Richa Singh

AAAI Conference 2026 Conference Paper

ACID Test: A Benchmark for Cultural Safety and Alignment in LALMs

• Bikash Dutta
• Adit Jain
• Rishabh Ranjan
• Mayank Vatsa
• Richa Singh

Large Audio Language Models (LALMs) are transforming AI by processing and generating human language directly from audio. As these models proliferate in real-world applications, it becomes critical to evaluate their performance to ensure equitable and safe use across diverse linguistic and cultural contexts. We present the first comprehensive study of cultural bias in LALMs, extending text-based harm frameworks to the audio modality to analyze how linguistic diversity influences model behavior and uncover challenges in interpreting audio nuances. To address this, we introduce the Audio Cultural Intelligence Dataset (ACID), a multilingual audio–text benchmark spanning 1,315 hours across diverse languages and cultural contexts, and we conduct a systematic evaluation of 10 open-source and two closed-source models. Our results reveal substantial performance disparities across languages and cultural settings and show that biases manifest distinctly when models process audio inputs. These findings highlight the need to evaluate LALMs not only for technical accuracy but also for fair and culturally sensitive behavior, motivating the development of inclusive datasets and culturally aware training practices for safer and more equitable audio language models.

TMLR Journal 2026 Journal Article

CatScreen: A Large MultiModal Benchmark Dataset for Cataract Screening

• Mahapara Khurshid
• Sonam Kumar
• Dr Anusuya Bhattacharyya
• Dhruve Kiyawat
• Anshul Chauhan
• Suklengmung Buragohain
• Harsha Bhattacharjee
• Limalemla Jamir

Low-cost slit-lamp imaging holds significant potential for transforming eye care by facilitating affordable and scalable cataract diagnosis. However, the development of robust, generalizable AI-based cataract screening solutions is currently constrained by the limited availability of large-scale, richly annotated datasets. To address this critical gap, we introduce CatScreen, a comprehensive multimodal benchmark dataset specifically designed for cataract screening, comprising approximately 18,000 slit-lamp images collected from 2,251 subjects using a portable slit-lamp camera. CatScreen is structured into three subsets: (i) a clean set meticulously annotated using a structured multi-tier framework involving trained optometrists with final validation by an experienced ophthalmologist across clinically relevant dimensions, including image gradability, quality assessment, illumination type, diagnostic classification, cataract subtype, and severity grading according to established standards; (ii) a noisy-labeled set that simulates real-world annotation inaccuracies; and (iii) an unlabeled set intended to foster the development of self-supervised and semi-supervised learning approaches. Furthermore, CatScreen integrates extensive subject-level metadata encompassing demographics, lifestyle factors, and detailed clinical histories, and includes a subset with anatomical and pathological annotations to support multimodal modeling and anatomically grounded analysis. We present baseline experiments under independent, structured sequential, and multitask prediction settings in both unimodal and multimodal configurations. These results establish initial benchmarks for CatScreen and demonstrate the value of metadata for selected diagnostic tasks, while also highlighting open challenges, such as class imbalance and fine-grained subtype discrimination. CatScreen is intended as a benchmark resource for future research in cataract screening, robust learning, semi-supervised learning, and interpretability-oriented analysis. The database is available at: https://iab-rubric.org/resources/healthcare-datasets/catscreen.

TIST Journal 2026 Journal Article

Introduction to the Special Issue on Transformers

• Feng Xia
• Tyler Derr
• Anh Tuan Luu
• Richa Singh
• Aline Villavicencio

AAAI Conference 2026 Conference Paper

NutriScreener: Retrieval Augmented Multi-Pose Graph Attention Network for Malnourishment Screening

