Source author record

Mohan Kumar

Mohan Kumar appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

astro-ph.GA astro-ph.IM Cryptography and Security Networking and Internet Architecture

Catalog footprint

What is connected

3works

4topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Explainable Galaxy Interaction Prediction with Hybrid Attention Mechanisms

Galaxy interaction classification remains challenging due to complex morphological patterns and the limited interpretability of deep learning models. We propose an attentive neural ensemble that combines AG-XCaps, H-SNN, and ResNet-GRU architectures, trained on the Galaxy Zoo DESI dataset and enhanced with LIME to enable explainable predictions. The model achieves Precision = 0.95, Recall = 1.00, F1 = 0.97, and Accuracy = 96%, outperforming a Random Forest baseline by significantly reducing false positives (23 vs. 70). This lightweight (0.45 MB) and scalable framework provides an interpretable and efficient solution for large-scale surveys such as Euclid and LSST, advancing data-driven studies of galaxy evolution.

preprint2022arXiv

Service Composition in Opportunistic Networks: A Load and Mobility Aware Solution

Pervasive networks formed by users' mobile devices have the potential to exploit a rich set of distributed service components that can be composed to provide each user with a multitude of application level services. However, in many challenging scenarios, opportunistic networking techniques are required to enable communication as devices suffer from intermittent connectivity, disconnections and partitions. This poses novel challenges to service composition techniques. While several works have discussed middleware and architectures for service composition in well-connected wired networks and in stable MANET environments, the underlying mechanism for selecting and forwarding service requests in the significantly challenging networking environment of opportunistic networks has not been entirely addressed. The problem comprises three stages: i) selecting an appropriate service sequence set out of available services to obtain the required application level service; ii) routing results of a previous stage in the composition to the next one through a multi-hop opportunistic path; and iii) routing final service outcomes back to the requester. The proposed algorithm derives efficiency and effectiveness by taking into account the estimated load at service providers and expected time to opportunistically route information between devices. Based on this information the algorithm estimates the best composition to obtain a required service. It is shown that using only local knowledge collected in a distributed manner, performance close to a real-time centralized system can be achieved. Applicability and performance guarantee of the service composition algorithm in a range of mobility characteristics are established through extensive simulations on real/synthetic traces.

preprint2014arXiv

FilterPlus: A real-time content filtering extension for Google Chrome

Content filtering in web browsers is a tedious process for most of the people, because of several reasons. By blocking JavaScript, Cookies and Popups, end users can ensure maximum protection from browser based attacks and vulnerabilities. In order to accomplish this, we build an extension for Google Chrome which allows users to have easy control over what they wish to recievce from a web page. We also build this extension in such a way that it remembers the choice of options made by the user for every URLs, thereby letting users create rules for websites they visit.