arXiv subjects

Abhinav Prakash

Publications and source records attributed to Abhinav Prakash.

At least 19 recordsLinked to original sources

arXiv2025

Sample-Efficient Language Model for Hinglish Conversational AI

This paper presents our process for developing a sample-efficient language model for a conversational Hinglish chatbot. Hinglish, a code-mixed language that combines Hindi and English, presents a unique computational challenge due to inconsistent spelling, lack of standardization, and limited quality of conversational data. This work evaluates multiple pre-trained cross-lingual language models, including Gemma3-4B and Qwen2.5-7B, and employs fine-tuning techniques to improve performance on Hinglish conversational tasks. The proposed approach integrates synthetically generated dialogues with insights from existing Hinglish datasets to address data scarcity. Experimental results demonstrate that models with fewer parameters, when appropriately fine-tuned on high-quality code-mixed data, can achieve competitive performance for Hinglish conversation generation while maintaining computational efficiency.

cs.CL↗

arXiv2025

LLM-driven Constrained Copy Generation through Iterative Refinement

Crafting a marketing message (copy), or copywriting is a challenging generation task, as the copy must adhere to various constraints. Copy creation is inherently iterative for humans, starting with an initial draft followed by successive refinements. However, manual copy creation is time-consuming and expensive, resulting in only a few copies for each use case. This limitation restricts our ability to personalize content to customers. Contrary to the manual approach, LLMs can generate copies quickly, but the generated content does not consistently meet all the constraints on the first attempt (similar to humans). While recent studies have shown promise in improving constrained generation through iterative refinement, they have primarily addressed tasks with only a few simple constraints. Consequently, the effectiveness of iterative refinement for tasks such as copy generation, which involves many intricate constraints, remains unclear. To address this gap, we propose an LLM-based end-to-end framework for scalable copy generation using iterative refinement. To the best of our knowledge, this is the first study to address multiple challenging constraints simultaneously in copy generation. Examples of these constraints include length, topics, keywords, preferred lexical ordering, and tone of voice. We demonstrate the performance of our framework by creating copies for e-commerce banners for three different use cases of varying complexity. Our results show that iterative refinement increases the copy success rate by $16.25-35.91$% across use cases. Furthermore, the copies generated using our approach outperformed manually created content in multiple pilot studies using a multi-armed bandit framework. The winning copy improved the click-through rate by $38.5-45.21$%.

cs.CL↗

arXiv2024

Chaining text-to-image and large language model: A novel approach for generating personalized e-commerce banners

Text-to-image models such as stable diffusion have opened a plethora of opportunities for generating art. Recent literature has surveyed the use of text-to-image models for enhancing the work of many creative artists. Many e-commerce platforms employ a manual process to generate the banners, which is time-consuming and has limitations of scalability. In this work, we demonstrate the use of text-to-image models for generating personalized web banners with dynamic content for online shoppers based on their interactions. The novelty in this approach lies in converting users' interaction data to meaningful prompts without human intervention. To this end, we utilize a large language model (LLM) to systematically extract a tuple of attributes from item meta-information. The attributes are then passed to a text-to-image model via prompt engineering to generate images for the banner. Our results show that the proposed approach can create high-quality personalized banners for users.

cs.HC↗

arXiv2022

The temporal overfitting problem with applications in wind power curve modeling

This paper is concerned with a nonparametric regression problem in which the input variables and the errors are autocorrelated in time. The motivation for the research stems from modeling wind power curves. Using existing model selection methods, like cross validation, results in model overfitting in presence of temporal autocorrelation. This phenomenon is referred to as temporal overfitting, which causes loss of performance while predicting responses for a time domain different from the training time domain. We propose a Gaussian process (GP)-based method to tackle the temporal overfitting problem. Our model is partitioned into two parts -- a time-invariant component and a time-varying component, each of which is modeled through a GP. We modify the inference method to a thinning-based strategy, an idea borrowed from Markov chain Monte Carlo sampling, to overcome temporal overfitting and estimate the time-invariant component. We extensively compare our proposed method with both existing power curve models and available ideas for handling temporal overfitting on real wind turbine datasets. Our approach yields significant improvement when predicting response for a time period different from the training time period. Supplementary material and computer code for this article is available online.

stat.AP↗

arXiv2022

Development of a Scalable Quantum Memory Platform -- Materials Science of Erbium-Doped TiO$_2$ Thin Films on Silicon