• Misaal Khan
• Mayank Vatsa
• Kuldeep Singh
• Richa Singh

Child malnutrition remains a global crisis, yet existing screening methods are laborious and poorly scalable, hindering early intervention. In this work, we present NutriScreener, a retrieval-augmented, multi-pose graph attention network that combines CLIP-based visual embeddings, class-boosted knowledge retrieval, and context awareness to enable robust malnutrition detection and anthropometric prediction from children's images, simultaneously addressing generalizability and class-imbalance. In a clinical study, doctors rated it 4.3/5 for accuracy and 4.6/5 for efficiency, confirming its deployment readiness in low-resource settings. NutriScreener was trained and tested on 2,141 children from AnthroVision and additionally evaluated on diverse cross-continent populations, including ARAN and an in-house collected CampusPose dataset, achieving 0.79 recall, 0.82 AUC, and significantly lower anthropometric RMSEs, demonstrating reliable measurement in unconstrained, pediatric settings. Cross-dataset results show up to 25\% recall gain and up to 2.3 cm reduction in head circumference RMSE using demographically matched knowledge bases. NutriScreener offers a scalable and accurate solution for early malnutrition detection in low-resource environments.

AAAI Conference 2026 Conference Paper

Right Looks, Wrong Reasons: Compositional Fidelity in Text-to-Image Generation

• Mayank Vatsa
• Aparna Bharati
• Richa Singh

The architectural blueprint of today’s leading text-to-image models contains a fundamental flaw: an inability to handle logical composition. This survey investigates this breakdown across three core primitives—negation, counting, and spatial relations. Our analysis reveals a dramatic performance collapse: models that are accurate on single primitives fail precipitously when these are combined, exposing severe interference. We trace this failure to three key factors. First, training data show a near-total absence of explicit negations. Second, continuous attention architectures are fundamentally unsuitable for discrete logic. Third, evaluation metrics reward visual plausibility over constraint satisfaction. By analyzing recent benchmarks and methods, we show that current solutions and simple scaling cannot bridge this gap. Achieving genuine compositionality, we conclude, will require fundamental advances in representation and reasoning rather than incremental adjustments to existing architectures.

AAAI Conference 2025 Conference Paper

AQUAFace: Age-Invariant Quality Adaptive Face Recognition for Unconstrained Selfie vs ID Verification

• Shivang Agarwal
• Jyoti Chaudhary
• Sadiq Siraj Ebrahim
• Mayank Vatsa
• Richa Singh
• Shyam Prasad Adhikari
• Sangeeth Reddy Battu

Face recognition in the presence of age and quality variations poses a formidable challenge. While recent margin-based loss functions have shown promise in addressing these variations individually, real-world scenarios such as selfie versus ID face matching often involve simultaneous variations of both age and quality. In response, we propose a comprehensive framework aimed at mitigating the impact of these variations while preserving vital identity-related information crucial for accurate face recognition. The proposed adaptive margin-based loss function AQUAFace exhibits adaptiveness towards hard samples characterized by significant age and quality variations. This loss function is meticulously designed to prioritize the preservation of identity-related features while simultaneously mitigating the adverse effects of age and quality variations on face recognition accuracy. To validate the effectiveness of our approach, we focus on the specific task of selfie versus ID document matching. Our results demonstrate that AQUAFace effectively handles age and quality differences, leading to enhanced recognition performance. Additionally, we explore the benefits of fine-tuning the recognition model with synthetic data, further boosting performance. As a result, our proposed model, AQUAFace, achieves state-of-the-art performance on six benchmark datasets (CALFW, CPLFW, CFP-FP, AgeDB, IJB-C, and TinyFace), each exhibiting diverse age and quality variations.

TMLR Journal 2025 Journal Article

DIVINE: Diverse-Inconspicuous Feature Learning to Mitigate Abridge Learning

• Saheb Chhabra
• Kartik Thakral
• Surbhi Mittal
• Mayank Vatsa
• Richa Singh

Deep learning algorithms aim to minimize overall error and exhibit impressive performance on test datasets across various domains. However, they often struggle with out-of-distribution (OOD) data samples. We posit that deep models primarily capture prominent features beneficial for the task while neglecting subtle yet discriminative features, a phenomenon we refer to as Abridge Learning. To address this issue and encourage more comprehensive feature utilization, we introduce DIVINE (DIVerse and INconspicuous FEature Learning), a novel approach that leverages iterative feature suppression guided by dominance maps to ensure that models engage with a diverse and complementary set of discriminative features. Through extensive experiments on multiple datasets, including MNIST, CIFAR-10, CIFAR-100, TinyImageNet, and their corrupted and perturbed variants (CIFAR-10-C/P, CIFAR-100-C/P, TinyImageNet-C/P), we demonstrate that DIVINE significantly improves model robustness and generalization. On perturbation benchmarks, DIVINE achieves mean Flip Rates (mFR) of 5.36%, 3.10%, and 21.85% on CIFAR-10-P, CIFAR-100-P, and TinyImageNet-P respectively, compared to 6.53%, 11.75%, and 31.90% for standard training methods exhibiting Abridge Learning. Moreover, DIVINE attains state-of-the-art results on CIFAR-100-P, demonstrating that addressing Abridge Learning leads to more robust models against real-world distribution variations.

