Deepak Khazanchi

Fairness in Machine Learning (ML) development is a critical but subjective concern, shaped by biases in data, practices, and assumptions. Although algorithmic fairness has gained attention, empirical insights into how ML practitioners and domain experts perceive fairness across development remain limited. Using a sociotechnical lens, this study examines how fairness perceptions evolve during ML development, shifting attention from outcomes to processes to support human-centered evaluation. This research introduces and validates the TAR framework, which operationalizes fairness perceptions across three dimensions: transparency, accountability, and representativeness. Grounded in theory, three validated survey scales measure these perceptions systematically. The findings offer a nuanced understanding of fairness, enabling more inclusive ML design. By incorporating practitioner and subject matter expert perspectives, the TAR framework provides a human-centered tool for evaluating fairness, guiding ethical design decisions, and fostering equitable ML practices.

Encrypted inference under homomorphic encryption (HE) imposes stringent constraints on computation, making the direct deployment of deep neural networks infeasible on resource-constrained edge devices. Existing approaches, such as FHEEDGE and UniHENN, primarily focus on improving the efficiency of encrypted computation through algorithmic and systems-level optimizations. In contrast, we adopt a complementary perspective by redesigning neural architectures to align with the inherent cost model of HE operations. In this work, we present a framework that reformulates HE-aware model design as a constrained neural architecture search problem, where the objective is to identify architectures that are both cryptographically feasible and computationally efficient while preserving task performance. Starting from a pretrained network, we perform a structured, iterative transformation process that includes HE-compatible operator substitutions, depth–width tradeoffs, and rotation-efficient design strategies such as successive average pooling. These transformations are guided by constraints on ciphertext slots, multiplicative depth, and rotation complexity, and are evaluated using short-horizon training to efficiently explore the architecture space. We provide a theoretical analysis showing that the transformed network approximates the original model with bounded error under layer-wise Lipschitz propagation, ensuring controlled performance degradation. As a case study, we apply the proposed framework to encrypted crack detection for structural health monitoring. The optimized model achieves an execution time of 86.5 seconds for single-threaded CPU inference in the encrypted domain, compared with an estimated 3.5 hours required by the original encrypted convolutional pipeline. Furthermore, deployment on a Raspberry Pi demonstrates end-to-end encrypted inference within approximately 7.5 minutes. These results highlight that architecture-level optimization, when framed as a constrained search problem, complements existing algorithmic acceleration approaches and provides a practical pathway for enabling efficient and privacy-preserving inference on edge devices across a broad class of sensitive AI workloads.

Perceived fairness in machine learning (ML) systems represents a sociotechnical phenomenon, shaped by both organizational dynamics and technical practices. Building on prior research that formalizes fairness as experienced within ML development, this study investigates perceptual gaps through a large-scale survey of around 123 ML practitioners. Shifting from outcome-focused fairness research, this work centers on how stakeholders interpret and experience fairness throughout the ML development lifecycle. Specifically, it examines how practitioner demographics influence fairness perceptions within real-world ML projects. Drawing from sociotechnical literature, this study develops and validates a survey instrument to capture these perceptions throughout the ML development lifecycle. Analysis reveals that demographic factors, including experience, gender, and education, significantly affect how fairness is perceived. The findings underscore the importance of inclusive design practices and point to actionable pathways for policy reform, targeted training, and organizational strategies aimed at promoting equity, transparency, and accountability in ML systems.

Bridge health monitoring is critical for infrastructure safety, especially with the growing deployment of IoT sensors. This work addresses the challenge of securely storing large volumes of sensor data and extracting actionable insights for timely damage detection. We propose BIONIB, a novel framework that combines an unsupervised machine learning approach called the Novelty Index (NI) with a scalable blockchain platform (EOSIO) for secure, real-time monitoring of bridges. BIONIB leverages EOSIO’s smart contracts for efficient, programmable, and secure data management across distributed sensor nodes. Experiments on real-world bridge sensor data under varying loads, climatic conditions, and health states demonstrate BIONIB’s practical effectiveness. Key findings include CPU utilization below 40% across scenarios, a twofold increase in storage efficiency, and acceptable latency degradation, which is not critical in this domain. Our comparative analysis suggests that BIONIB fills a unique niche by coupling NI-based detection with a decentralized architecture, offering real-time alerts and transparent, verifiable records across sensor nodes.

