arXiv subjects

Mykola Maksymenko

Publications and source records attributed to Mykola Maksymenko.

13 recordsLinked to original sources

arXiv2026

Demonstration of a Structured Agentic Workflow for Applied Quantum Computing Research

Quantum hardware advances now make it practical to explore quantum approaches to scientific problems. Applied quantum computing research sits at the intersection of domain knowledge, algorithms, theoretical physics, device physics, and quantum and classical computational methods. The breadth of expertise required and the inherent project complexity challenge not only any individual researcher, but also generic AI systems. Here, we introduce the Quantum Agentic Operating System (QAOS), a structured agentic research system designed to address this challenge for quantum and computational R&D projects. QAOS is built around a nonlinear workflow graph of heterogeneous multi-agent nodes, with a controlled flow of context and computational resources. It decomposes a complex research project into sandboxed multi-agent tasks executed under constraints from the project graph, task-relevant methodologies, and necessary verification steps. We demonstrate QAOS on a quantum chemistry prototyping task, in which it is asked to set up a minimal model for a proton-transfer problem and to implement its solution using sample-based quantum diagonalization (SQD), together with classical reference methods such as FCI (exact diagonalization) and density functional theory (DFT). We compare the QAOS solution to those obtained with baseline agents (Claude Sonnet 5 with the Claude Code harness). In contrast to QAOS, the baseline agents silently ignore explicit instructions, and the runs show large variability.

quant-ph↗

arXiv2026

Circuit Depth Reduction for Executable Hamiltonian Dynamics of Covalent Inhibitor Reactivity on Quantum Hardware

Quantum chemistry applications in the noisy intermediate-scale quantum era require end-to-end approaches that balance algorithmic fidelity with practical executability on existing hardware. We present an end-to-end Hamiltonian dynamics case study for predicting the reactivity of pharmaceutically relevant covalent inhibitors containing sulfonyl fluoride warheads, using a quantum-centric data-driven research and development framework that combines Hamiltonian time evolution with classical machine learning. To make such simulations executable on current quantum processors, we introduce a systematic circuit reduction strategy based on Hamiltonian term truncation with observable error bounds, Clifford Decomposition and Transformation, and hardware-aware transpilation. Across representative molecular fragments, this approach achieves circuit depth reductions of up to 28.5x under all-to-all connectivity assumptions and up to 15.5x on IBM Heron-class architectures. For an eight-qubit Hamiltonian dynamics simulation, a transpiled instruction set architecture (ISA) circuit depth of 1330 is rendered executable through middleware-enabled circuit decomposition, enabling the execution of sub-circuits with depths up to 371 and containing up to 216 two-qubit gates on real hardware. We evaluate the impact of circuit reduction on downstream reactivity prediction accuracy and show that chemically meaningful predictions can be retained despite aggressive circuit simplifications, clarifying the trade-offs that govern practical quantum chemistry workflows on near-term quantum systems.

quant-ph↗

arXiv2026

Entanglement scaling in matrix product state representation of smooth functions and their shallow quantum circuit approximations

Encoding classical data in a quantum state is a key prerequisite of many quantum algorithms. Recently matrix product state (MPS) methods emerged as the most promising approach for constructing shallow quantum circuits approximating input functions, including probability distributions, with only linear number of gates. We derive rigorous asymptotic expansions for the decay of entanglement across bonds in the MPS representation depending on the smoothness of the input function, real or complex. We also consider the dependence of the entanglement on localization properties and function support. Based on these analytical results we construct an improved MPS-based algorithm yielding shallow and accurate encoding quantum circuits. By using Tensor Cross Interpolation we are able to construct utility-scale quantum circuits in a compute- and memory-efficient way. We validate our methods by loading heavy-tailed distributions, including Levy, important in finance, but they apply to any smooth function inputs. We test the performance of the resulting quantum circuits by executing and sampling from them on IBM quantum devices, for up to 156 qubits.