TMLR Journal 2025 Journal Article

IndicFake Meets SAFARI-LLM: Unifying Semantic and Acoustic Intelligence for Multilingual Deepfake Detection

• Rishabh Ranjan
• Mayank Vatsa
• Richa Singh

Audio deepfakes pose a growing threat, particularly in linguistically diverse and low-resource settings where existing detection methods often struggle. This work introduces two transformative contributions to address these challenges. First, we present \textbf{IndicFake}, a pioneering audio deepfake dataset with over 4.2 million samples (7,350 hours) spanning English and 17 Indian languages across Indo-European, Dravidian, and Sino-Tibetan families. With minimal overlap (Jaccard similarity: 0.00--0.06) with existing datasets, IndicFake offers an unparalleled benchmark for multilingual deepfake detection. Second, we propose \textbf{SAFARI-LLM} (Semantic Acoustic Feature Adaptive Router with Integrated LLM), a novel framework that integrates Whisper’s semantic embeddings and m-HuBERT’s acoustic features through an adaptive Audio Feature Unification Module (AFUM). Enhanced by LoRA-fine-tuned LLaMA-7B, SAFARI-LLM achieves unmatched cross-lingual and cross-family generalization. Evaluations across IndicFake, DECRO, and WaveFake datasets demonstrate its superiority, outperforming 14 state-of-the-art models with standout accuracies of 94.21\% (English-to-Japanese transfer on WaveFake) and 84.48\% (English-to-Chinese transfer on DECRO), alongside robust performance across diverse linguistic contexts. These advancements establish a new standard for reliable, scalable audio deepfake detection. Code and resources are publicly available at: https://anonymousillusion.github.io/indicfake/.

IJCAI Conference 2025 Conference Paper

SHIELD: A Self-supervised, Silicosis-focused Hierarchical Imaging Framework for Occupational Lung Disease Diagnosis

• Yasmeena Akhter
• Rishabh Ranjan
• Richa Singh
• Mayank Vatsa

Silicosis is an irreversible lung disease caused by silica dust exposure in industrial settings. Early detection is crucial, but automatic diagnostic methods are hindered by limited data availability. We propose SHIELD - a self-supervised, Silicosis-focused Hierarchical Imaging framework for early occupational Lung disease Diagnosis. Our method leverages a multi-resolution jigsaw puzzle pretext task on CXR images to extract and preserve features for lung region analysis. By employing a pyramidal strategy to generate pretrained models at various resolutions, followed by fine-tuning and a two-level ensembling across diverse deep learning architectures, SHIELD achieves enhanced diagnostic accuracy. We validate our approach on a publicly collected CXR dataset of 3044 samples from public health centers in India. SHIELD achieves 72% accuracy, demonstrating up to 20% improvement over baseline approaches. This work advances medical image analysis and supports UN Sustainable Development Goal 3 by providing cost-effective early screening in resource-limited settings.

IJCAI Conference 2025 Conference Paper

Words Over Pixels? Rethinking Vision in Multimodal Large Language Models

• Anubhooti Jain
• Mayank Vatsa
• Richa Singh

Multimodal Large Language Models (MLLMs) promise seamless integration of vision and language understanding. However, despite their strong performance, recent studies reveal that MLLMs often fail to effectively utilize visual information, frequently relying on textual cues instead. This survey provides a comprehensive analysis of the vision component in MLLMs, covering both application-level and architectural aspects. We investigate critical challenges such as weak spatial reasoning, poor fine-grained visual perception, and suboptimal fusion of visual and textual modalities. Additionally, we explore limitations in current vision encoders, benchmark inconsistencies, and their implications for downstream tasks. By synthesizing recent advancements, we highlight key research opportunities to enhance visual understanding, improve cross-modal alignment, and develop more robust and efficient MLLMs. Our observations emphasize the urgent need to elevate vision to an equal footing with language, paving the path for more reliable and perceptually aware multimodal models.

