Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Anindya Bijoy Das has a total of 21 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Explainable Artificial Intelligence for Predicting Gastrointestinal Adverse Effects of GLP-1 Receptor Agonists

    by Tadesse M. Abegaz, Gabriel Frietze, Anindya Bijoy Das

    Gastrointestinal (GI) adverse drug reactions (ADRs) are common among glucagon-like peptide-1 receptor agonist (GLP-1 RA) users and frequently contribute to treatment discontinuation and reduced therapeutic benefit. This retrospective study aimed to develop and validate an explainable artificial intelligence (XAI) model to predict GI ADR risk among GLP-1 RA users using real-world clinical data from the NIH All of Us Research Program. Adults prescribed GLP-1 RAs were identified and classified according to the occurrence of GI ADRs following treatment initiation. Multiple supervised machine learning models, including logistic regression, random forest, extreme gradient boosting (XGBoost), support vector machine, neural network, LightGBM, and CatBoost, were evaluated using demographic, socioeconomic, clinical, medication, and laboratory variables. Model performance was assessed using area under the receiver operating characteristic curve (AUC), accuracy, precision, recall, and F1-score. A total of 8697 participants were included, of whom 59.1% experienced GI ADRs. All models demonstrated reasonable predictive performance, with AUC values ranging from 0.82 to 0.84. The XGBoost achieved discrimination of (AUC: 0.84 ± 0.01). SHapley Additive exPlanations (SHAP) identified gastroesophageal reflux disease, hemorrhoids, and elevated HbA1c as important predictors of GI ADR risk. These findings demonstrate the potential utility of explainable machine learning approaches for predicting the safety of GLP-1 RA therapy.

  • On Targeted Output Generation via Adversarial Reframing in Language Models

    by Shahnewaz Karim Sakib, Swati Kar, Anindya Bijoy Das

  • A Survey on Coded Matrix Computations and Gradient Coding

    by Aditya Ramamoorthy, Anindya Bijoy Das

    In modern times, processing and storage needs have made distributed computing the technology of choice in various domains. How ever, large-scale distributed clusters (especially within the cloud) often suffer from the problem of stragglers (slow or failed workers). Coded distributed computation focuses on straggler mitigation ideas that have their roots in coding theory, i.e., it judiciously introduces redundancy across tasks assigned to workers such that the desired result can be obtained even in the presence of stragglers. Within distributed matrix computations and distributed parameter learning, it has led to key results on the trade-off between redundancy and the level of straggler mitigation that can be supported. In this survey, we overview the key ideas underpinning coded computation for both these applications. Following this, we offer insights about important considerations such as the numerical stability of recovery, leveraging partial work from slow (but not failed) workers, and considering scenarios where the inputs are sparse; we also discuss several works that have addressed these issues. We conclude with a discussion of open problems and future directions within these broad areas.

  • Cooperative Decentralized Backdoor Attacks on Vertical Federated Learning

    by Seohyun Lee, Wenzhi Fang, Anindya Bijoy Das, Seyyedali Hosseinalipour, David J. Love, Christopher G. Brinton

    Federated learning (FL) is vulnerable to backdoor attacks, where adversaries alter model behavior on target classification labels by embedding triggers into data samples. While these attacks have received considerable attention in horizontal FL, they are less understood for vertical FL (VFL), where devices hold different features of the samples, and only the server holds the labels. In this work, we propose a novel backdoor attack on VFL which (i) does not rely on gradient information from the server and (ii) considers potential collusion among multiple adversaries for sample selection and trigger embedding. Our label inference model augments variational autoencoders with metric learning, which adversaries can train locally. A consensus process over the adversary graph topology determines which datapoints to poison. We further propose methods for trigger splitting across the adversaries, with an intensity-based implantation scheme skewing the server towards the trigger. Our convergence analysis reveals the impact of backdoor perturbations on VFL indicated by a stationarity gap for the trained model, which we verify empirically as well. We conduct experiments comparing our attack with recent backdoor VFL approaches, finding that ours obtains significantly higher success rates for the same main task performance despite not using server information. Additionally, our results verify the impact of collusion on attack performance.


  • Open Access

    Trustworthy Medical Imaging with Large Language Models: A Study of Hallucinations Across Modalities

    by Anindya Bijoy Das, Shahnewaz Karim Sakib, Shibbir Ahmed

    Large Language Models (LLMs) are increasingly applied to medical imaging tasks, including image interpretation and synthetic image generation. However, these models often produce hallucinations, which are confident but incorrect outputs that can mislead clinical decisions. This study examines hallucinations in two directions: image to text, where LLMs generate reports from X-ray, CT, or MRI scans, and text to image, where models create medical images from clinical prompts. We analyze errors such as factual inconsistencies and anatomical inaccuracies, evaluating outputs using expert informed criteria across imaging modalities. Our findings reveal common patterns of hallucination in both interpretive and generative tasks, with implications for clinical reliability. We also discuss factors contributing to these failures, including model architecture and training data. By systematically studying both image understanding and generation, this work provides insights into improving the safety and trustworthiness of LLM driven medical imaging systems.


  • Learning-Based Two-Way Communications: Algorithmic Framework and Comparative Analysis

    by David R. Nickel, Anindya Bijoy Das, David J. Love, Christopher G. Brinton

    Machine learning (ML)-based feedback channel coding has garnered significant research interest in the past few years. However, there has been limited research exploring ML approaches in the so-called “two-way" setting where two users jointly encode messages and feedback over a shared channel. In this work, we present a general architecture for ML-based two-way feedback coding, and show how several popular one-way schemes can be converted to the two-way setting through our algorithmic framework. We compare such schemes against one-way counterparts, revealing error-rate benefits of ML-based two-way coding in certain signal-to-noise ratio (SNR) regimes. We then analyze the tradeoffs between error performance and computational overhead for three state-of-the-art neural network coding models instantiated in the two-way paradigm.

