arXiv subjects

Shreya Shukla

Publications and source records attributed to Shreya Shukla.

14 recordsLinked to original sources

arXiv2026

RA-CoA: Training-free Fashion Image Captioning via Retrieval-Augmented Chain-of-Attributes

Fashion Image Captioning (FIC) plays a vital role in enhancing user experience and product search in e-commerce platforms. Unlike natural scene image captioning, FIC requires fine-grained visual reasoning and knowledge of domain-specific terminology to capture subtle attributes such as neckline and closure types, graphic patterns, and dress silhouettes. Moreover, as fashion inventories evolve rapidly with new trends, styles, and frequently emerging vocabulary, developing training-free captioning solution becomes essential for scalability and real-world adaptability. Instruction-tuned vision-language models (VLMs) offer a promising solution to fashion image captioning dueto their strong zero-shot capabilities and natural language fluency. However, these general-purpose models often lack attribute-level coverage and precision, and tend to hallucinate or misidentify fine-grained fashion details, making them less suitable for high-fidelity applications like product cataloging or personalized recommendations. To address this, we propose RA-CoA (Retrieval-Augmented Chain-of-Attributes), a novel, training-free framework that disentangles fashion image captioning into two interpretable stages: (i) retrieval of relevant attribute sets from a product knowledge base, and (ii) attribute-level reasoning to generate the final caption. RA-CoA is a model-agnostic approach that works with frozen VLMs to improve fine-grained attribute precision in product captions without the need for fine-tuning. Extensive evaluations across diverse VLM model families under different prompting paradigms demonstrate that RA-CoA significantly improves caption quality, achieving an average gain of 26.3% METEOR score over zero-shot captioning. We make our code publicly available.

cs.CV↗

arXiv2026

Computationally sufficient statistics for Ising models

Learning Gibbs distributions using only sufficient statistics has long been recognized as a computationally hard problem. On the other hand, computationally efficient algorithms for learning Gibbs distributions rely on access to full sample configurations generated from the model. For many systems of interest that arise in physical contexts, expecting a full sample to be observed is not practical, and hence it is important to look for computationally efficient methods that solve the learning problem with access to only a limited set of statistics. We examine the trade-offs between the power of computation and observation within this scenario, employing the Ising model as a paradigmatic example. We demonstrate that it is feasible to reconstruct the model parameters for a model with $\ell_1$ width $γ$ by observing statistics up to an order of $O(γ)$. This approach allows us to infer the model's structure and also learn its couplings and magnetic fields. We also discuss a setting where prior information about structure of the model is available and show that the learning problem can be solved efficiently with even more limited observational power.

cs.LG↗

arXiv2025

Efficient Learning of Lattice Gauge Theories with Fermions

We introduce a learning method for recovering action parameters in lattice field theories. Our method is based on the minimization of a convex loss function constructed using the Schwinger-Dyson relations. We show that score matching, a popular learning method, is a special case of our construction of an infinite family of valid loss functions. Importantly, our general Schwinger-Dyson-based construction applies to gauge theories and models with Grassmann-valued fields used to represent dynamical fermions. In particular, we extend our method to realistic lattice field theories including quantum chromodynamics.

hep-lat↗

arXiv2025

qa-FLoRA: Data-free query-adaptive Fusion of LoRAs for LLMs

The deployment of large language models for specialized tasks often requires domain-specific parameter-efficient finetuning through Low-Rank Adaptation (LoRA) modules. However, effectively fusing these adapters to handle complex, multi-domain composite queries remains a critical challenge. Existing LoRA fusion approaches either use static weights, which assign equal relevance to each participating LoRA, or require data-intensive supervised training for every possible LoRA combination to obtain respective optimal fusion weights. We propose qa-FLoRA, a novel query-adaptive data-and-training-free method for LoRA fusion that dynamically computes layer-level fusion weights by measuring distributional divergence between the base model and respective adapters. Our approach eliminates the need for composite training data or domain-representative samples, making it readily applicable to existing adapter collections. Extensive experiments across nine multilingual composite tasks spanning mathematics, coding, and medical domains, show that qa-FLoRA outperforms static fusion by ~5% with LLaMA-2 and ~6% with LLaMA-3, and the training-free baselines by ~7% with LLaMA-2 and ~10% with LLaMA-3, while significantly closing the gap with supervised baselines. Further, layer-level analysis of our fusion weights reveals interpretable fusion patterns, demonstrating the effectiveness of our approach for robust multi-domain adaptation.

