ABAdarsh Prasad Behera

Adarsh Prasad Behera

*This is an algorithmically generated author record based on publicly available information.

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


Adarsh Prasad Behera has a total of 28 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Simulation Driven Evaluation of Three-tier Hierarchical Inference Learning Framework

    by Adarsh Prasad Behera, Jaya Prakash Varma Champati, James Gross

    Distributed inference systems leverage hierarchical architectures spanning device, edge, and cloud to balance accuracy, latency, and resource cost. While prior work on Hierarchical Inference (HI) has primarily focused on two-tier designs, practical deployments often involve multiple models with varying capabilities. In this paper, we present a simulation-based study of a three-tier HI framework comprising two on-device models and a high-accuracy cloud model, enabling finer-grained inference decisions. We formulate the problem as an online threshold-based HI framework, where confidence scores from each model determine whether inputs are processed locally or escalated to the upper tier. An exponential weighted forecaster (EWF) strategy is used to adapt thresholds over time. We evaluate the proposed framework on CIFAR-10 and ImageNet-1K datasets under varying cost regimes. Results show that the threetier system consistently achieves near-oracle performance and outperforms all two-tier baselines by effectively handling moderate-difficulty samples at the intermediate tier. Furthermore, we observe that different feedback mechanisms yield similar average costs despite producing distinct inference acceptance patterns, indicating the presence of multiple cost-equivalent operating points. These findings demonstrate the practical advantages of three-tier HI and provide insights for designing efficient distributed ML systems.

  • Open Access

    A Cost-Aware Hierarchical Cascade for Anomaly Detection at the Edge in Connected Vehicles

    by Cheng-Hsun Chang, Adarsh Prasad Behera, Sophia Zhang Pettersson, James Gross

  • Exploring the Boundaries of On-Device Inference: When Tiny Falls Short, Go Hierarchical

    by Adarsh Prasad Behera, Paulius Daubaris, Iñaki Bravo, José Gallego, Roberto Morabito, Joerg Widmer, Jaya Prakash Champati

    On-device inference offers significant benefits in edge ML systems, such as improved energy efficiency, responsiveness, and privacy, compared to traditional centralized approaches. However, the resource constraints of embedded devices limit their use to simple inference tasks, creating a tradeoff between efficiency and capability. In this context, the hierarchical inference (HI) system has emerged as a promising solution that augments the capabilities of the local ML by offloading selected samples to an edge server/cloud for remote ML inference. Existing works, primarily based on simulations, demonstrate that HI improves accuracy. However, they fail to account for the latency and energy consumption in real-world deployments, nor do they consider three key heterogeneous components that characterize ML-enabled IoT systems, such as hardware, network connectivity, and models. To bridge this gap, this article systematically evaluates HI against standalone on-device inference by analyzing accuracy, latency, and energy tradeoffs across five devices and three image classification datasets. Our findings show that, for a given accuracy requirement, the HI approach we designed achieved up to 73% lower latency and up to 77% lower device energy consumption than an on-device inference system. Despite these gains, HI introduces a fixed energy and latency overhead from on-device inference for all samples. To address this, we propose a hybrid system called early exit with HI (EE-HI) and demonstrate that, compared to HI, EE-HI reduces the latency up to 59.7% and lowers the device’s energy consumption up to 60.4%. These findings demonstrate the potential of HI and EE-HI to enable more efficient ML in IoT systems.


  • Intensity Based Event Detection in Sensor Based IoT

    by Anubhav Shivhare, Adarsh Prasad Behera, Manish Kumar

    Finding an optimum trade-off between event detection and network lifetime is a major problem in the sensor-based Internet of Things framework. Further, reliable, effective, and accurate event detection is a perennial research problem explored in the domain of Sensor Based Internet of Things (SBIoT). Major research problems focusing on event detection depend upon models like Boolean and probabilistic sensing models. However, event detection is practically dependent upon the intensity and persistence of the event. The traditional non-intensity-based event sensing models fix a predefined sensing radius. Any occurrence outside the sensing radius is not considered an event, independent of its severity. The present work argues that the intensity and persistence of the event are also relevant parameters for event detection. This paper proposes two novel event intensity and persistence-based models for detecting different types of events and improving upon the quality of detection. The proposed ’Improved' model proves to be more efficient than the proposed ’Conventional' model. Further, the simulation results indicate the proposed algorithm's efficiency and effectiveness, and compare it with Non-intensity based models. Additionally, the results are compared in terms of detection accuracy, node activation, and network lifetime to show the efficiency and trade-offs of the proposed scheme.

