Searcharxiv⌕ Search

arXiv subjects

Ayana Ghosh

Publications and source records attributed to Ayana Ghosh.

At least 19 recordsLinked to original sources
arXiv2026

Symmetry-electronic fingerprints reveal competing magnetic phases in two-dimensional materials

Two-dimensional magnets offer compelling platforms for spintronics and quantum technologies, yet predicting their magnetic ground states, moments, and anisotropy remains challenging. This limitation primarily arises because existing machine-learning representations encode chemical environments without capturing the symmetry or exchange physics that govern magnetism. In this work, we introduce the symmetry-electronic fingerprint (SEF), a physically interpretable representation that encodes crystallographic symmetry operations, Wyckoff-site geometry, together with site-resolved electronic structure. Combined with ensemble learning with random forests, the SEF accurately classifies magnetic ordering while regressing moments alongside anisotropy energies while simultaneously resolving the distinct regimes of itinerant Stoner ferromagnetism from localized superexchange. What sets the SEF-trained models apart is that regions of elevated model uncertainty are not a failure but a diagnostic, identifying materials where these mechanisms compete. First-principles calculations on Co- and Ni-based halides and oxides confirm that these regions correspond to genuine near-degenerate FM and AFM phases with magnetic frustration, suppressed anisotropy, and emergent non-collinear ordering. By encoding symmetry together with exchange physics directly into the representation unlike conventional descriptors, the SEF transforms model uncertainty into a compass pointing toward two-dimensional materials where small perturbations drive transitions between collinear, frustrated, or non-collinear magnetic phases.

cond-mat.mtrl-sci↗
arXiv2026

Towards "on-demand" van der Waals epitaxy with adaptive ensemble sampling atomistic workflows

Traditional approaches to achieve targeted epitaxial growth involves exploring a vast parameter space of thermodynamical and kinetic drivers (e.g., temperature, pressure, chemical potential etc). This tedious and time-consuming approach becomes particularly cumbersome to accelerate synthesis and characterization of novel materials with complex dependencies on local chemical environment, temperature and lattice-strains, specifically nanoscale heterostructures of layered 2D materials. We combine the strength of next generation supercomputers at the extreme scale, machine learning and classical molecular dynamics simulations within an adaptive real time closed-loop virtual environment steered by Bayesian optimization to enable asynchronous ensemble sampling of the synthesis space, and apply it to the recrystallization phenomena of amorphous transition-metal dichalcogenide (TMDC) bilayer to form stack moiré heterostructures under various growth parameters. We show that such asynchronous ensemble sampling frameworks for materials simulations can be promising towards achieving on-demand epitaxy of van der Waals stacked moiré devices, paving the way towards a robust autonomous materials synthesis pipeline to enable unprecedented discovery of new functionalities.

cond-mat.mtrl-sci↗
arXiv2026

Born-Qualified: An Autonomous Framework for Deploying Advanced Energy and Electronic Materials

Autonomous science is transforming how we discover materials and chemical systems for advanced energy technologies. However, many initially promising systems never reach deployment. This "valley of death" stems from optimization that prioritizes laboratory metrics over industrial viability. We propose a new strategy: "born-qualified" autonomous development, which embeds manufacturability, cost, and durability constraints from the outset. This approach is enabled by four pillars, including the development of multi-objective metrics, causal models, a modular infrastructure, and embedding manufacturing in the discovery loop. Realizing this vision will require sustained, community-wide commitment, but the potential return on that investment is commensurate with the scale of the challenge.

cond-mat.mtrl-sci↗
arXiv2025

Intervention Strategies for Polarization Switching in Hybrid Improper Ferroelectrics

The potential of hybrid improper ferroelectrics (HIFs) in electronic and spintronic devices hinges on their ability to switch polarization. Although the coupling between octahedral rotation and tilt is well established, the factors that govern switching barriers remain elusive. In this study, we explore this area to demonstrate the critical role of causal reasoning in uncovering the mechanisms to control the ferroelectric switching barrier in HIFs. By combining causal discovery, causal interventions, and first-principles simulations, we identify tolerance factor, A-site cation radii mismatch, epitaxial strain, and octahedral rotation/tilt as key parameters and quantify how their interplay directly influences switching barrier. Three key insights emerge from our work: (a) the analysis identifies the structural descriptors controlling polarization reversal across a broad family of A-site-layered double perovskites and superlattices, (b) it uncovers non-trivial, material-specific rotation-tilt mechanisms, including a counterintuitive cooperative pathway where both rotation and tilt change while lowering the barrier, an effect mostly inaccessible to conventional Landau or first-principles-based approaches and (c) it maps these material-specific mechanisms to experimentally realizable parameters, showing that epitaxial strain from orthorhombic substrates (e.g., NdScO$_3$, NdGaO$_3$) selectively tunes octahedral distortions to achieve barrier reduction across varied compositions. These results establish actionable, materials-by-design principles linking composition, structure, and strain to polarization switching, while highlighting the potential of causal reasoning to guide intelligent, mechanism-driven strategies for engineering complex functional oxides.