Rare-earth ions (REI) have emerged as an attractive candidate for solid-state qubits, particularly as a quantum memory. Their 4f-4f transitions are shielded by filled 5s and 5p orbitals, offering a degree of protection from external electric fields. Embedded within a thin film oxide host, REIs could enable a qubit platform with significant memory capabilities. Furthermore, a silicon-compatible thin film form factor would enable the use of standard semiconductor fabrication processes to achieve chip-based integrability and scalability for functional quantum networks. Towards this goal, we have carried out optical and microstructural studies of erbium-doped polycrystalline and epitaxial TiO$_2$ thin films on Si (100), r-sapphire, and SrTiO$_3$ (100). We observe that the inhomogeneous optical linewidth of the Er photoluminescence is comparable or better for polycrystalline Er:TiO$_2$(grown on Si) in comparison to single crystal epitaxial films on sapphire or SrTiO$_3$, implying a relative insensitivity to extended defects. We investigated the effect of the film/substrate and film/air interface and found that the inhomogeneous linewidth and spectral diffusion can be significantly improved via bottom buffer and top capping layers of undoped TiO$_2$. Using such approaches, we obtain inhomogeneous linewidths of 5.2 GHz and spectral diffusion of 180 MHz in Er:TiO$_2$ /Si(100) films and have demonstrated the engineerability of quantum-relevant properties in these thin films.

cond-mat.mtrl-sci↗

arXiv2021

Direct observation and consequences of dopant segregation inside and outside dislocation cores in perovskite BaSnO3

Distinct dopant behaviors inside and outside dislocation cores are identified by atomic-resolution electron microscopy in perovskite BaSnO3 with considerable consequences on local atomic and electronic structures. Driven by elastic strain, when A-site designated La dopants segregate near a dislocation core, the dopant atoms accumulate at the Ba sites in compressively strained regions. This triggers formation of Ba-vacancies adjacent to the core atomic sites resulting in reconstruction of the core. Notwithstanding the presence of extremely large tensile strain fields, when La atoms segregate inside the dislocation core, they become B-site dopants, replacing Sn atoms and compensating the positive charge of the core oxygen vacancies. Electron energy-loss spectroscopy shows that the local electronic structure of these dislocations changes dramatically due to the segregation of the dopants inside and around the core ranging from formation of strong La-O hybridized electronic states near the conduction band minimum to insulator-to-metal transition.

cond-mat.mtrl-sci↗

arXiv2021

A Case Study of Space-time Performance Comparison of Wind Turbines on a Wind Farm

This paper presents an academia-industry joint case study, which was conducted to quantify and compare multi-year changes in power production performance of multiple turbines scattered over a mid-size wind farm. This analysis is referred to as a space-time performance comparison. One key aspect in power performance analysis is to have the wind and environmental inputs controlled for. This research employs, in a sequential fashion, two principal modeling components to exercise tight control of multiple input conditions -- a covariate matching method, followed by a Gaussian process model-based functional comparison. The analysis method is applied to a wind farm that houses 66 turbines on a moderately complex terrain. The power production and environmental data span nearly four years, during which period the turbines have gone through multiple technical upgrades. The space-time analysis presents a quantitative and global picture showing how turbines differ relative to each other as well as how each of them changes over time.

stat.AP↗

arXiv2021

Gaussian process aided function comparison using noisy scattered data

This work proposes a nonparametric method to compare the underlying mean functions given two noisy datasets. The motivation for the work stems from an application of comparing wind turbine power curves. Comparing wind turbine data presents new problems, namely the need to identify the regions of difference in the input space and to quantify the extent of difference that is statistically significant. Our proposed method, referred to as funGP, estimates the underlying functions for different data samples using Gaussian process models. We build a confidence band using the probability law of the estimated function differences under the null hypothesis. Then, the confidence band is used for the hypothesis test as well as for identifying the regions of difference. This identification of difference regions is a distinct feature, as existing methods tend to conduct an overall hypothesis test stating whether two functions are different. Understanding the difference regions can lead to further practical insights and help devise better control and maintenance strategies for wind turbines. The merit of funGP is demonstrated by using three simulation studies and four real wind turbine datasets.

stat.ME↗

arXiv2021

Surface state at $BaSnO_3$ evidenced by angle-resolved photoemission spectroscopy and ab initio calculations

Perovskite alkaline earth stannates, such as $BaSnO_3$ and $SrSnO_3$, showing light transparency and high electrical conductivity (when doped), have become promising candidates for novel optoelectrical devices. Such devices are mostly based on hetero-structures and understanding of their electronic structure, which usually deviates from the bulk, is mandatory for exploring a full application potential. Employing angle-resolved photoemission spectroscopy and ab initio calculations we reveal the existence of a 2-dimensional metallic state at the $SnO_2$-terminated surface of a 1\% La-doped $BaSnO_3$ thin film. The observed surface state is characterized by distinct carrier density and a smaller effective mass in comparison with the corresponding bulk values. The small surface effective mass of about $0.12m_e$ can cause an improvement of the electrical conductivity of BSO based heterostructures.