AAAI Conference 2024 Conference Paper

Adventures of Trustworthy Vision-Language Models: A Survey

• Mayank Vatsa
• Anubhooti Jain
• Richa Singh

Recently, transformers have become incredibly popular in computer vision and vision-language tasks. This notable rise in their usage can be primarily attributed to the capabilities offered by attention mechanisms and the outstanding ability of transformers to adapt and apply themselves to a variety of tasks and domains. Their versatility and state-of-the-art performance have established them as indispensable tools for a wide array of applications. However, in the constantly changing landscape of machine learning, the assurance of the trustworthiness of transformers holds utmost importance. This paper conducts a thorough examination of vision-language transformers, employing three fundamental principles of responsible AI: Bias, Robustness, and Interpretability. The primary objective of this paper is to delve into the intricacies and complexities associated with the practical use of transformers, with the overarching goal of advancing our comprehension of how to enhance their reliability and accountability.

AAAI Conference 2024 Conference Paper

BirdCollect: A Comprehensive Benchmark for Analyzing Dense Bird Flock Attributes

• Kshitiz.
• Sonu Shreshtha
• Bikash Dutta
• Muskan Dosi
• Mayank Vatsa
• Richa Singh
• Saket Anand
• Sudeep Sarkar

Automatic recognition of bird behavior from long-term, un controlled outdoor imagery can contribute to conservation efforts by enabling large-scale monitoring of bird populations. Current techniques in AI-based wildlife monitoring have focused on short-term tracking and monitoring birds individually rather than in species-rich flocks. We present Bird-Collect, a comprehensive benchmark dataset for monitoring dense bird flock attributes. It includes a unique collection of more than 6,000 high-resolution images of Demoiselle Cranes (Anthropoides virgo) feeding and nesting in the vicinity of Khichan region of Rajasthan. Particularly, each image contains an average of 190 individual birds, illustrating the complex dynamics of densely populated bird flocks on a scale that has not previously been studied. In addition, a total of 433 distinct pictures captured at Keoladeo National Park, Bharatpur provide a comprehensive representation of 34 distinct bird species belonging to various taxonomic groups. These images offer details into the diversity and the behaviour of birds in vital natural ecosystem along the migratory flyways. Additionally, we provide a set of 2,500 point-annotated samples which serve as ground truth for benchmarking various computer vision tasks like crowd counting, density estimation, segmentation, and species classification. The benchmark performance for these tasks highlight the need for tailored approaches for specific wildlife applications, which include varied conditions including views, illumination, and resolutions. With around 46.2 GBs in size encompassing data collected from two distinct nesting ground sets, it is the largest birds dataset containing detailed annotations, showcasing a substantial leap in bird research possibilities. We intend to publicly release the dataset to the research community. The database is available at: https://iab-rubric.org/resources/wildlife-dataset/birdcollect

TMLR Journal 2023 Journal Article

Dropped Scheduled Task: Mitigating Negative Transfer in Multi-task Learning using Dynamic Task Dropping

• Aakarsh Malhotra
• Mayank Vatsa
• Richa Singh

In Multi-Task Learning (MTL), K distinct tasks are jointly optimized. With the varying nature and complexities of tasks, few tasks might dominate learning. For other tasks, their respective performances may get compromised due to a negative transfer from dominant tasks. We propose a Dropped-Scheduled Task (DST) algorithm, which probabilistically “drops” specific tasks during joint optimization while scheduling others to reduce negative transfer. For each task, a scheduling probability is decided based on four different metrics: (i) task depth, (ii) number of ground-truth samples per task, (iii) amount of training completed, and (iv) task stagnancy. Based on the scheduling probability, specific tasks get joint computation cycles while others are “dropped”. To demonstrate the effectiveness of the proposed DST algorithm, we perform multi-task learning on three applications and two architectures. Across unilateral (single input) and bilateral (multiple input) multi-task net- works, the chosen applications are (a) face (AFLW), (b) fingerprint (IIITD MOLF, MUST, and NIST SD27), and (c) character recognition (Omniglot) applications. Experimental results show that the proposed DST algorithm has the minimum negative transfer and overall least errors across different state-of-the-art algorithms and tasks.