cond-mat.mtrl-sci↗
arXiv2025

Operationalizing Serendipity: Multi-Agent AI Workflows for Enhanced Materials Characterization with Theory-in-the-Loop

The history of science is punctuated by serendipitous discoveries, where unexpected observations, rather than targeted hypotheses, opened new fields of inquiry. While modern autonomous laboratories excel at accelerating hypothesis testing, their optimization for efficiency risks overlooking these crucial, unplanned findings. To address this gap, we introduce SciLink, an open-source, multi-agent artificial intelligence framework designed to operationalize serendipity in materials research by creating a direct, automated link between experimental observation, novelty assessment, and theoretical simulations. The framework employs a hybrid AI strategy where specialized machine learning models perform quantitative analysis of experimental data, while large language models handle higher-level reasoning. These agents autonomously convert raw data from materials characterization techniques into falsifiable scientific claims, which are then quantitatively scored for novelty against the published literature. We demonstrate the framework's versatility across diverse research scenarios, showcasing its application to atomic-resolution and hyperspectral data, its capacity to integrate real-time human expert guidance, and its ability to close the research loop by proposing targeted follow-up experiments. By systematically analyzing all observations and contextualizing them, SciLink provides a practical framework for AI-driven materials research that not only enhances efficiency but also actively cultivates an environment ripe for serendipitous discoveries, thereby bridging the gap between automated experimentation and open-ended scientific exploration.

cs.AI↗
arXiv2025

Active Deep Kernel Learning of Molecular Properties: Realizing Dynamic Structural Embeddings

As vast databases of chemical identities become increasingly available, the challenge shifts to how we effectively explore and leverage these resources to study molecular properties. This paper presents an active learning approach for molecular discovery using Deep Kernel Learning (DKL), demonstrated on the QM9 dataset. DKL links structural embeddings directly to properties, creating organized latent spaces that prioritize relevant property information. By iteratively recalculating embedding vectors in alignment with target properties, DKL uncovers concentrated maxima representing key molecular properties and reveals unexplored regions with potential for innovation. This approach underscores DKL's potential in advancing molecular research and discovery.

cs.LG↗
arXiv2025

Autonomous Fabrication of Tailored Defect Structures in 2D Materials using Machine Learning-enabled Scanning Transmission Electron Microscopy

Materials with tailored quantum properties can be engineered from atomic scale assembly techniques, but existing methods often lack the agility and accuracy to precisely and intelligently control the manufacturing process. Here we demonstrate a fully autonomous approach for fabricating atomic-level defects using electron beams in scanning transmission electron microscopy (STEM) that combines advanced machine learning and automated beam control. As a proof of concept, we achieved controlled fabrication of MoS-nanowire (MoS-NW) edge structures by iterative and targeted exposure of $MoS_2$ monolayer to a focused electron beam to selectively eject sulfur atoms, utilizing high-angle annular dark-field (HAADF) imaging for feedback-controlled monitoring structural evolution of defects. A machine learning framework combining a random forest model and convolutional neural networks (CNN) was developed to decode the HAADF image and accurately identify atomic positions and species. This atomic-level information was then integrated into an autonomous decision-making platform, which applied predefined fabrication strategies to instruct beam control about atomic sites to be ejected. The selected sites were subsequently exposed to localized electron beam using an FPGA-controlled scan routine with precise control over beam positioning and duration. While the MoS-NW edge structures produced exhibit promising mechanical and electronic properties, the proposed autonomous fabrication framework is material-agnostic and can be extended to other 2D materials for the creation of diverse defect structures and heterostructures beyond $MoS_2$.

cond-mat.mtrl-sci↗
arXiv2024

Active Causal Learning for Decoding Chemical Complexities with Targeted Interventions