cond-mat.mtrl-sci↗

arXiv2020

Self-Assembled Periodic Nanostructures Using Martensitic Phase Transformations

We describe a novel approach for the rational design and synthesis of self-assembled periodic nanostructures using martensitic phase transformations. We demonstrate this approach in a thin film of perovskite SrSnO3 with reconfigurable periodic nanostructures consisting of regularly spaced regions of sharply contrasted dielectric properties. The films can be designed to have different periodicities and relative phase fractions via chemical doping or strain engineering. The dielectric contrast within a single film can be tuned using temperature and laser wavelength, effectively creating a variable photonic crystal. Our results show the realistic possibility of designing large-area self-assembled periodic structures using martensitic phase transformations with the potential of implementing "built-to-order" nanostructures for tailored optoelectronic functionalities.

cond-mat.mtrl-sci↗

arXiv2020

Electronic Structure and Small Hole Polarons in YTiO3

As a prototypical Mott insulator with ferromagnetic ordering, YTiO3 (YTO) is of great interest in the study of strong electron correlation effects and orbital ordering. Here we report the first molecular beam epitaxy (MBE) growth of YTO films, combined with theoretical and experimental characterization of the electronic structure and charge transport properties. The obstacles of YTO MBE growth are discussed and potential routes to overcome them are proposed. DC transport and Seebeck measurements on thin films and bulk single crystals identify p-type Arrhenius transport behavior, with an activation energy of ~ 0.17 eV in thin films, consistent with the energy barrier for small hole polaron migration from hybrid density functional theory (DFT) calculations. Hard X-ray photoelectron spectroscopy measurements (HAXPES) show the lower Hubbard band (LHB) at 1.1 eV below the Fermi level, whereas a Mott-Hubbard band gap of ~1.5 eV is determined from photoluminescence (PL) measurements. These findings provide critical insight into the electronic band structure of YTO and related materials.

cond-mat.mtrl-sci↗

arXiv2020

Precursor Selection in Hybrid Molecular Beam Epitaxy of Alkaline-Earth Stannates

One of the challenges of oxide molecular beam epitaxy (MBE) is the synthesis of oxides containing metals with high electronegativity (metals that are hard to oxidize). The use of reactive organometallic precursors can potentially address this issue. To investigate the formation of radicals in MBE, we explored three carefully chosen metal-organic precursors of tin for SnO2 and BaSnO3 growth: tetramethyltin (TMT), tetraethyltin (TET), and hexamethylditin (HMDT). All three precursors produced single-crystalline, atomically smooth, and epitaxial SnO2 (101) films on r-Al2O3 in the presence of an oxygen plasma. The study of growth kinetics revealed reaction-limited and flux-limited regimes except for TET, which also exhibited a decrease in deposition rate with increasing temperature above 800 C. Contrary to these similarities, the performance of these precursors was dramatically different for BaSnO3 growth. TMT and TET were ineffective in supplying adequate tin whereas HMDT yielded phase-pure, stoichiometric BaSnO3 films. Significantly, HMDT resulted in phase-pure and stoichiometric BaSnO3 films even without the use of an oxygen plasma (i.e., with molecular oxygen alone). These results are discussed using the ability of HMDT to form tin radicals and therefore, assisting with Sn to Sn4+ oxidation reaction. Structural and electronic transport properties of films grown using HMDT with and without oxygen plasma are compared. This study provides guideline for the choice of precursors that will enable synthesis of metal oxides containing hard-to-oxidize metals using reactive radicals in MBE.

cond-mat.mtrl-sci↗

arXiv2020

Machine Learning Analysis of Perovskite Oxides Grown by Molecular Beam Epitaxy

Reflection high-energy electron diffraction (RHEED) is a ubiquitous in situ molecular beam epitaxial (MBE) characterization tool. Although RHEED can be a powerful means for crystal surface structure determination, it is often used as a static qualitative surface characterization method at discrete intervals during a growth. A full analysis of RHEED data collected during the entirety of MBE growths is made possible using principle component analysis (PCA) and k-means clustering to examine significant boundaries that occur in the temporal clusters grouped from RHEED data and identify statistically significant patterns. This process is applied to data from homoepitaxial SrTiO$_{3}$ growths, heteroepitaxial SrTiO$_{3}$ grown on scandate substrates, BaSnO$_{3}$ films grown on SrTiO$_{3}$ substrates, and LaNiO$_{3}$ films grown on LaAlO$_{3}$ substrates. This analysis may provide additional insights into the surface evolution and transitions in growth modes at precise times and depths during growth, and that video archival of an entire RHEED image sequence may be able to provide more insight and control over growth processes and film quality.