AAAI Conference 2023 Short Paper

IdProv: Identity-Based Provenance for Synthetic Image Generation (Student Abstract)

• Harshil Bhatia
• Jaisidh Singh
• Gaurav Sangwan
• Aparna Bharati
• Richa Singh
• Mayank Vatsa

Recent advancements in Generative Adversarial Networks (GANs) have made it possible to obtain high-quality face images of synthetic identities. These networks see large amounts of real faces in order to learn to generate realistic looking synthetic images. However, the concept of a synthetic identity for these images is not very well-defined. In this work, we verify identity leakage from the training set containing real images into the latent space and propose a novel method, IdProv, that uses image composition to trace the source of identity signals in the generated image.

IJCAI Conference 2023 Conference Paper

Long-term Monitoring of Bird Flocks in the Wild

• Kshitiz
• Sonu Shreshtha
• Ramy Mounir
• Mayank Vatsa
• Richa Singh
• Saket Anand
• Sudeep Sarkar
• Sevaram Mali Parihar

Monitoring and analysis of wildlife are key to conservation planning and conflict management. The widespread use of camera traps coupled with AI-based analysis tools serves as an excellent example of successful and non-invasive use of technology for design, planning, and evaluation of conservation policies. As opposed to the typical use of camera traps that capture still images or short videos, in this project, we propose to analyze longer term videos monitoring a large flock of birds. This project, which is part of the NSF-TIH Indo-US joint R&D partnership, focuses on solving challenges associated with the analysis of long-term videos captured at feeding grounds and nesting sites, among other such locations that host large flocks of migratory birds. We foresee that the objectives of this project would lead to datasets and benchmarking tools as well as novel algorithms that would be instrumental in developing automated video analysis tools that could in turn help understand individual and social behavior of birds. The first of the key outcomes of this research will include the curation of challenging, real-world datasets for benchmarking various image and video analytics algorithms for tasks such as counting, detection, segmentation, and tracking. Our recent efforts towards this outcome is a curated dataset of 812 high-resolution, point-annotated, images (4K - 32MP) of a flock of Demoiselle cranes (Anthropoides virgo) taken from their feeding site at Khichan, Rajasthan, India. The average number of birds in each image is about 207, with a maximum count of 1500. The benchmark experiments show that state-of-the-art vision techniques struggle with tasks such as segmentation, detection, localization, and density estimation for the proposed dataset. Over the execution of this open science research, we will be scaling this dataset for segmentation and tracking in videos, as well as developing novel techniques for video analytics for wildlife monitoring.

IJCAI Conference 2023 Conference Paper

NutriAI: AI-Powered Child Malnutrition Assessment in Low-Resource Environments

• Misaal Khan
• Shivang Agarwal
• Mayank Vatsa
• Richa Singh
• Kuldeep Singh

Malnutrition among infants and young children is a pervasive public health concern, particularly in developing countries where resources are limited. Millions of children globally suffer from malnourishment and its complications1. Despite the best efforts of governments and organizations, malnourishment persists and remains a leading cause of morbidity and mortality among children under five. Physical measurements, such as weight, height, middle-upper-arm-circumference (muac), and head circumference are commonly used to assess the nutritional status of children. However, this approach can be resource-intensive and challenging to carry out on a large scale. In this research, we are developing NutriAI, a low-cost solution that leverages small sample size classification approach to detect malnutrition by analyzing 2D images of the subjects in multiple poses. The proposed solution will not only reduce the workload of health workers but also provide a more efficient means of monitoring the nutritional status of children. On the dataset prepared as part of this research, the baseline results highlight that the modern deep learning approaches can facilitate malnutrition detection via anthropometric indicators in the presence of diversity with respect to age, gender, physical characteristics, and accessories including clothing.

