arXiv subjects

Sarthak Jain

Publications and source records attributed to Sarthak Jain.

At least 19 recordsLinked to original sources

arXiv2026

AlphaWiSE: Adaptive Weight Interpolation for Continual Multimodal Representation Learning

Multimodal models such as CLIP learn a shared embedding space for cross-modal retrieval, but continual adaptation to sequentially arriving data can disrupt the cross-modal alignment acquired from earlier phases. Conventional continual-learning methods return a single checkpoint, which commits every retrieval direction to the same stability-plasticity trade-off. We propose AlphaWiSE, a post-hoc weight-space interpolation method that composes two frozen source checkpoints. For each aligned parameter tensor identified by its checkpoint key, AlphaWiSE fits one scalar interpolation coefficient shared by all tensor entries. The coefficients are fitted on a smaller exemplar memory and used to materialize one interpolated checkpoint. The deployed model has the same architecture and parameter count as either source checkpoint, which does not require additional inference time. Extensive experiments on audio-image-text retrieval show consistent improvements over strong continual-learning baselines across multiple retrieval directions and evaluation metrics.

cs.CV↗

arXiv2026

AC3S: Adaptive Conditioning for 3D-Aware Synthetic Data Generation

Synthetic data generation has emerged as a powerful tool for improving data scalability in computer vision. Recent diffusion-based pipelines have demonstrated strong photorealism. However, how to enforce precise 3D structure and pose consistency in generated images remains challenging. Existing methods leverage visual prompts such as edge maps to guide diffusion models, but often suffer from over-conditioning artifacts that degrade image realism and limit dataset quality. In this paper, we present a diffusion-based image generation framework that enforces 3D structural alignment while preserving photorealism through adaptive conditioning. Our framework, Adaptive Conditioning for 3D-Aware Synthetic Data Generation (AC3S), introduces a self-supervised visual prompt modulator that dynamically adjusts the strength of ControlNet conditioning, preventing over-conditioning and enabling the diffusion model to retain its generative expressiveness. To further enhance diversity and semantic consistency, we develop a multi-agent vision language model framework that composes detailed and 3D-aware prompts aligned with the underlying geometric structure. Together, these components enable the scalable generation of high-quality synthetic datasets with accurate 2D and 3D annotations. Extensive experiments demonstrate that our method significantly improves image quality and downstream utility.

cs.CV↗

arXiv2026

Temporal Data Requirement for Predicting Unplanned Hospital Readmissions

With the proliferation of Electronic Health Records (EHRs), a critical challenge in building predictive models is determining the optimal historical data time window to maximize accuracy. This study investigates the impact of various observation windows ranging from the day of surgery to three years prior on predicting 30-day readmission following hip and knee arthroplasties. The dataset encompasses both structured encounter records (over 4 million) and unstructured clinical notes (80,000) from 7,174 patients. To extract meaning from the clinical notes, we employed a suite of non neural (BOW, count BOW, TF IDF, LDA) and neural encoders (BERT, 1D CNN, BiLSTM, Average). We subsequently evaluated models utilizing clinical notes alone, structured data alone, and a combination of both modalities. Our results demonstrate that the optimal time window for unstructured clinical notes is significantly shorter than for structured data, maximum predictive performance was achieved using notes from just three to six months prior to surgery. In contrast, performance using structured data improved as the time window lengthened, but strictly plateaued after twelve months. These modality-specific temporal patterns remained consistent regardless of model complexity or encoder type. Ultimately, these findings challenge the general assumption that more historical data inherently yields better machine learning predictions, establishing targeted time-window guidelines for optimizing readmission prediction models.

cs.LG↗

arXiv2026

Generative deep learning for foundational video translation in ultrasound

Deep learning (DL) has the potential to revolutionize image acquisition and interpretation across medicine, however, attention to data imbalance and missingness is required. Ultrasound data presents a particular challenge because in addition to different views and structures, it includes several sub-modalities-such as greyscale and color flow doppler (CFD)-that are often imbalanced in clinical studies. Image translation can help balance datasets but is challenging for ultrasound sub-modalities to date. Here, we present a generative method for ultrasound CFD-greyscale video translation, trained on 54,975 videos and tested on 8,368. The method developed leveraged pixel-wise, adversarial, and perceptual loses and utilized two networks: one for reconstructing anatomic structures and one for denoising to achieve realistic ultrasound imaging. Average pairwise SSIM between synthetic videos and ground truth was 0.91+/-0.04. Synthetic videos performed indistinguishably from real ones in DL classification and segmentation tasks and when evaluated by blinded clinical experts: F1 score was 0.9 for real and 0.89 for synthetic videos; Dice score between real and synthetic segmentation was 0.97. Overall clinician accuracy in distinguishing real vs synthetic videos was 54+/-6% (42-61%), indicating realistic synthetic videos. Although trained only on heart videos, the model worked well on ultrasound spanning several clinical domains (average SSIM 0.91+/-0.05), demonstrating foundational abilities. Together, these data expand the utility of retrospectively collected imaging and augment the dataset design toolbox for medical imaging.