  • Coding for Gaussian Two-Way Channels: Linear and Learning-Based Approaches

    by Junghoon Kim, Taejoon Kim, Anindya Bijoy Das, Seyyedali Hosseinalipour, David J. Love, Christopher G. Brinton

    Although user cooperation cannot improve the capacity of Gaussian two-way channels (GTWCs) with independent noises, it can improve communication reliability. In this work, we aim to enhance and balance the communication reliability in GTWCs by minimizing the sum of error probabilities via joint design of encoders and decoders at the users. We first formulate general encoding/decoding functions, where the user cooperation is captured by the coupling of user encoding processes. The coupling effect renders the encoder/decoder design non-trivial, requiring effective decoding to capture this effect, as well as efficient power management at the encoders within power constraints. To address these challenges, we propose two different twoway coding strategies: linear coding and learning-based coding. For linear coding, we propose optimal linear decoding and discuss new insights on encoding regarding user cooperation to balance reliability. We then propose an efficient algorithm for joint encoder/decoder design. For learning-based coding, we introduce a novel recurrent neural network (RNN)-based coding architecture, where we propose interactive RNNs and a power control layer for encoding, and we incorporate bi-directional RNNs with an attention mechanism for decoding. Through simulations, we show that our two-way coding methodologies outperform conventional channel coding schemes (that do not utilize user cooperation) significantly in sum-error performance. We also demonstrate that our linear coding excels at high signal-to-noise ratios (SNRs), while our RNN-based coding performs best at low SNRs. We further investigate our two-way coding strategies in terms of power distribution, two-way coding benefit, different coding rates, and block-length gain.


  • Open Access

    Breaking the Shield: Vulnerabilities in Content Moderation for Multimodal Language Models

    by Anindya Bijoy Das, Shahnewaz Karim Sakib

    Text-to-image generation models, especially with recent breakthroughs like GPT-4o's image generation, have rapidly transformed how visual content is created and shared across domains. Their ability to produce visually compelling images from natural language prompts has made them increasingly influential in digital media. However, a growing concern is the generation of misinformation or discriminatory information (e.g., images that visually encode or propagate widely debunked or false information) often in a persuasive and authoritative format. In this paper, we demonstrate how multimodal large language model (MMLLM) based image generation systems can be prompted to produce misinformation and discriminatory content. Additionally, we formulate the problem of developing a generalizable prompt template that adversaries could exploit to bypass the system guardrails. Our hypothesis is that prompts framed with narrative context are more successful in evading safeguards than overtly harmful ones. We conduct extensive simulations to validate these strategies and demonstrate the potential impact of the threat.


  • Multi-Agent Reinforcement Learning for Graph Discovery in D2D-Enabled Federated Learning

    by Satyavrat Wagle, Anindya Bijoy Das, David J. Love, Christopher G. Brinton

    Augmenting federated learning (FL) with deviceto- device (D2D) communications can help improve convergence speed and reduce model bias through local information exchange. However, data privacy concerns, trust constraints between devices, and unreliable wireless channels each pose challenges in finding an effective yet resource efficient D2D graph structure. In this paper, we develop a decentralized reinforcement learning (RL) method for D2D graph discovery that promotes communication of impactful datapoints over reliable links for multiple learning paradigms, while following both data and device-specific trust constraints. An independent RL agent at each device trains a policy to predict the impact of incoming links in a decentralized manner without exposure of local data or significant communication overhead. For supervised settings, the D2D graph aims to improve device-specific label diversity without compromising system-level performance. For semi-supervised settings, we enable this by incorporating distributed label propagation. For unsupervised settings, we develop a variation-based diversity metric which estimates data diversity in terms of occupied latent space. Numerical experiments on five widely used datasets confirm that the data diversity improvements induced by our method increase convergence speed by up to 3× while reducing energy consumption by up to 5×. They also show that our method is resilient to stragglers and changes in the aggregation interval. Finally, we show that our method offers scalability benefits for larger system sizes without increases in relative overhead, and adaptability to various downstream FL architectures and to dynamic wireless environments.


  • Open Access

    Battling Misinformation: An Empirical Study on Adversarial Factuality in Open-Source Large Language Models

    by Shahnewaz Karim Sakib, Anindya Bijoy Das, Shibbir Ahmed

    Adversarial factuality refers to the deliberate insertion of misinformation into input prompts by an adversary, characterized by varying levels of expressed confidence. In this study, we systematically evaluate the performance of several open-source large language models (LLMs) when exposed to such adversarial inputs. Three tiers of adversarial confidence are considered: strongly confident, moderately confident, and limited confidence. Our analysis encompasses eight LLMs: LLaMA 3.1 (8B), Phi 3 (3.8B), Qwen 2.5 (7B), Deepseek-v2 (16B), Gemma2 (9B), Falcon (7B), Mistrallite (7B), and LLaVA (7B). Empirical results indicate that LLaMA 3.1 (8B) exhibits a robust capability in detecting adversarial inputs, whereas Falcon (7B) shows comparatively lower performance. Notably, for the majority of the models, detection success improves as the adversary's confidence decreases; however, this trend is reversed for LLaMA 3.1 (8B) and Phi 3 (3.8B), where a reduction in adversarial confidence corresponds with diminished detection performance. Further analysis of the queries that elicited the highest and lowest rates of successful attacks reveals that adversarial attacks are more effective when targeting less commonly referenced or obscure information.