cs.CL↗

arXiv2025

Learning of Statistical Field Theories

Recovering microscopic couplings directly from data provides a route to solving the inverse problem in statistical field theories, one that complements the traditional-often computationally intractable-forward approach of predicting observables from an action or Hamiltonian. Here, we propose an approach for the inverse problem that uniformly accommodates systems with discrete, continuous, and hybrid variables. We demonstrate accurate parameter recovery in several benchmark systems-including Wegner's Ising gauge theory, $ϕ^4$ theory, Schwinger and Sine-Gordon models, and mixed spin-gauge systems, and show how iterating the procedure under coarse-graining reconstructs full non-perturbative renormalization-group flows. This gives direct access to phase boundaries, fixed points, and emergent interactions without relying on perturbation theory. We also address a realistic setting where full gauge configurations may be unavailable, and reformulate learning algorithms for multiple field theories so that they are recovered directly using observables such as correlations from scattering data or quantum simulators. We anticipate that our methodology will find widespread use in practical learning of field theories in strongly coupled regimes where analytical tools might fail.

cond-mat.stat-mech↗

arXiv2025

A Survey on Bridging EEG Signals and Generative AI: From Image and Text to Beyond

Decoding neural activity into human-interpretable representations is a key research direction in brain-computer interfaces (BCIs) and computational neuroscience. Recent progress in machine learning and generative AI has driven growing interest in transforming non-invasive Electroencephalography (EEG) signals into images, text, and audio. This survey consolidates and analyzes developments across EEG-to-image synthesis, EEG-to-text generation, and EEG-to-audio reconstruction. We conducted a structured literature search across major databases (2017-2025), extracting key information on datasets, generative architectures (GANs, VAEs, transformers, diffusion models), EEG feature-encoding techniques, evaluation metrics, and the major challenges shaping current work in this area. Our review finds that EEG-to-image models predominantly employ encoder-decoder architectures built on GANs, VAEs, or diffusion models; EEG-to-text approaches increasingly leverage transformer-based language models for open-vocabulary decoding; and EEG-to-audio methods commonly map EEG signals to mel-spectrograms that are subsequently rendered into audio using neural vocoders. Despite promising advances, the field remains constrained by small and heterogeneous datasets, limited cross-subject generalization, and the absence of standardized benchmarks. By consolidating methodological trends and available datasets, this survey provides a foundational reference for advancing EEG-based generative AI and supporting reproducible research. We further highlight open-source datasets and baseline implementations to facilitate systematic benchmarking and accelerate progress in EEG-driven neural decoding.

cs.AI↗

arXiv2025

Thought2Text: Text Generation from EEG Signal using Large Language Models (LLMs)

Decoding and expressing brain activity in a comprehensible form is a challenging frontier in AI. This paper presents Thought2Text, which uses instruction-tuned Large Language Models (LLMs) fine-tuned with EEG data to achieve this goal. The approach involves three stages: (1) training an EEG encoder for visual feature extraction, (2) fine-tuning LLMs on image and text data, enabling multimodal description generation, and (3) further fine-tuning on EEG embeddings to generate text directly from EEG during inference. Experiments on a public EEG dataset collected for six subjects with image stimuli and text captions demonstrate the efficacy of multimodal LLMs (LLaMA-v3, Mistral-v0.3, Qwen2.5), validated using traditional language generation evaluation metrics, as well as fluency and adequacy measures. This approach marks a significant advancement towards portable, low-cost "thoughts-to-text" technology with potential applications in both neuroscience and natural language processing.

cs.CL↗

arXiv2025

Towards Making Flowchart Images Machine Interpretable

Computer programming textbooks and software documentations often contain flowcharts to illustrate the flow of an algorithm or procedure. Modern OCR engines often tag these flowcharts as graphics and ignore them in further processing. In this paper, we work towards making flowchart images machine-interpretable by converting them to executable Python codes. To this end, inspired by the recent success in natural language to code generation literature, we present a novel transformer-based framework, namely FloCo-T5. Our model is well-suited for this task,as it can effectively learn semantics, structure, and patterns of programming languages, which it leverages to generate syntactically correct code. We also used a task-specific pre-training objective to pre-train FloCo-T5 using a large number of logic-preserving augmented code samples. Further, to perform a rigorous study of this problem, we introduce theFloCo dataset that contains 11,884 flowchart images and their corresponding Python codes. Our experiments show promising results, and FloCo-T5 clearly outperforms related competitive baselines on code generation metrics. We make our dataset and implementation publicly available.

cs.CV↗

arXiv2025

PatentLMM: Large Multimodal Model for Generating Descriptions for Patent Figures

Writing comprehensive and accurate descriptions of technical drawings in patent documents is crucial to effective knowledge sharing and enabling the replication and protection of intellectual property. However, automation of this task has been largely overlooked by the research community. To this end, we introduce PatentDesc-355K, a novel large-scale dataset containing ~355K patent figures along with their brief and detailed textual descriptions extracted from more than 60K US patent documents. In addition, we propose PatentLMM - a novel multimodal large language model specifically tailored to generate high-quality descriptions of patent figures. Our proposed PatentLMM comprises two key components: (i) PatentMME, a specialized multimodal vision encoder that captures the unique structural elements of patent figures, and (ii) PatentLLaMA, a domain-adapted version of LLaMA fine-tuned on a large collection of patents. Extensive experiments demonstrate that training a vision encoder specifically designed for patent figures significantly boosts the performance, generating coherent descriptions compared to fine-tuning similar-sized off-the-shelf multimodal models. PatentDesc-355K and PatentLMM pave the way for automating the understanding of patent figures, enabling efficient knowledge sharing and faster drafting of patent documents. We make the code and data publicly available.