cs.CV↗

arXiv2025

SecureBERT 2.0: Advanced Language Model for Cybersecurity Intelligence

Effective analysis of cybersecurity and threat intelligence data demands language models that can interpret specialized terminology, complex document structures, and the interdependence of natural language and source code. Encoder-only transformer architectures provide efficient and robust representations that support critical tasks such as semantic search, technical entity extraction, and semantic analysis, which are key to automated threat detection, incident triage, and vulnerability assessment. However, general-purpose language models often lack the domain-specific adaptation required for high precision. We present SecureBERT 2.0, an enhanced encoder-only language model purpose-built for cybersecurity applications. Leveraging the ModernBERT architecture, SecureBERT 2.0 introduces improved long-context modeling and hierarchical encoding, enabling effective processing of extended and heterogeneous documents, including threat reports and source code artifacts. Pretrained on a domain-specific corpus more than thirteen times larger than its predecessor, comprising over 13 billion text tokens and 53 million code tokens from diverse real-world sources, SecureBERT 2.0 achieves state-of-the-art performance on multiple cybersecurity benchmarks. Experimental results demonstrate substantial improvements in semantic search for threat intelligence, semantic analysis, cybersecurity-specific named entity recognition, and automated vulnerability detection in code within the cybersecurity domain.

cs.CR↗

arXiv2025

Can Argus Judge Them All? Comparing VLMs Across Domains

Vision-Language Models (VLMs) are increasingly used in industry VLM applications such as retrieval systems, content generation platforms, and decision-support workflows, where model selection is commonly guided by benchmark rankings. These rankings are largely determined by retrieval, captioning, and reasoning downstream tasks; however, models with similar task performance often show substantially different behavior across datasets. This creates a Capability-Reliability Gap between benchmark performance and observed model stability. We present ARGUS-EVAL, a capability-reliability-oriented evaluation framework for VLMs that characterizes model behavior through Benchmark Capability P(M), Cross-Dataset Consistency CDC(M), Robustness Retention RR(M), and Efficiency E(M). We evaluate CLIP, BLIP, LXMERT, Gemma-3-4B, and Qwen-2.5VL-3B-Instruct across retrieval, captioning, and reasoning downstream tasks. The results reveal notable differences between capability-oriented and reliability-oriented rankings. Qwen-2.5VL-3BInstruct achieves the strongest overall capability (R@1 = 82.7%, BLEU-4 = 47.2%, CIDEr = 141.6, CDC = 0.91), whereas CLIP records the lowest latency (31 ms) and memory footprint (0.9 GB).

cs.IR↗

arXiv2025

SNIFR : Boosting Fine-Grained Child Harmful Content Detection Through Audio-Visual Alignment with Cascaded Cross-Transformer

As video-sharing platforms have grown over the past decade, child viewership has surged, increasing the need for precise detection of harmful content like violence or explicit scenes. Malicious users exploit moderation systems by embedding unsafe content in minimal frames to evade detection. While prior research has focused on visual cues and advanced such fine-grained detection, audio features remain underexplored. In this study, we embed audio cues with visual for fine-grained child harmful content detection and introduce SNIFR, a novel framework for effective alignment. SNIFR employs a transformer encoder for intra-modality interaction, followed by a cascaded cross-transformer for inter-modality alignment. Our approach achieves superior performance over unimodal and baseline fusion methods, setting a new state-of-the-art.

eess.AS↗

arXiv2025

ViBe: A Text-to-Video Benchmark for Evaluating Hallucination in Large Multimodal Models

Recent advances in Large Multimodal Models (LMMs) have expanded their capabilities to video understanding, with Text-to-Video (T2V) models excelling in generating videos from textual prompts. However, they still frequently produce hallucinated content, revealing AI-generated inconsistencies. We introduce ViBe (https://vibe-t2v-bench.github.io/): a large-scale dataset of hallucinated videos from open-source T2V models. We identify five major hallucination types: Vanishing Subject, Omission Error, Numeric Variability, Subject Dysmorphia, and Visual Incongruity. Using ten T2V models, we generated and manually annotated 3,782 videos from 837 diverse MS COCO captions. Our proposed benchmark includes a dataset of hallucinated videos and a classification framework using video embeddings. ViBe serves as a critical resource for evaluating T2V reliability and advancing hallucination detection. We establish classification as a baseline, with the TimeSFormer + CNN ensemble achieving the best performance (0.345 accuracy, 0.342 F1 score). While initial baselines proposed achieve modest accuracy, this highlights the difficulty of automated hallucination detection and the need for improved methods. Our research aims to drive the development of more robust T2V models and evaluate their outputs based on user preferences.

