Explanipedia

Scaling Laws Revisited: Modeling the Role of Data Quality in Language Model Pretraining Open

A V Subramanyam, Yuxin Chen, Robert L. Grossman · 2025

Scaling laws for language model training traditionally characterize how performance scales with model size and dataset volume. Prior work has explored architecture variants and data treatments such as dataset filtering and noise injection …

Nearest Neighbor Projection Removal Adversarial Training Open

Himanshu Singh, A V Subramanyam, S.K.S. Rajput, Mohan Kankanhalli · 2025

Deep neural networks have exhibited impressive performance in image classification tasks but remain vulnerable to adversarial examples. Standard adversarial training enhances robustness but typically fails to explicitly address inter-class…

Boosting Weak Positives for Text Based Person Search Open

Akshay Modi, Ashhar Aziz, Nilanjana Chatterjee, A V Subramanyam · 2025

Large vision-language models have revolutionized cross-modal object retrieval, but text-based person search (TBPS) remains a challenging task due to limited data and fine-grained nature of the task. Existing methods primarily focus on alig…

LaB-RAG: Label Boosted Retrieval Augmented Generation for Radiology Report Generation Open

Steven Song, A V Subramanyam, Irene Madejski, Robert L. Grossman · 2024

In the current paradigm of image captioning, deep learning models are trained to generate text from image embeddings of latent features. We challenge the assumption that fine-tuning of large, bespoke models is required to improve model gen…

Keypoint Aware Masked Image Modelling Open

Madhava Krishna, A V Subramanyam · 2024

SimMIM is a widely used method for pretraining vision transformers using masked image modeling. However, despite its success in fine-tuning performance, it has been shown to perform sub-optimally when used for linear probing. We propose an…

Resource Efficient Perception for Vision Systems Open

A V Subramanyam, Niyati Singal, Vinay Kumar Verma · 2024

Despite the rapid advancement in the field of image recognition, the processing of high-resolution imagery remains a computational challenge. However, this processing is pivotal for extracting detailed object insights in areas ranging from…

Image Synthesis with Graph Conditioning: CLIP-Guided Diffusion Models for Scene Graphs Open

Rameshwar Mishra, A V Subramanyam · 2024

Advancements in generative models have sparked significant interest in generating images while adhering to specific structural guidelines. Scene graph to image generation is one such task of generating images which are consistent with the …

Language Guided Adversarial Purification Open

Himanshu Singh, A V Subramanyam · 2023

Adversarial purification using generative models demonstrates strong adversarial defense performance. These methods are classifier and attack-agnostic, making them versatile but often computationally intensive. Recent strides in diffusion …

IIITD-20K: Dense captioning for Text-Image ReID Open

A V Subramanyam, N. Sundararajan, Vibhu Dubey, Brejesh Lall · 2023

Text-to-Image (T2I) ReID has attracted a lot of attention in the recent past. CUHK-PEDES, RSTPReid and ICFG-PEDES are the three available benchmarks to evaluate T2I ReID methods. RSTPReid and ICFG-PEDES comprise of identities from MSMT17 b…

Certified Zeroth-order Black-Box Defense with Robust UNet Denoiser Open

Astha Verma, Siddhesh Bangar, A V Subramanyam, Naman Lal, Rajiv Ratn Shah , et al. · 2023

Certified defense methods against adversarial perturbations have been recently investigated in the black-box setting with a zeroth-order (ZO) perspective. However, these methods suffer from high model variance with low performance on high-…

Meta Generative Attack on Person Reidentification Open

A V Subramanyam · 2023

Adversarial attacks have been recently investigated in person re-identification. These attacks perform well under cross dataset or cross model setting. However, the challenges present in cross-dataset cross-model scenario does not allow th…

Antenna Design Optimization using GAN-based Surrogate Model Open

Kainat Yasmeen, Kumar Vijay Mishra, A V Subramanyam, Shobha Sundar Ram · 2023

Deep neural network (DNN) based surrogate models have been used to rapidly create antenna characteristics as a function of design geometry parameters in place of computationally expensive electromagnetic (EM) solvers.The limitation of sing…