cond-mat.mtrl-sci↗

arXiv2019

Dopant Solubility, and Charge Compensation in La-doped SrSnO3 Films

We investigate lanthanum (La) as an n-type dopant in the strain-stabilized tetragonal phase of SrSnO3 grown on GdScO3 (110) using a radical-based hybrid molecular beam epitaxy approach. Fully coherent, epitaxial films with atomically smooth film surface were obtained irrespective of doping density. By combining secondary ion mass spectroscopy and Hall measurements, we demonstrate that each La atom contributes to one electron to the film confirming it occupies Sr-site in SrSnO3 and that it is completely activated. Carrier density exceeding 1 x 10^20 cm-3 was achieved in LSSO films, which is in excellent agreement with the dopant-solubility limit predicted by the density functional theory calculations. A record-high room-temperature mobility of 70 cm2V-1s-1 at 1 x 10^20 cm-3 was obtained in 12 nm La-doped SrSnO3 film making this the thinnest perovskite oxide semiconductor with a reasonably high electron mobility at room temperature. We discuss the structure-dopant-transport property relationships providing essential knowledge for the design of electronic devices using these materials.

cond-mat.mtrl-sci↗

arXiv2019

STEM beam channeling in BaSnO3/LaAlO3 perovskite bilayers and visualization of 2D misfit dislocation network

A study of the STEM probe channeling in a heterostructures crystalline specimen is presented here with a goal to guide appropriate STEM-based characterization for complex structures. STEM analysis of perovskite BaSnO3/LaAlO3 bilayers is performed and the dominating effects of beam channeling on HAADF- and LAADF-STEM are illustrated. To study the electron beam propagating through BaSnO3/LaAlO3 bilayers, probe intensity depth profiles are calculated, and the effects of probe defocus and atomic column alignment are discussed. Characteristics of the beam channeling are correlated to resulting ADF-STEM images, which is then tested by comparing focal series of plan-view HAADF-STEM images to those recorded experimentally. Additionally, discussions on how to visualize the misfit dislocation network at the BaSnO3/LaAlO3 interface using HAADF- and LAADF-STEM images are provided.

cond-mat.mtrl-sci↗

arXiv2019

Unraveling the Effect of Electron-Electron Interaction on Electronic Transport in High-Mobility Stannate Films

Contrary to the common belief that electron-electron interaction (EEI) should be negligible in s-orbital-based conductors, we demonstrated that the EEI effect could play a significant role on electronic transport leading to the misinterpretation of the Hall data. We show that the EEI effect is primarily responsible for an increase in the Hall coefficient in the La-doped SrSnO3 films below 50 K accompanied by an increase in the sheet resistance. The quantitative analysis of the magnetoresistance (MR) data yielded a large phase coherence length of electrons exceeding 450 nm at 1.8 K and revealed the electron-electron interaction being accountable for breaking of electron phase coherency in La-doped SrSnO3 films. These results while providing critical insights into the fundamental transport behavior in doped stannates also indicate the potential applications of stannates in quantum coherent electronic devices owing to their large phase coherence length.

cond-mat.mtrl-sci↗

arXiv2019

Separating Electrons and Donors in BaSnO3 via Band Engineering

Through a combination of thin film growth, hard X-ray photoelectron spectroscopy (HAXPES), scanning transmission electron microscopy/electron energy loss spectroscopy (STEM/EELS), magneto-transport measurements, and transport modeling, we report on the demonstration of modulation-doping of BaSnO3 (BSO) using a wider bandgap La-doped SrSnO3 (LSSO) layer. Hard X-ray photoelectron spectroscopy (HAXPES) revealed a valence band offset of 0.71 +/- 0.02 eV between LSSO and BSO resulting in a favorable conduction band offset for remote doping of BSO using LSSO. Nonlinear Hall effect of LSSO/BSO heterostructure confirmed two-channel conduction owing to electron transfer from LSSO to BSO and remained in good agreement with the results of self-consistent solution to one-dimensional Poisson and Schrödinger equations. Angle-dependent HAXPES measurements revealed a spatial distribution of electrons over 2-3 unit cells in BSO. These results bring perovskite oxides a step closer to room-temperature oxide electronics by establishing modulation-doping approaches in non-SrTiO3-based oxide heterostructure.

cond-mat.mtrl-sci↗

arXiv2016

Electronic structure of BaSnO3 investigated by high-energy-resolution electron energy-loss spectroscopy and ab initio calculations

There has been growing interest in perovskite BaSnO3 due to its desirable properties for oxide electronic devices including high electron mobility at room temperature and optical transparency. As these electronic and optical properties originate largely from the electronic structure of the material, here the basic electronic structure of epitaxially-grown BaSnO3 films is studied using high-energy-resolution electron energy-loss spectroscopy in a transmission electron microscope and ab initio calculations. This study provides a detailed description of the dielectric function of BaSnO3, including the energies of bulk plasmon excitations and critical interband electronic transitions, the band structure and partial densities of states, the measured band gap, and more. To make the study representative of a variety of deposition methods, results from BaSnO3 films grown by both hybrid molecular beam epitaxy and high pressure oxygen sputter deposition are reported.

cond-mat.mtrl-sci↗