertices is given and each of these vertices has to be reached by some agent. The problem is to find an assignment of the goals to the agents as well as the collision-free paths, and we are interested in finding the solution with the optimal makespan. A well-established approach to solve this problem is to reduce it to a special type of a graph search problem, i.e. to the problem of finding a maximum flow on an auxiliary graph induced by the input one. The size of the former graph may be very large and the search on it may become a bottleneck. To this end, we suggest a specific search algorithm that leverages the idea of exploring the search space not through considering separate search states but rather bulks of them simultaneously. That is, we implicitly compress, store and expand bulks of the search states as single states, which results in high reduction in runtime and memory. Empirically, the resultant AMAPF solver demonstrates superior performance compared to the state-of-the-art competitor and is able to solve all publicly available MAPF instances from the well-known MovingAI benchmark in less than 30 seconds., Comment: Accepted at AAAI24

Magnetars form a special class of neutron stars possessing superstrong magnetic fields and demonstrating power flares triggered likely by these fields. Observations of such flares reveal the presence of quasi-periodic oscillations (QPOs) at certain frequencies; they are thought to be excited in the flares. QPOs carry potentially important information on magnetar structure, magnetic field, and mechanisms of magnetar activity. We calculate frequencies of torsional (magneto-elastic) oscillations of the magnetar crust treating the magnetic field effects in the first order of perturbation theory. The theory predicts splitting of non-magnetic oscillation frequencies into Zeeman components. Zeeman splitting of torsional oscillation spectrum of magnetars was suggested, clearly described and estimated by Shaisultanov and Eichler (2009) but their work has not been given considerable attention. To extend it we suggest the technique of calculating oscillation frequencies including Zeeman splitting at not too strong magnetic fields for arbitrary magnetic field configuration. Zeeman splitting enriches the oscillation spectrum and simplifies theoretical interpretation of observations. We calculate several low-frequency oscillations of magnetars with pure dipole magnetic field in the crust. The results qualitatively agree with low-frequency QPOs detected in the hyperflare of SGR 1806--20, and in the giant flare of SGR 1900+14., Comment: 15 pages, 3 figures, typos in eqs. (26) and (27) corrected

We study the near threshold behavior of cross sections of low-energy antiproton scattering off the ground and excited states of positronium with zero total orbital momentum $L=0$. In our computational experiment, the existence of singularities called the Gailitis-Damburg oscillations above the thresholds of excited states of positronium and antihydrogen atoms is confirmed. In the future the obtained results can be useful for developing proposals for improving the conditions of experiments with antimatter., Comment: 7 pages, 4 figures

Superbursts of neutron stars are rare but powerful events explained by the explosive burning of carbon in the deep layers of the outer envelope of the star. In this paper we perform a simulation of superbursts and propose a simple method for describing the neutrino stage of their cooling, as well as a method for describing the evolution of the burst energy on a scale of several months. We note a universal relation for the temperature distribution in the burnt layer at its neutrino cooling stage, as well as the unification of bolometric light curves and neutrino heat loss rates for deep and powerful bursts. We point out the possibility of long-term retention of the burst energy in the star's envelope. The results can be useful for interpretation of superburst observations., Comment: 8 pages, 6 figures, to be published in Astronomy Letters, 49, No.12 (2023)