quant-ph↗

arXiv2022

Towards realistic symmetry-based completion of previously unseen point clouds

3D scanning is a complex multistage process that generates a point cloud of an object typically containing damaged parts due to occlusions, reflections, shadows, scanner motion, specific properties of the object surface, imperfect reconstruction algorithms, etc. Point cloud completion is specifically designed to fill in the missing parts of the object and obtain its high-quality 3D representation. The existing completion approaches perform well on the academic datasets with a predefined set of object classes and very specific types of defects; however, their performance drops significantly in the real-world settings and degrades even further on previously unseen object classes. We propose a novel framework that performs well on symmetric objects, which are ubiquitous in man-made environments. Unlike learning-based approaches, the proposed framework does not require training data and is capable of completing non-critical damages occurring in customer 3D scanning process using e.g. Kinect, time-of-flight, or structured light scanners. With thorough experiments, we demonstrate that the proposed framework achieves state-of-the-art efficiency in point cloud completion of real-world customer scans. We benchmark the framework performance on two types of datasets: properly augmented existing academic dataset and the actual 3D scans of various objects.

cs.CV↗

arXiv2020

ActGAN: Flexible and Efficient One-shot Face Reenactment

This paper introduces ActGAN - a novel end-to-end generative adversarial network (GAN) for one-shot face reenactment. Given two images, the goal is to transfer the facial expression of the source actor onto a target person in a photo-realistic fashion. While existing methods require target identity to be predefined, we address this problem by introducing a "many-to-many" approach, which allows arbitrary persons both for source and target without additional retraining. To this end, we employ the Feature Pyramid Network (FPN) as a core generator building block - the first application of FPN in face reenactment, producing finer results. We also introduce a solution to preserve a person's identity between synthesized and target person by adopting the state-of-the-art approach in deep face recognition domain. The architecture readily supports reenactment in different scenarios: "many-to-many", "one-to-one", "one-to-another" in terms of expression accuracy, identity preservation, and overall image quality. We demonstrate that ActGAN achieves competitive performance against recent works concerning visual quality.

cs.CV↗

arXiv2020

Mid-Air Haptic Bio-Holograms in Mixed Reality

We present a prototype demonstrator that integrates three technologies, mixed reality head-mounted displays, wearable bio-sensors, and mid-air haptic projectors to deliver an interactive tactile experience with a bio-hologram. Users of this prototype are able to see, touch and feel a hologram of a heart that is beating at the same rhythm as their own. The demo uses an Ultrahaptics device, a Magic Leap One Mixed Reality headset, and an Apple Watch that measures the wearer's heart rate, all synchronized and networked together such that updates from the wristband dynamically change the haptic feedback and the animation speed of the beating heart thus creating a more personalised experience.

cs.HC↗

arXiv2019

Training Deep Neural Networks by optimizing over nonlocal paths in hyperparameter space

Hyperparameter optimization is both a practical issue and an interesting theoretical problem in training of deep architectures. Despite many recent advances the most commonly used methods almost universally involve training multiple and decoupled copies of the model, in effect sampling the hyperparameter space. We show that at a negligible additional computational cost, results can be improved by sampling nonlocal paths instead of points in hyperparameter space. To this end we interpret hyperparameters as controlling the level of correlated noise in training, which can be mapped to an effective temperature. The usually independent instances of the model are coupled and allowed to exchange their hyperparameters throughout the training using the well established parallel tempering technique of statistical physics. Each simulation corresponds then to a unique path, or history, in the joint hyperparameter/model-parameter space. We provide empirical tests of our method, in particular for dropout and learning rate optimization. We observed faster training and improved resistance to overfitting and showed a systematic decrease in the absolute validation error, improving over benchmark results.

cs.LG↗

arXiv2017

Persistence of the flat band in a kagome magnet with dipolar interactions

The weathervane modes of the classical Heisenberg antiferromagnet on the kagome lattice constitute possibly the earliest and certainly the most celebrated example of a flat band of zero-energy excitations. Such modes arise from the underconstraint that has since become a defining criterion of strong geometrical frustration. We investigate the fate of this flat band when dipolar interactions are added. These change the nearest-neighbour model fundamentally as they remove the Heisenberg spin-rotational symmetry while also introducing a long- range component to the interaction. We explain how the modes continue to remain approximately dispersionless, while being lifted to finite energy as well as being squeezed: they change their ellipticity described by the ratio of the amplitudes of the canonically conjugate variables comprising them. This phenomenon provides interesting connections between concepts such as constraint counting and self-screening underpinning the field of frustrated magnetism. We discuss variants of these phenomena for different interactions, lattices and dimension.