Predicting and enhancing inherent properties based on molecular structures is paramount to design tasks in medicine, materials science, and environmental management. Most of the current machine learning and deep learning approaches have become standard for predictions, but they face challenges when applied across different datasets due to reliance on correlations between molecular representation and target properties. These approaches typically depend on large datasets to capture the diversity within the chemical space, facilitating a more accurate approximation, interpolation, or extrapolation of the chemical behavior of molecules. In our research, we introduce an active learning approach that discerns underlying cause-effect relationships through strategic sampling with the use of a graph loss function. This method identifies the smallest subset of the dataset capable of encoding the most information representative of a much larger chemical space. The identified causal relations are then leveraged to conduct systematic interventions, optimizing the design task within a chemical space that the models have not encountered previously. While our implementation focused on the QM9 quantum-chemical dataset for a specific design task-finding molecules with a large dipole moment-our active causal learning approach, driven by intelligent sampling and interventions, holds potential for broader applications in molecular, materials design and discovery.

cs.LG↗
arXiv2023

Direct Fabrication of Atomically Defined Pores in MXenes

Controlled fabrication of nanopores in atomically thin two-dimensional material offers the means to create robust membranes needed for ion transport, nanofiltration, and DNA sensing. Techniques for creating nanopores have relied upon either plasma etching or direct irradiation using electrons or ions; however, aberration-corrected scanning transmission electron microscopy (STEM) offers the advantage of combining a highly energetic, sub-angstrom sized electron beam for atomic manipulation along with atomic resolution imaging. Here, we utilize a method for automated nanopore fabrication with real-time atomic visualization to enhance our mechanistic understanding of beam-induced transformations. Additionally, an electron beam simulation technique, Electron-Beam Simulator (E-BeamSim) was developed to observe the atomic movements and interactions resulting from electron beam irradiation. Using the 2D MXene Ti3C2Tx, we explore the influence of temperature on nanopore fabrication by tracking atomic transformation pathways and find that at room temperature, electron beam irradiation induces random displacement of atoms and results in a pileup of titanium atoms at the nanopore edge. This pileup was confirmed and demonstrated in E-BeamSim simulations around the small, milled area in the MXene monolayer. At elevated temperatures, the surface functional groups on MXene are effectively removed, and the mobility of atoms increases, which results in atomic transformations that lead to the selective removal of atoms layer by layer. Through controllable manufacture using e-beam milling fabrication, the production and then characterization of the fabricated defects can be better understood for future work. This work can lead to the development of defect engineering techniques within functionalized MXene layers.

cond-mat.mtrl-sci↗
arXiv2023

When the atoms dance: exploring mechanisms of electron-beam induced modifications of materials with machine-learning assisted high temporal resolution electron microscopy

Directed atomic fabrication using an aberration-corrected scanning transmission electron microscope (STEM) opens new pathways for atomic engineering of functional materials. In this approach, the electron beam is used to actively alter the atomic structure through electron beam induced irradiation processes. One of the impediments that has limited widespread use thus far has been the ability to understand the fundamental mechanisms of atomic transformation pathways at high spatiotemporal resolution. Here, we develop a workflow for obtaining and analyzing high-speed spiral scan STEM data, up to 100 fps, to track the atomic fabrication process during nanopore milling in monolayer MoS2. An automated feedback-controlled electron beam positioning system combined with deep convolution neural network (DCNN) was used to decipher fast but low signal-to-noise datasets and classify time-resolved atom positions and nature of their evolving atomic defect configurations. Through this automated decoding, the initial atomic disordering and reordering processes leading to nanopore formation was able to be studied across various timescales. Using these experimental workflows a greater degree of speed and information can be extracted from small datasets without compromising spatial resolution. This approach can be adapted to other 2D materials systems to gain further insights into the defect formation necessary to inform future automated fabrication techniques utilizing the STEM electron beam.

cond-mat.mtrl-sci↗
arXiv2023

Identification of novel organic ferroelectrics: A study combining importance sampling with machine learning

Recent advances in the synthesis of polar molecular materials have produced practical alternatives to ferroelectric ceramics, opening up exciting new avenues for their incorporation into modern electronic devices. However, in order to realize the full potential of polar polymer and molecular crystals for modern technological applications, it is paramount to assemble and evaluate all the available data for such compounds, identifying descriptors that could be associated with an emergence of ferroelectricity. In this work, we utilized data-driven approaches to judiciously shortlist candidate materials from a wide chemical space that could possess ferroelectric functionalities. An importance-sampling based method was utilized to address the challenge of having a limited amount of available data on already known organic ferroelectrics. Sets of molecular- and crystal-level descriptors were combined with a Random Forest Regression algorithm in order to predict spontaneous polarization of the shortlisted compounds. First-principles simulations were performed to further validate the predictions obtained from the machine learning model.