IJCAI Conference 2023 Conference Paper

On AI-Assisted Pneumoconiosis Detection from Chest X-rays

• Yasmeena Akhter
• Rishabh Ranjan
• Richa Singh
• Mayank Vatsa
• Santanu Chaudhury

According to theWorld Health Organization, Pneumoconiosis affects millions of workers globally, with an estimated 260, 000 deaths annually. The burden of Pneumoconiosis is particularly high in low-income countries, where occupational safety standards are often inadequate, and the prevalence of the disease is increasing rapidly. The reduced availability of expert medical care in rural areas, where these diseases are more prevalent, further adds to the delayed screening and unfavourable outcomes of the disease. This paper aims to highlight the urgent need for early screening and detection of Pneumoconiosis, given its significant impact on affected individuals, their families, and societies as a whole. With the help of low-cost machine learning models, early screening, detection, and prevention of Pneumoconiosis can help reduce healthcare costs, particularly in low-income countries. In this direction, this research focuses on designing AI solutions for detecting different kinds of Pneumoconiosis from chest X-ray data. This will contribute to the Sustainable Development Goal 3 of ensuring healthy lives and promoting well-being for all at all ages, and present the framework for data collection and algorithm for detecting Pneumoconiosis for early screening. The baseline results show that the existing algorithms are unable to address this challenge. Therefore, it is our assertion that this research will improve state-of-the-art algorithms of segmentation, semantic segmentation, and classification not only for this disease but in general medical image analysis literature.

IJCAI Conference 2023 Conference Paper

Uncovering the Deceptions: An Analysis on Audio Spoofing Detection and Future Prospects

• Rishabh Ranjan
• Mayank Vatsa
• Richa Singh

Audio has become an increasingly crucial biometric modality due to its ability to provide an intuitive way for humans to interact with machines. It is currently being used for a range of applications including person authentication to banking to virtual assistants. Research has shown that these systems are also susceptible to spoofing and attacks. Therefore, protecting audio processing systems against fraudulent activities such as identity theft, financial fraud, and spreading misinformation, is of paramount importance. This paper reviews the current state-of-the-art techniques for detecting audio spoofing and discusses the current challenges along with open research problems. The paper further highlights the importance of considering the ethical and privacy implications of audio spoofing detection systems. Lastly, the work aims to accentuate the need for building more robust and generalizable methods, the integration of automatic speaker verification and countermeasure systems, and better evaluation protocols.

AAAI Conference 2022 Conference Paper

Anatomizing Bias in Facial Analysis

• Richa Singh
• Puspita Majumdar
• Surbhi Mittal
• Mayank Vatsa

Existing facial analysis systems have been shown to yield biased results against certain demographic subgroups. Due to its impact on society, it has become imperative to ensure that these systems do not discriminate based on gender, identity, or skin tone of individuals. This has led to research in the identification and mitigation of bias in AI systems. In this paper, we encapsulate bias detection/estimation and mitigation algorithms for facial analysis. Our main contributions include a systematic review of algorithms proposed for understanding bias, along with a taxonomy and extensive overview of existing bias mitigation algorithms. We also discuss open challenges in the field of biased facial analysis.

AAAI Conference 2021 Short Paper

Detection of Digital Manipulation in Facial Images (Student Abstract)

• Aman Mehra
• Akshay Agarwal
• Mayank Vatsa
• Richa Singh

Advances in deep learning have enabled the creation of photo-realistic DeepFakes by switching the identity or expression of individuals. Such technology in the wrong hands can seed chaos through blackmail, extortion, and forging false statements of influential individuals. This work proposes a novel approach to detect forged videos by magnifying their temporal inconsistencies. A study is also conducted to understand role of ethnicity bias due to skewed datasets on deepfake detection. A new dataset comprising forged videos of Indian ethnicity individuals is presented to facilitate this study.

