arXiv subjects

Aditya Mehta

Publications and source records attributed to Aditya Mehta.

8 recordsLinked to original sources

arXiv2026

Interface and Strain Control of Emergent Weyl Semimetallic Phase in SrNbO$_{3}$/LaFeO$_{3}$ Heterostructures

Realizing correlated topological semimetallic phases in bulk transition-metal oxides remains challenging due to rigid lattice symmetry, correlation-induced gap opening, and limited structural tunability. However, complex-oxide thin films and heterostructures provide a powerful platform to stabilize topological phases by tailoring the requisite lattice symmetry through strain control and interface design. In this study, we demonstrate the emergence of Weyl-like electronic states and associated chiral transport in SrNbO$_3$ (SNO)/LaFeO$_3$ (LFO) bilayers. Transport measurements reveal signatures consistent with nontrivial topology, including large non-saturating MR, a nonlinear Hall response, and a chiral anomaly like feature in longitudinal magnetotransport under parallel electric and magnetic fields ($\mathbf{B} \parallel \mathbf{I}$). In addition, we observe a \textcolor{black}{signature} of anomalous Hall contribution, likely arising from \textcolor{black}{proximity effect induced by LFO layers at the interface}. First-principles calculations reveal an $a^0a^0c^-$ rotation pattern of the NbO$_6$ octahedra, together with interfacial lattice distortions in the SNO layer that drive the emergence of a twofold degenerate Weyl semimetallic phase protected by screw axis lattice symmetry. This is further confirmed by Berry curvature calculations, which show opposite sign Berry curvature peaks for the upper and lower band characteristic of a Weyl node. Our combined experimental and theoretical results highlight the critical role of strain and interfacial octahedral distortions in stabilizing Weyl phase in transition metal based perovskite bilayer.

cond-mat.str-el↗

arXiv2025

Same or Not? Enhancing Visual Perception in Vision-Language Models

Vision-language models (VLMs) excel at broad visual understanding but remain coarse-grained, exhibit visual biases, and miss subtle visual details. Existing training corpora reinforce this limitation by emphasizing general recognition ("Is it a cat or a dog?") over fine-grained perception. To address this, we introduce a new training corpus and task designed to enhance the perceptual abilities of VLMs. TWIN is a large-scale dataset of 561,000 image-pair queries that task models to determine whether two visually similar images depict the same object, encouraging attention to nuanced visual cues. The dataset spans a diverse range of everyday objects across contexts, viewpoints, and appearances. Fine-tuning VLMs on TWIN yields notable gains in fine-grained recognition, even on unseen domains such as art, animals, plants, and landmarks. To quantify these gains, we introduce FGVQA, a benchmark suite of 12,000 queries that repurposes fine-grained recognition and retrieval datasets from multiple domains. While existing VLMs struggle on FGVQA, when fine-tuned on TWIN they improve by up to 19.3%, without compromising performance on general VQA benchmarks. Finally, our TWIN dataset scales favorably with object annotations, and our analysis shows that scale is key to performance. We envision TWIN as a drop-in addition to open-source VLM training corpora, advancing perceptual precision of future models. Project webpage: https://glab-caltech.github.io/twin/

cs.CV↗

arXiv2025

Exposing DeepFakes via Hyperspectral Domain Mapping

Modern generative and diffusion models produce highly realistic images that can mislead human perception and even sophisticated automated detection systems. Most detection methods operate in RGB space and thus analyze only three spectral channels. We propose HSI-Detect, a two-stage pipeline that reconstructs a 31-channel hyperspectral image from a standard RGB input and performs detection in the hyperspectral domain. Expanding the input representation into denser spectral bands amplifies manipulation artifacts that are often weak or invisible in the RGB domain, particularly in specific frequency bands. We evaluate HSI-Detect across FaceForensics++ dataset and show the consistent improvements over RGB-only baselines, illustrating the promise of spectral-domain mapping for Deepfake detection.

cs.CV↗

arXiv2025

Remote Labor Index: Measuring AI Automation of Remote Work

AIs have made rapid progress on research-oriented benchmarks of knowledge and reasoning, but it remains unclear how these gains translate into economic value and automation. To measure this, we introduce the Remote Labor Index (RLI), a broadly multi-sector benchmark comprising real-world, economically valuable projects designed to evaluate end-to-end agent performance in practical settings. AI agents perform near the floor on RLI, with the highest-performing agent achieving an automation rate of 2.5%. These results help ground discussions of AI automation in empirical evidence, setting a common basis for tracking AI impacts and enabling stakeholders to proactively navigate AI-driven labor automation.