cs.CV↗

arXiv2023

Chirality Changing RG Flows: Dynamics and Models

Chirality plays an important role in understanding the dynamics of quantum field theories. In this paper, we study the dynamics of models where renormalization group flows change the chiral structure of the theory. We introduce model building tools and construct models with a variety of chirality flows: from the appearance of new massless composite matter, to the development of mass gaps to completely general changes in the chiral matter content. The stability of chirally symmetric vacua is sensitive to the interplay between non-perturbative dynamics and deformations necessary to generate chirality flows. In particular, we show that chirality flows can be easily induced by deformations of s-confining models. On the other hand, in the absence of true s-confinement, the required deformations destabilize chirally symmetric ground states.

hep-th↗

arXiv2022

Neutrino Flavor Model Building and the Origins of Flavor and CP Violation: A Snowmass White Paper

The neutrino sector offers one of the most sensitive probes of new physics beyond the Standard Model of Particle Physics. The mechanism of neutrino mass generation is still unknown. The observed suppression of neutrino masses hints at a large scale, conceivably of the order of the scale of a Grand Unified Theory (GUT), a unique feature of neutrinos that is not shared by the charged fermions. The origin of neutrino masses and mixing is part of the outstanding puzzle of fermion masses and mixings, which is not explained in the SM. Flavor model building for both quark and lepton sectors is important in order to gain a better understanding of the origin of the structure of mass hierarchy and flavor mixing, which constitute the dominant fraction of the SM parameters. Recent activities in neutrino flavor model building based on non-Abelian discrete flavor symmetries and modular flavor symmetries have been shown to be a promising direction to explore. The emerging models provide a framework that has a significantly reduced number of undetermined parameters in the flavor sector. Model building based on non-Abelian discrete flavor symmetries and their modular variants enables the particle physics community to interpret the current and anticipated upcoming data from neutrino experiments. Pursuit of flavor model building based on such frameworks can also provide connections to possible UV completions, in particular to string theory. We emphasize the importance of constructing models in which the uncertainties of theoretical predictions are smaller than, or at most compatible with, the error bars of measurements in neutrino experiments.

hep-ph↗

arXiv2021

Generation flow in field theory and strings

Nontrivial strong dynamics often leads to the appearance of chiral composites. In phenomenological applications, these can either play the role of Standard Model particles or lift chiral exotics by partnering with them in mass terms. As a consequence, the RG flow may change the effective number of chiral generations, a phenomenon we call generation flow. We provide explicit constructions of globally consistent string models exhibiting generation flow. Since such constructions were misclassified in the traditional model searches, our results imply that more care than usually appreciated has to be taken when scanning string compactifications for realistic models.

hep-th↗

arXiv2021

Quasi-Eclectic Modular Flavor Symmetries

Modular flavor symmetries provide us with a new, promising approach to the flavor problem. However, in their original formulation the kinetic terms of the standard model fields do not have a preferred form, thus introducing additional parameters, which limit the predictive power of this scheme. In this work, we introduce the scheme of quasi-eclectic flavor symmetries as a simple fix. These symmetries are the direct product of a modular and a traditional flavor symmetry, which are spontaneously broken to a diagonal modular flavor subgroup. This allows us to construct a version of Feruglio's model with the Kaehler terms under control. At the same time, the starting point is reminiscent of what one obtains from explicit string models.

hep-ph↗

arXiv2021

Metaplectic Flavor Symmetries from Magnetized Tori

We revisit the flavor symmetries arising from compactifications on tori with magnetic background fluxes. Using Euler's Theorem, we derive closed form analytic expressions for the Yukawa couplings that are valid for arbitrary flux parameters. We discuss the modular transformations for even and odd units of magnetic flux, M, and show that they give rise to finite metaplectic groups the order of which is determined by the least common multiple of the number of zero-mode flavors involved. Unlike in models in which modular flavor symmetries are postulated, in this approach they derive from an underlying torus. This allows us to retain control over parameters, such as those governing the kinetic terms, that are free in the bottom-up approach, thus leading to an increased predictivity. In addition, the geometric picture allows us to understand the relative suppression of Yukawa couplings from their localization properties in the compact space. We also comment on the role supersymmetry plays in these constructions, and outline a path towards non-supersymmetric models with modular flavor symmetries.

hep-th↗