cs.CV↗

arXiv2025

Mitigating Bad Ground Truth in Supervised Machine Learning based Crop Classification: A Multi-Level Framework with Sentinel-2 Images

In agricultural management, precise Ground Truth (GT) data is crucial for accurate Machine Learning (ML) based crop classification. Yet, issues like crop mislabeling and incorrect land identification are common. We propose a multi-level GT cleaning framework while utilizing multi-temporal Sentinel-2 data to address these issues. Specifically, this framework utilizes generating embeddings for farmland, clustering similar crop profiles, and identification of outliers indicating GT errors. We validated clusters with False Colour Composite (FCC) checks and used distance-based metrics to scale and automate this verification process. The importance of cleaning the GT data became apparent when the models were trained on the clean and unclean data. For instance, when we trained a Random Forest model with the clean GT data, we achieved upto 70\% absolute percentage points higher for the F1 score metric. This approach advances crop classification methodologies, with potential for applications towards improving loan underwriting and agricultural decision-making.

cs.CV↗

arXiv2025

Self-supervised Analogical Learning using Language Models

Large language models have been shown to suffer from reasoning inconsistency issues. That is, they fail more in situations unfamiliar to the training data, even though exact or very similar reasoning paths exist in more common cases that they can successfully solve. Such observations motivate us to propose methods that encourage models to understand the high-level and abstract reasoning processes during training instead of only the final answer. This way, models can transfer the exact solution to similar cases, regardless of their relevance to the pre-training data distribution. In this work, we propose SAL, a self-supervised analogical learning framework. SAL mimics the human analogy process and trains models to explicitly transfer high-quality symbolic solutions from cases that they know how to solve to other rare cases in which they tend to fail more. We show that the resulting models after SAL learning outperform base language models on a wide range of reasoning benchmarks, such as StrategyQA, GSM8K, and HotpotQA, by 2% to 20%. At the same time, we show that our model is more generalizable and controllable through analytical studies.

cs.CL↗

arXiv2024

SeQuiFi: Mitigating Catastrophic Forgetting in Speech Emotion Recognition with Sequential Class-Finetuning

In this work, we introduce SeQuiFi, a novel approach for mitigating catastrophic forgetting (CF) in speech emotion recognition (SER). SeQuiFi adopts a sequential class-finetuning strategy, where the model is fine-tuned incrementally on one emotion class at a time, preserving and enhancing retention for each class. While various state-of-the-art (SOTA) methods, such as regularization-based, memory-based, and weight-averaging techniques, have been proposed to address CF, it still remains a challenge, particularly with diverse and multilingual datasets. Through extensive experiments, we demonstrate that SeQuiFi significantly outperforms both vanilla fine-tuning and SOTA continual learning techniques in terms of accuracy and F1 scores on multiple benchmark SER datasets, including CREMA-D, RAVDESS, Emo-DB, MESD, and SHEMO, covering different languages.

eess.AS↗

arXiv2024

MMLU-SR: A Benchmark for Stress-Testing Reasoning Capability of Large Language Models

We propose MMLU-SR, a novel dataset designed to measure the true comprehension abilities of Large Language Models (LLMs) by challenging their performance in question-answering tasks with modified terms. We reasoned that an agent that "truly" understands a concept can still evaluate it when key terms are replaced by suitably defined alternate terms, and sought to differentiate such comprehension from mere text replacement. In our study, we modified standardized test questions by replacing a key term with a dummy word along with its definition. The key term could be in the context of questions, answers, or both questions and answers. Notwithstanding the high scores achieved by recent popular LLMs on the MMLU leaderboard, we found a substantial reduction in model performance after such replacement, suggesting poor comprehension. This new benchmark provides a rigorous benchmark for testing true model comprehension, and poses a challenge to the broader scientific community.

cs.CL↗

arXiv2024

Are Music Foundation Models Better at Singing Voice Deepfake Detection? Far-Better Fuse them with Speech Foundation Models

In this study, for the first time, we extensively investigate whether music foundation models (MFMs) or speech foundation models (SFMs) work better for singing voice deepfake detection (SVDD), which has recently attracted attention in the research community. For this, we perform a comprehensive comparative study of state-of-the-art (SOTA) MFMs (MERT variants and music2vec) and SFMs (pre-trained for general speech representation learning as well as speaker recognition). We show that speaker recognition SFM representations perform the best amongst all the foundation models (FMs), and this performance can be attributed to its higher efficacy in capturing the pitch, tone, intensity, etc, characteristics present in singing voices. To our end, we also explore the fusion of FMs for exploiting their complementary behavior for improved SVDD, and we propose a novel framework, FIONA for the same. With FIONA, through the synchronization of x-vector (speaker recognition SFM) and MERT-v1-330M (MFM), we report the best performance with the lowest Equal Error Rate (EER) of 13.74 %, beating all the individual FMs as well as baseline FM fusions and achieving SOTA results.