cond-mat.stat-mech↗

arXiv2016

Dynamics of a Many-Body-Localized System Coupled to a Bath

Coupling a many-body-localized system to a dissipative bath necessarily leads to delocalization. Here, we investigate the nature of the ensuing relaxation dynamics and the information it holds on the many-body-localized state. We formulate the relevant Lindblad equation in terms of the local integrals of motion of the underlying localized Hamiltonian. This allows to map the quantum evolution deep in the localized state to tractable classical rate equations. We consider two different types of dissipation relevant to systems of ultra-cold atoms: dephasing due to inelastic scattering on the lattice lasers and particle loss. Only the latter mechanism shows a pronounced effects of interactions on the relaxation of observables.

cond-mat.str-el↗

arXiv2015

Strongly correlated flat-band systems: The route from Heisenberg spins to Hubbard electrons

In this review we recapitulate the basic features of the flat-band spin systems and briefly summarize earlier studies in the field. Main emphasis is made on recent developments which include results for both spin and electron flat-band models. In particular, for flat-band spin systems we highlight field-driven phase transitions for frustrated quantum Heisenberg antiferromagnets at low temperatures, chiral flat-band states, as well as the effect of a slight dispersion of a previously strictly flat band due to nonideal lattice geometry. For electronic systems, we discuss the universal low-temperature behavior of several flat-band Hubbard models, the emergence of ground-state ferromagnetism in the square-lattice Tasaki-Hubbard model and the related Pauli-correlated percolation problem, as well as the dispersion-driven ground-state ferromagnetism in flat-band Hubbard systems. Closely related studies and possible experimental realizations of the flat-band physics are also described briefly.

cond-mat.str-el↗

arXiv2015

Classical dipoles on the kagome lattice

Motivated by recent developments in magnetic materials, frustrated nanoarrays and cold atomic systems, we investigate the behaviour of dipolar spins on the frustrated two-dimensional kagome lattice. By combining the Luttinger-Tisza approach, numerical energy minimization, spin-wave analysis and parallel tempering Monte-Carlo, we study long-range ordering and finite-temperature phase transitions for a Hamiltonian containing both dipolar and nearest-neighbor interactions. For both weak and moderate dipolar interactions, the system enters a three-sublattice long-range ordered state, with each triangle having vanishing dipole and quadrupole moments; while for dominating dipolar interactions we uncover ferrimagnetic three-sublattice order. These are also the ground states for XY spins. We discuss excitations of, as well as phase transitions into, these states. We find behaviour consistent with Ising criticality for the 120-degree state, while the ferrimagnetic state appears to be associated with drifting exponents. The celebrated flat band of zero-energy excitations of the kagome nearest-neighbour Heisenberg model is lifted to finite energies but acquires only minimal dispersion as dipolar interactions are added.

cond-mat.str-el↗

arXiv2014

Reversible first-order transition in Pauli percolation

Percolation plays an important role in fields and phenomena as diverse as the study of social networks, the dynamics of epidemics, the robustness of electricity grids, conduction in disordered media, and geometric properties in statistical physics. We analyse a new percolation problem in which the first order nature of an equilibrium percolation transition can be established analytically and verified numerically. The rules for this site percolation model are physical and very simple, requiring only the introduction of a weight $W(n)=n+1$ for a cluster of size $n$. This establishes that a discontinuous percolation transition can occur with qualitatively more local interactions than in all currently considered examples of explosive percolation; and that, unlike these, it can be reversible. This greatly extends both the applicability of such percolation models in principle, and their reach in practice.

cond-mat.stat-mech↗

arXiv2011

Localized states on triangular traps and low-temperature properties of the antiferromagnetic Heisenberg and repulsive Hubbard models

We consider the antiferromagnetic Heisenberg and the repulsive Hubbard model on two $N$-site one-dimensional lattices, which support dispersionless one-particle states corresponding to localized states on triangular trapping cells. We calculate the degeneracy of the ground states in the subspaces with $n\le n_{\max}$, $n_{\max}\propto N$ magnons or electrons as well as the contribution of these states (independent localized states) to thermodynamic quantities. Moreover, we discuss another class of low-lying eigenstates (so-called interacting localized states) and calculate their contribution to the partition function. We also discuss the effect of extra interactions, which lift the degeneracy present due to the chirality of the localized states on triangles. The localized states set an extra low-energy scale in the system and lead to a nonzero residual ground-state entropy and to one (or more) additional low-temperature peak(s) in the specific heat. Low-energy degrees of freedom in the presence of perturbations removing degeneracy owing to the chirality can be described in terms of an effective (pseudo)spin-1/2 transverse $XX$ chain.

cond-mat.str-el↗