Novel deep learning framework for wideband spectrum characterization at sub-Nyquist rate Open

Shivam Chandhok, Himani Joshi, A V Subramanyam, Sumit J. Darak · 2021

OneStopTuner: An End to End Architecture for JVM Tuning of Spark Applications Open

V Venktesh, Pooja B Bindal, Devesh Singhal, A V Subramanyam, Vivek Kumar · 2020

Java is the backbone of widely used big data frameworks, such as Apache Spark, due to its productivity, portability from JVM-based execution, and support for a rich set of libraries. However, the performance of these applications can widel…

Novel Deep Learning Framework for Wideband Spectrum Characterization at\n Sub-Nyquist Rate Open

Shivam Chandhok, Himani Joshi, A V Subramanyam, Sumit J. Darak · 2019

Introduction of spectrum-sharing in 5G and subsequent generation networks\ndemand base-station(s) with the capability to characterize the wideband\nspectrum spanned over licensed, shared and unlicensed non-contiguous frequency\nbands. Spec…

SenseNet: Deep Learning based Wideband spectrum sensing and modulation classification network. Open

Shivam Chandhok, Himani Joshi, A V Subramanyam, Sumit J. Darak · 2019

Next generation networks are expected to operate in licensed, shared as well as unlicensed spectrum to support spectrum demands of a wide variety of services.Due to shortage of radio spectrum, the need for communication systems(like cognit…

Attentional Road Safety Networks Open

Sonu Gupta, Deepak Srivatsav, A V Subramanyam, Ponnurangam Kumaraguru · 2019

Road safety mapping using satellite images is a cost-effective but a challenging problem for smart city planning. The scarcity of labeled data, misalignment and ambiguity makes it hard for supervised deep networks to learn efficient embedd…

CARF-Net: CNN attention and RNN fusion network for video-based person reidentification Open

Kajal Kansal, A V Subramanyam, Dilip K. Prasad, Mohan Kankanhalli · 2019

Video-based person reidentification is a challenging and important task in surveillance-based applications. Toward this, several shallow and deep networks have been proposed. However, the performance of existing shallow networks does not g…

Hdrnet: Person Re-Identification Using Hybrid Sampling in Deep Reconstruction Network Open

Kajal Kansal, A V Subramanyam · 2019

Person re-identification (re-id) is the task of identifying a person across non-overlapping cameras. Most of the current techniques apply deep learning and achieve a significant accuracy. However, learning a deep model that can generalize …

Joint Spatial and Discrete Cosine Transform Domain-Based Counter Forensics for Adaptive Contrast Enhancement Open

Ambuj Mehrish, A V Subramanyam, Sabu Emmanuel · 2019

Contrast enhancement (CE) is a common post-processing step in image forgery to create visually convincing tampered images. However, the artifacts embedded during this process can be captured to determine the presence of CE. To overcome the…

Framework on Quality of Service in Mobile Ad Hoc Networks Open

C. Siva Krishnaiah, A V Subramanyam · 2017

Mobile Ad Hoc Network (MANET) has been continuously self-configuring network composed of a set of mobile devices that can communicate between them without infrastructure connected wirelessly.With the expanding scope of MANET applications t…

PicHunt: Social Media Image Retrieval for Improved Law Enforcement Open

Sonal Goel, Niharika Sachdeva, Ponnurangam Kumaraguru, A V Subramanyam, Divam Gupta · 2016

First responders are increasingly using social media to identify and reduce crime for well-being and safety of the society. Images shared on social media hurting religious, political, communal and other sentiments of people, often instigat…

Prevalence of hypothyroidism in common bile duct stone patients Open

Srinivas Sidduri, Babulreddy Hanmayyagari, Vivekanand Bongi, Mythili Ayyagari, A V Subramanyam · 2015

Background: Hypothyroidism affects bile content, bile flow, and functions of the sphincter of Oddi, thereby increases the formation of common bile duct (CBD) stones. The exact prevalence of hypothyroidism in CBD stone patients is not known…

A V Subramanyam YOU? Author Swipe