eess.AS↗

arXiv2024

LLM Agents Improve Semantic Code Search

Code Search is a key task that many programmers often have to perform while developing solutions to problems. Current methodologies suffer from an inability to perform accurately on prompts that contain some ambiguity or ones that require additional context relative to a code-base. We introduce the approach of using Retrieval Augmented Generation (RAG) powered agents to inject information into user prompts allowing for better inputs into embedding models. By utilizing RAG, agents enhance user queries with relevant details from GitHub repositories, making them more informative and contextually aligned. Additionally, we introduce a multi-stream ensemble approach which when paired with agentic workflow can obtain improved retrieval accuracy, which we deploy on application called repo-rift.com. Experimental results on the CodeSearchNet dataset demonstrate that RepoRift significantly outperforms existing methods, achieving an 78.2% success rate at Success@10 and a 34.6% success rate at Success@1. This research presents a substantial advancement in semantic code search, highlighting the potential of agentic LLMs and RAG to enhance code retrieval systems.

cs.SE↗

arXiv2024

ComFeAT: Combination of Neural and Spectral Features for Improved Depression Detection

In this work, we focus on the detection of depression through speech analysis. Previous research has widely explored features extracted from pre-trained models (PTMs) primarily trained for paralinguistic tasks. Although these features have led to sufficient advances in speech-based depression detection, their performance declines in real-world settings. To address this, in this paper, we introduce ComFeAT, an application that employs a CNN model trained on a combination of features extracted from PTMs, a.k.a. neural features and spectral features to enhance depression detection. Spectral features are robust to domain variations, but, they are not as good as neural features in performance, suprisingly, combining them shows complementary behavior and improves over both neural and spectral features individually. The proposed method also improves over previous state-of-the-art (SOTA) works on E-DAIC benchmark.

eess.AS↗

arXiv2024

PERSONA: An Application for Emotion Recognition, Gender Recognition and Age Estimation

Emotion Recognition (ER), Gender Recognition (GR), and Age Estimation (AE) constitute paralinguistic tasks that rely not on the spoken content but primarily on speech characteristics such as pitch and tone. While previous research has made significant strides in developing models for each task individually, there has been comparatively less emphasis on concurrently learning these tasks, despite their inherent interconnectedness. As such in this demonstration, we present PERSONA, an application for predicting ER, GR, and AE with a single model in the backend. One notable point is we show that representations from speaker recognition pre-trained model (PTM) is better suited for such a multi-task learning format than the state-of-the-art (SOTA) self-supervised (SSL) PTM by carrying out a comparative study. Our methodology obviates the need for deploying separate models for each task and can potentially conserve resources and time during the training and deployment phases.

eess.AS↗

arXiv2024

The Reasonable Effectiveness of Speaker Embeddings for Violence Detection

In this paper, we focus on audio violence detection (AVD). AVD is necessary for several reasons, especially in the context of maintaining safety, preventing harm, and ensuring security in various environments. This calls for accurate AVD systems. Like many related applications in audio processing, the most common approach for improving the performance, would be by leveraging self-supervised (SSL) pre-trained models (PTMs). However, as these SSL models are very large models with million of parameters and this can hinder real-world deployment especially in compute-constraint environment. To resolve this, we propose the usage of speaker recognition models which are much smaller compared to the SSL models. Experimentation with speaker recognition model embeddings with SVM & Random Forest as classifiers, we show that speaker recognition model embeddings perform the best in comparison to state-of-the-art (SOTA) SSL models and achieve SOTA results.

eess.AS↗

arXiv2024

Sparsity-Constrained Community-Based Group Testing

In this work, we consider the sparsity-constrained community-based group testing problem, where the population follows a community structure. In particular, the community consists of $F$ families, each with $M$ members. A number $k_f$ out of the $F$ families are infected, and a family is said to be infected if $k_m$ out of its $M$ members are infected. Furthermore, the sparsity constraint allows at most $ρ_T$ individuals to be grouped in each test. For this sparsity-constrained community model, we propose a probabilistic group testing algorithm that can identify the infected population with a vanishing probability of error and we provide an upper-bound on the number of tests. When $k_m = Θ(M)$ and $M \gg \log(FM)$, our bound outperforms the existing sparsity-constrained group testing results trivially applied to the community model. If the sparsity constraint is relaxed, our achievable bound reduces to existing bounds for community-based group testing. Moreover, our scheme can also be applied to the classical dilution model, where it outperforms existing noise-level-independent schemes in the literature.

cs.IT↗