AAAI Conference 2021 Short Paper

NEAP-F: Network Epoch Accuracy Prediction Framework (Student Abstract)

• Arushi Chauhan
• Mayank Vatsa
• Richa Singh

Recent work in neural architecture search has spawned interest in algorithms that can predict the performance neural networks using minimum time and computation resources. We propose a new framework, Network Epoch Accuracy Prediction Framework (NEAP-F) which can predict the testing accuracy achieved by a convolutional neural network in one or more epochs. We introduce a novel approach to generate vector representations for networks, and encode “ease” of classifying image datasets into a vector. For vector representations of networks, we focus on the layer parameters and connections between the network layers. A network achieves different accuracy on different image datasets; therefore, we use the image dataset characteristics to create a vector signifying the “ease” of classifying the image dataset. After generating these vectors, the prediction models are trained with architectures having skip connections seen in current state-of-the-art architectures. The framework predicts accuracies in order of milliseconds, demonstrating its computational efficiency. It can be easily applied to neural architecture search methods to predict the performance of candidate networks and can work on unseen datasets as well.

AAAI Conference 2021 Short Paper

On Learning Deep Models with Imbalanced Data Distribution

• Puspita Majumdar
• Richa Singh
• Mayank Vatsa

The availability of large training data has led to the development of sophisticated deep learning algorithms to achieve state-of-the-art performance on various tasks and several applications have been benefited immensely. Despite the unparalleled success, the performance of deep learning algorithms depends significantly on the training data distribution. An imbalance in training data distribution affects the performance of deep models. Our research focuses on designing and developing solutions for different real-world problems, specifically related to facial analytic tasks, with imbalanced data distribution. These problems include injured face recognition, fake image detection, and estimation and mitigation of bias in model prediction.

AAAI Conference 2021 Short Paper

Role of Optimizer on Network Fine-tuning for Adversarial Robustness (Student Abstract)

• Akshay Agarwal
• Mayank Vatsa
• Richa Singh

The solutions proposed in the literature for adversarial robustness are either not effective against the challenging gradientbased attacks or are computationally demanding, such as adversarial training. Adversarial training or network training based data augmentation shows the potential to increase the adversarial robustness. While the training seems compelling, it is not feasible for resource-constrained institutions, especially academia, to train the network from scratch multiple times. The two fold contributions are: (i) providing an effective solution against white-box adversarial attacks via network fine-tuning steps and (ii) observing the role of different optimizers towards robustness. Extensive experiments are performed on a range of databases, including Fashion- MNIST and a subset of ImageNet. It is found that the few steps of network fine-tuning effectively increases the robustness of both shallow and deep architectures. To know other interesting observations, especially regarding the role of the optimizer, refer to the paper.

AAAI Conference 2021 Short Paper

Semi-Supervised Learning via Triplet Network Based Active Learning (Student Abstract)

• Divyanshu Sundriyal
• Soumyadeep Ghosh
• Mayank Vatsa
• Richa Singh

In recent years deep learning models have pushed state-ofthe-art accuracies for several machine learning tasks. However, such models require a large amount of data for training. Active learning techniques help us in utilizing unlabelled data which may result in an improved classification model. In this research, we present an active learning algorithm which can help in increasing performance of deep learning models by using large amount of available unlabelled data. A novel active learning algorithm (Triplet AL) is proposed which uses a triplet network to select samples from an unlabelled data set. Previous active learning methods rely on classification model’s final prediction scores as a measure of confidence for an unlabelled sample. We propose a more reliable confidence measure called Top-Two-Margin which is given by Triplet Network. The proposed algorithm shows improved performance compared to other active learning approaches.

AAAI Conference 2020 Conference Paper

On the Robustness of Face Recognition Algorithms Against Attacks and Bias

• Richa Singh
• Akshay Agarwal
• Maneet Singh
• Shruti Nagpal
• Mayank Vatsa

Face recognition algorithms have demonstrated very high recognition performance, suggesting suitability for real world applications. Despite the enhanced accuracies, robustness of these algorithms against attacks and bias has been challenged. This paper summarizes different ways in which the robustness of a face recognition algorithm is challenged, which can severely affect its intended working. Different types of attacks such as physical presentation attacks, disguise/makeup, digital adversarial attacks, and morphing/tampering using GANs have been discussed. We also present a discussion on the effect of bias on face recognition models and showcase that factors such as age and gender variations affect the performance of modern algorithms. The paper also presents the potential reasons for these challenges and some of the future research directions for increasing the robustness of face recognition models.