cs.LG↗

arXiv2024

LDFaceNet: Latent Diffusion-based Network for High-Fidelity Deepfake Generation

Over the past decade, there has been tremendous progress in the domain of synthetic media generation. This is mainly due to the powerful methods based on generative adversarial networks (GANs). Very recently, diffusion probabilistic models, which are inspired by non-equilibrium thermodynamics, have taken the spotlight. In the realm of image generation, diffusion models (DMs) have exhibited remarkable proficiency in producing both realistic and heterogeneous imagery through their stochastic sampling procedure. This paper proposes a novel facial swapping module, termed as LDFaceNet (Latent Diffusion based Face Swapping Network), which is based on a guided latent diffusion model that utilizes facial segmentation and facial recognition modules for a conditioned denoising process. The model employs a unique loss function to offer directional guidance to the diffusion process. Notably, LDFaceNet can incorporate supplementary facial guidance for desired outcomes without any retraining. To the best of our knowledge, this represents the first application of the latent diffusion model in the face-swapping task without prior training. The results of this study demonstrate that the proposed method can generate extremely realistic and coherent images by leveraging the potential of the diffusion model for facial swapping, thereby yielding superior visual outcomes and greater diversity.

cs.CV↗

arXiv2021

Learning to Enhance Visual Quality via Hyperspectral Domain Mapping

Deep learning based methods have achieved remarkable success in image restoration and enhancement, but most such methods rely on RGB input images. These methods fail to take into account the rich spectral distribution of natural images. We propose a deep architecture, SpecNet, which computes spectral profile to estimate pixel-wise dynamic range adjustment of a given image. First, we employ an unpaired cycle-consistent framework to generate hyperspectral images (HSI) from low-light input images. HSI is further used to generate a normal light image of the same scene. We incorporate a self-supervision and a spectral profile regularization network to infer a plausible HSI from an RGB image. We evaluate the benefits of optimizing the spectral profile for real and fake images in low-light conditions on the LOL Dataset.

eess.IV↗

arXiv2020

Domain-Aware Unsupervised Hyperspectral Reconstruction for Aerial Image Dehazing

Haze removal in aerial images is a challenging problem due to considerable variation in spatial details and varying contrast. Changes in particulate matter density often lead to degradation in visibility. Therefore, several approaches utilize multi-spectral data as auxiliary information for haze removal. In this paper, we propose SkyGAN for haze removal in aerial images. SkyGAN consists of 1) a domain-aware hazy-to-hyperspectral (H2H) module, and 2) a conditional GAN (cGAN) based multi-cue image-to-image translation module (I2I) for dehazing. The proposed H2H module reconstructs several visual bands from RGB images in an unsupervised manner, which overcomes the lack of hazy hyperspectral aerial image datasets. The module utilizes task supervision and domain adaptation in order to create a "hyperspectral catalyst" for image dehazing. The I2I module uses the hyperspectral catalyst along with a 12-channel multi-cue input and performs effective image dehazing by utilizing the entire visual spectrum. In addition, this work introduces a new dataset, called Hazy Aerial-Image (HAI) dataset, that contains more than 65,000 pairs of hazy and ground truth aerial images with realistic, non-homogeneous haze of varying density. The performance of SkyGAN is evaluated on the recent SateHaze1k dataset as well as the HAI dataset. We also present a comprehensive evaluation of HAI dataset with a representative set of state-of-the-art techniques in terms of PSNR and SSIM.

cs.CV↗

arXiv2020

Does the First Mover Advantage Exist on GitHub?

Collaborative consensus-finding is an integral element of many Web services and greatly determines the quality of information, content, and products that are available through the Web. That also means that the dynamics of democratic consensus-finding strengthen collective resilience against potential threats that attempt to degrade information, content, and products and affect Web data, users, behaviors, and even beyond as well as offline life. Even on Web platforms that are open to all, the influence of some first mover authors may shape future discussion and collaboration, which is comparable to academic citation networks for instance. In a social coding network such as GitHub, activities of a set of users can have influence on other users who can get interested in further actions, possibly contributing to a new project together with influential users. In this paper, we analyze the effect of contribution activities on gaining influence in this and comparable networks that provide users the functionality and aims for reaching collaborative goals on the Web.For this purpose, we present an empirical approach to identify the top influential users by using network features and contribution characteristics, which we find in existing and newly collected data set. We find that early adopter dynamics exist in the GitHub community, where early adopters have more followers in the end as expected. However, we also see counterexamples that arise due to the social networking behavior of late adopters, and due to the aging effect of older repositories and users. We publicly share the source code and the data sets for reproducing our paper.

cs.SI↗