Structural Health Monitoring (SHM) of bridges is essential for maintaining safe and reliable transportation networks. This study integrates the EOSIO blockchain with Unmanned Aerial Vehicles (UAVs) and Deep Reinforcement Learning (DRL) using the Proximal Policy Optimization (PPO) algorithm to optimize UAV-based bridge inspections and crack detection. The blockchain ensures secure and tamper-proof crack data storage, enhancing transparency and data integrity. Experimental results show a 28.57% latency improvement in low-traffic scenarios, demonstrating the method’s potential to improve SHM efficiency and support infrastructure modernization efforts.

In machine learning (ML) applications, unfairness is triggered due to bias in the data, the data curation process, erroneous assumptions, and implicit bias rendered during the development process. It is also well-accepted by researchers that fairness in ML application development is highly subjective, with a lack of clarity of what it means from an ML development and implementation perspective. Thus, in this research, we investigate and formalize the notion of the perceived fairness of ML development from a sociotechnical lens. Our goal in this research is to understand the characteristics of perceived fairness in ML applications. We address this research goal using a three-pronged strategy: 1) conducting virtual focus groups with ML developers, 2) reviewing existing literature on fairness in ML, and 3) incorporating aspects of justice theory relating to procedural and distributive justice. Based on our theoretical exposition, we propose operational attributes of perceived fairness to be transparency, accountability, and representativeness. These are described in terms of multiple concepts that comprise each dimension of perceived fairness. We use this operationalization to empirically validate the notion of perceived fairness of machine learning (ML) applications from both the ML practioners and users perspectives. The multidimensional framework for perceived fairness offers a comprehensive understanding of perceived fairness, which can guide the creation of fair ML systems with positive implications for society and businesses.

This work focuses on using advanced techniques for Structural Health Monitoring (SHM) for bridges with Traffic (SHM-Traffic). We propose an approach using Deep Reinforcement Learning (DRL)-based control for Unmanned Aerial Vehicle (UAV). Our approach conducts a concrete bridge deck survey while traffic is ongoing and detects cracks. The UAV performs the crack detection, and the location of cracks is initially unknown. We use two edge detection techniques. First, we use canny edge detection for crack detection. We also use a Convolutional Neural Network (CNN) for crack detection and compare it with canny edge detection. Transfer learning is applied using CNN with pre-trained weights obtained from a crack image dataset. This enables the model to adapt and improve its performance in identifying and localizing cracks. Proximal Policy Optimization (PPO) is applied for UAV control and bridge surveys. The experimentation across various scenarios is performed to evaluate the performance of the proposed methodology. Key metrics such as task completion time and reward convergence are observed to gauge the effectiveness of the approach. We observe that the Canny edge detector offers up to 40% lower task completion time, while the CNN excels in up to 12% better damage detection and 1.8 times better rewards.

Fifteen years ago, Davis et al. (2009) defined and described the metaverse’s capabilities and presented a foundation for research in this environment. Since then, particularly in the past few years, researchers have paid considerable attention to metaverses in both research and practice. This growth in attention suggests it is time to assess what we learned about metaverses and explore future opportunities. Therefore, in this paper, we review what we learned over the past 15 years related to metaverse research and present important possibilities and opportunities for future research. We summarize foundational research findings in relation to the metaverse, people/avatars, metaverse technology capabilities, behaviors, and virtual world outcomes. Additionally, we present research opportunities in metaverse education, development, lifestyle, work and job design, and gaming—all areas where metaverse technology innovation and application have received attention since Davis et al. (2009). The research opportunities that we outline illustrate how the metaverse continues to present rich possibilities and promise for the future.