AAAI Conference 2019 Conference Paper

Data Fine-Tuning

• Saheb Chhabra
• Puspita Majumdar
• Mayank Vatsa
• Richa Singh

In real-world applications, commercial off-the-shelf systems are utilized for performing automated facial analysis including face recognition, emotion recognition, and attribute prediction. However, a majority of these commercial systems act as black boxes due to the inaccessibility of the model parameters which makes it challenging to fine-tune the models for specific applications. Stimulated by the advances in adversarial perturbations, this research proposes the concept of Data Fine-tuning to improve the classification accuracy of a given model without changing the parameters of the model. This is accomplished by modeling it as data (image) perturbation problem. A small amount of “noise” is added to the input with the objective of minimizing the classification loss without affecting the (visual) appearance. Experiments performed on three publicly available datasets LFW, CelebA, and MUCT, demonstrate the effectiveness of the proposed concept.

AAAI Conference 2019 Conference Paper

Guided Dropout

• Rohit Keshari
• Richa Singh
• Mayank Vatsa

Dropout is often used in deep neural networks to prevent over-fitting. Conventionally, dropout training invokes random drop of nodes from the hidden layers of a Neural Network. It is our hypothesis that a guided selection of nodes for intelligent dropout can lead to better generalization as compared to the traditional dropout. In this research, we propose “guided dropout” for training deep neural network which drop nodes by measuring the strength of each node. We also demonstrate that conventional dropout is a specific case of the proposed guided dropout. Experimental evaluation on multiple datasets including MNIST, CIFAR10, CIFAR100, SVHN, and Tiny ImageNet demonstrate the efficacy of the proposed guided dropout.

IJCAI Conference 2018 Conference Paper

Anonymizing k Facial Attributes via Adversarial Perturbations

• Saheb Chhabra
• Richa Singh
• Mayank Vatsa
• Gaurav Gupta

A face image not only provides details about the identity of a subject but also reveals several attributes such as gender, race, sexual orientation, and age. Advancements in machine learning algorithms and popularity of sharing images on the World Wide Web, including social media websites, have increased the scope of data analytics and information profiling from photo collections. This poses a serious privacy threat for individuals who do not want to be profiled. This research presents a novel algorithm for anonymizing selective attributes which an individual does not want to share without affecting the visual quality of images. Using the proposed algorithm, a user can select single or multiple attributes to be surpassed while preserving identity information and visual content. The proposed adversarial perturbation based algorithm embeds imperceptible noise in an image such that attribute prediction algorithm for the selected attribute yields incorrect classification result, thereby preserving the information according to user's choice. Experiments on three popular databases i. e. MUCT, LFWcrop, and CelebA show that the proposed algorithm not only anonymizes \textit{k}-attributes, but also preserves image quality and identity information.

AAAI Conference 2018 Conference Paper

Unravelling Robustness of Deep Learning Based Face Recognition Against Adversarial Attacks

• Gaurav Goswami
• Nalini Ratha
• Akshay Agarwal
• Richa Singh
• Mayank Vatsa

Deep neural network (DNN) architecture based models have high expressive power and learning capacity. However, they are essentially a black box method since it is not easy to mathematically formulate the functions that are learned within its many layers of representation. Realizing this, many researchers have started to design methods to exploit the drawbacks of deep learning based algorithms questioning their robustness and exposing their singularities. In this paper, we attempt to unravel three aspects related to the robustness of DNNs for face recognition: (i) assessing the impact of deep architectures for face recognition in terms of vulnerabilities to attacks inspired by commonly observed distortions in the real world that are well handled by shallow learning methods along with learning based adversaries; (ii) detecting the singularities by characterizing abnormal filter response behavior in the hidden layers of deep networks; and (iii) making corrections to the processing pipeline to alleviate the problem. Our experimental evaluation using multiple open-source DNN-based face recognition networks, including OpenFace and VGG-Face, and two publicly available databases (MEDS and PaSC) demonstrates that the performance of deep learning based face recognition algorithms can suffer greatly in the presence of such distortions. The proposed method is also compared with existing detection algorithms and the results show that it is able to detect the attacks with very high accuracy by suitably designing a classifier using the response of the hidden layers in the network. Finally, we present several effective countermeasures to mitigate the impact of adversarial attacks and improve the overall robustness of DNN-based face recognition.