cond-mat.mtrl-sci↗
arXiv2023

Bending-induced isostructural transitions in ultrathin layers of van der Waals ferrielectrics

Using Landau-Ginzburg-Devonshire (LGD) phenomenological approach we analyze the bending-induced re-distribution of electric polarization and field, elastic stresses and strains inside ultrathin layers of van der Waals ferrielectrics. We consider a CuInP2S6 (CIPS) thin layer with fixed edges and suspended central part, the bending of which is induced by external forces. The unique aspect of CIPS is the existence of two ferrielectric states, FI1 and FI2, corresponding to big and small polarization values, which arise due to the specific four-well potential of the eighth-order LGD functional. When the CIPS layer is flat, the single-domain FI1 state is stable in the central part of the layer, and the FI2 states are stable near the fixed edges. With an increase of the layer bending below the critical value, the sizes of the FI2 states near the fixed edges decreases, and the size of the FI1 region increases. When the bending exceeds the critical value, the edge FI2 states disappear being substituted by the FI1 state, but they appear abruptly near the inflection regions and expand as the bending increases. The bending-induced isostructural FI1-FI2 transition is specific for the bended van der Waals ferrielectrics described by the eighth (or higher) order LGD functional with consideration of linear and nonlinear electrostriction couplings. The isostructural transition, which is revealed in the vicinity of room temperature, can significantly reduce the coercive voltage of ferroelectric polarization reversal in CIPS nanoflakes, allowing for the curvature-engineering control of various flexible nanodevices.

cond-mat.mtrl-sci↗
arXiv2023

Disentangling stress and curvature effects in layered 2D ferroelectric CuInP2S6

Nanoscale ferroelectric 2D materials offer unique opportunity to investigate curvature and strain effects on materials functionalities. Among these, CuInP2S6 (CIPS) has attracted tremendous research interest in recent years due to combination of room temperature ferroelectricity, scalability to a few layers thickness, and unique ferrielectric properties due to coexistence of 2 polar sublattices. Here, we explore the local curvature and strain effect on the polarization in CIPS via piezoresponse force microscopy and spectroscopy. To explain the observed behaviors and decouple the curvature and strain effects in 2D CIPS, we introduce finite element Landau-Ginzburg-Devonshire model. The results show that bending induces ferrielectric domains in CIPS, and the polarization-voltage hysteresis loops differ in bending and non-bending regions. Our simulation indicates that the flexoelectric effect can affect local polarization hysteresis. These studies open a novel pathway for the fabrication of curvature-engineered nanoelectronic devices.

cond-mat.mtrl-sci↗
arXiv2023

Discovery of structure-property relations for molecules via hypothesis-driven active learning over the chemical space

Discovery of the molecular candidates for applications in drug targets, biomolecular systems, catalysts, photovoltaics, organic electronics, and batteries, necessitates development of machine learning algorithms capable of rapid exploration of the chemical spaces targeting the desired functionalities. Here we introduce a novel approach for the active learning over the chemical spaces based on hypothesis learning. We construct the hypotheses on the possible relationships between structures and functionalities of interest based on a small subset of data and introduce them as (probabilistic) mean functions for the Gaussian process. This approach combines the elements from the symbolic regression methods such as SISSO and active learning into a single framework. The primary focus of constructing this framework is to approximate physical laws in an active learning regime toward a more robust predictive performance, as traditional evaluation on hold-out sets in machine learning doesn't account for out-of-distribution effects and may lead to a complete failure on unseen chemical space. Here, we demonstrate it for the QM9 dataset, but it can be applied more broadly to datasets from both domains of molecular and solid-state materials sciences.