  • CNN-Based Metrics for Performance Evaluation of Generative Adversarial Networks

    by Adarsh Prasad Behera, Satya Prakash, Siddhant Khanna, Shivangi Nigam, Shekhar Verma

    In this work, we propose two Convolutional Neural Network (CNN) based metrics, Classification Score (CS) and Distribution Score (DS), for performance evaluation of Generative Adversarial Networks (GANs). Though GAN-generated images can be evaluated through manual assessment of visual fidelity, it is prolonged, subjective, challenging, tiresome, and can be misleading. Existing quantitative methods are biased towards memory GAN and fail to detect over-fitting. CS and DS allow us to experimentally prove that training of GANs is actually guided by the data set, that it improves with every epoch and gets closer to following the distribution of the data set. Both methods are based on GAN-generated image classification by CNN. CS is the root mean square (RMS) value of three different classification techniques, Direct Classification (DC), Indirect Classification (IC), and Blind Classification (BC). It exhibits the degree to which GAN can learn the features and generate fake images similar to real data sets. DS shows the contrast between the mean distribution of GAN-generated data and the real data. It indicates the extent to which GANs can create synthetic images with similar distribution to real data sets. We evaluated CS and DS metrics for different variants of GANs and compared their performances with existing metrics. Results show that CS and DS can evaluate the different variants of GANs quantitatively and qualitatively while detecting over-fitting and mode collapse.


  • Open Access

    Improved Decision Module Selection for Hierarchical Inference in Resource-Constrained Edge Devices

    by Adarsh Prasad Behera, Roberto Morabito, Joerg Widmer, Jaya Prakash Champati

    The Hierarchical Inference (HI) paradigm has recently emerged as an effective method for balancing inference accuracy, data processing, transmission throughput, and offloading cost. This approach proves particularly efficient in scenarios involving resource-constrained edge devices like micro controller units (MCUs), tasked with executing tinyML inference. Notably, it outperforms strategies such as local inference execution, inference offloading, and split inference (i.e., inference execution distributed between two endpoints). Building upon the HI paradigm, this work explores different techniques aimed at further optimizing inference task execution. We propose three distinct HI approaches and evaluate their utility for image classification.


  • Open Access

    The Case for Hierarchical Deep Learning Inference at the Network Edge

    by Ghina Al-Atat, Andrea Fresa, Adarsh Prasad Behera, Vishnu Narayanan Moothedath, James Gross, Jaya Prakash Champati

    Resource-constrained Edge Devices (EDs), e.g., IoT sensors and microcontroller units, are expected to make intelligent decisions using Deep Learning (DL) inference at the edge of the network. Toward this end, developing tinyML models is an area of active research - DL models with reduced computation and memory storage requirements - that can be embedded on these devices. However, tinyML models have lower inference accuracy. On a different front, DNN partitioning and inference offloading techniques were studied for distributed DL inference between EDs and Edge Servers (ESs). In this paper, we explore Hierarchical Inference (HI), a novel approach proposed in [19] for performing distributed DL inference at the edge. Under HI, for each data sample, an ED first uses a local algorithm (e.g., a tinyML model) for inference. Depending on the application, if the inference provided by the local algorithm is incorrect or further assistance is required from large DL models on edge or cloud, only then the ED offloads the data sample. At the outset, HI seems infeasible as the ED, in general, cannot know if the local inference is sufficient or not. Nevertheless, we present the feasibility of implementing HI for image classification applications. We demonstrate its benefits using quantitative analysis and show that HI provides a better trade-off between offloading cost, throughput, and inference accuracy compared to alternate approaches.


  • Strike off removal in Indic scripts with transfer learning

    by Shivangi Nigam, Adarsh Prasad Behera, Manas Gogoi, Shekhar Verma, P. Nagabhushan


  • Curvature Regularized Similarity Based Ricci Flow Embedding

    by Adarsh Prasad Behera, Puja Rani, Shekhar Verma, Manish Kumar

    This work proposes a novel curvature regularization method to regularize the individual sectional curvatures in Similarity-Based Ricci Flow Embedding (SBRFE) to reduce the non-Euclidean artefacts and compute the Euclidean embedding of similarity-based data sets. In pattern recognition, pairwise similarity or dissimilarity data have been used as alternative to the more conventional feature vector representation. While similarity representations are rarely Euclidean, most methods involving statistical analysis or learning of such data require them to be Euclidean. In SBRFE, each similarity is considered an individual entity, and the respective sectional curvature is calculated and updated separately. This compromises the smoothness of the manifold and causes numerical instability. To overcome this problem, we used regularized Newton's method (RNM) to regularize the sectional curvatures of each patch obtained from the initial curvature computation. It ensures numerical stability in the embedding and smoothens the manifold. It also preserves both the local and global geometry of the original data sets. Results show that proposed curvature regularized similarity-based Ricci Flow Embedding (CRRFE) is able to estimate the Euclidean embedding of similarity data sets with much lower computation cost and time complexity than the existing regularization method. Comparison results show that our proposed methodology outperforms other existing embedding methods in most data sets with a lower classification error rate.