cs.LG↗
arXiv2023

Anomalous Polarization Reversal in Strained Thin Films of CuInP$_2$S$_6$

Strain-induced transitions of polarization reversal in thin films of a ferrielectric CuInP$_2$S$_6$ (CIPS) with ideally-conductive electrodes is explored using the Landau-Ginzburg-Devonshire (LGD) approach with an eighth-order free energy expansion in polarization powers. Due to multiple potential wells, the height and position of which are temperature- and strain-dependent, the energy profiles of CIPS can flatten in the vicinity of the non-zero polarization states. This behavior differentiates these materials from classical ferroelectrics with the first or second order ferroelectric-paraelectric phase transition, for which potential energy profiles can be shallow or flat near the transition point only, corresponding to zero spontaneous polarization. Thereby we reveal an unusually strong effect of the mismatch strain on the out-of-plane polarization reversal, hysteresis loops shape, dielectric susceptibility, and piezoelectric response of CIPS films. In particular, by varying the sign of the mismatch strain and its magnitude in a narrow range, quasi-static hysteresis-less paraelectric curves can transform into double, triple, and other types of pinched and single hysteresis loops. The strain effect on the polarization reversal is opposite, i.e., "anomalous", in comparison with many other ferroelectric films in that the out-of-plane remanent polarization and coercive field increases strongly for tensile strains, meanwhile the polarization decreases or vanish for compressive strains. We explain the effect by "inverted" signs of linear and nonlinear electrostriction coupling coefficients of CIPS and their strong temperature dependence. For definite values of temperature and mismatch strain, the low-frequency hysteresis loops of polarization may exhibit negative slope in the relatively narrow range of external field amplitude and frequency.

cond-mat.mtrl-sci↗
arXiv2023

Deep Learning for Automated Experimentation in Scanning Transmission Electron Microscopy

Machine learning (ML) has become critical for post-acquisition data analysis in (scanning) transmission electron microscopy, (S)TEM, imaging and spectroscopy. An emerging trend is the transition to real-time analysis and closed-loop microscope operation. The effective use of ML in electron microscopy now requires the development of strategies for microscopy-centered experiment workflow design and optimization. Here, we discuss the associated challenges with the transition to active ML, including sequential data analysis and out-of-distribution drift effects, the requirements for the edge operation, local and cloud data storage, and theory in the loop operations. Specifically, we discuss the relative contributions of human scientists and ML agents in the ideation, orchestration, and execution of experimental workflows and the need to develop universal hyper languages that can apply across multiple platforms. These considerations will collectively inform the operationalization of ML in next-generation experimentation.

cond-mat.mtrl-sci↗
arXiv2023

Designing Workflows for Materials Characterization

Experimental science is enabled by the combination of synthesis, imaging, and functional characterization. Synthesis of a new material is typically followed by a set of characterization methods aiming to provide feedback for optimization or discover fundamental mechanisms. However, the sequence of synthesis and characterization methods and their interpretation, or research workflow, has traditionally been driven by human intuition and is highly domain specific. Here we explore concepts of scientific workflows that emerge at the interface between theory, characterization, and imaging. We discuss the criteria by which these workflows can be constructed for special cases of multi-resolution structural imaging and structural and functional characterization. Some considerations for theory-experiment workflows are provided. We further pose that the emergence of user facilities and cloud labs disrupt the classical progression from ideation, orchestration, and execution stages of workflow development and necessitate development of universal frameworks for workflow design, including universal hyper-languages describing laboratory operation, reward functions and their integration between domains, and policy development for workflow optimization. These tools will enable knowledge-based workflow optimization, enable lateral instrumental networks, sequential and parallel orchestration of characterization between dissimilar facilities, and empower distributed research.

cond-mat.mtrl-sci↗
arXiv2022

Microscopy is All You Need

We pose that microscopy offers an ideal real-world experimental environment for the development and deployment of active Bayesian and reinforcement learning methods. Indeed, the tremendous progress achieved by machine learning (ML) and artificial intelligence over the last decade has been largely achieved via the utilization of static data sets, from the paradigmatic MNIST to the bespoke corpora of text and image data used to train large models such as GPT3, DALLE and others. However, it is now recognized that continuous, minute improvements to state-of-the-art do not necessarily translate to advances in real-world applications. We argue that a promising pathway for the development of ML methods is via the route of domain-specific deployable algorithms in areas such as electron and scanning probe microscopy and chemical imaging. This will benefit both fundamental physical studies and serve as a test bed for more complex autonomous systems such as robotics and manufacturing. Favorable environment characteristics of scanning and electron microscopy include low risk, extensive availability of domain-specific priors and rewards, relatively small effects of exogeneous variables, and often the presence of both upstream first principles as well as downstream learnable physical models for both statics and dynamics. Recent developments in programmable interfaces, edge computing, and access to APIs facilitating microscope control, all render the deployment of ML codes on operational microscopes straightforward. We discuss these considerations and hope that these arguments will lead to creating a novel set of development targets for the ML community by accelerating both real-world ML applications and scientific progress.

cond-mat.dis-nn↗