Lexicon or dictionary generation across domains has the potential for societal impact, as it can potentially enhance information accessibility for a diverse user base while preserving language identity. Prior work in the field primarily focuses on bilingual lexical induction, which deals with word alignments using mapping-based or corpora-based approaches. However, these approaches do not cater to domain-specific lexicon generation that consists of domain-specific terminology. This task becomes particularly important in specialized medical, engineering, and other technical domains, owing to the highly infrequent usage of the terms and scarcity of data involving domain-specific terms especially for low-resource languages. We propose a new model to generate dictionary words for 6 Indian languages in the multi-domain setting. Our model consists of domain-specific and domain-generic layers that encode information, and these layers are invoked via a learnable routing technique. We also release a new benchmark dataset consisting of >75K translation pairs across 6 Indian languages spanning 8 diverse domains. We conduct both zero-shot and few-shot experiments across multiple domains to show the efficacy of our proposed model in generalizing to unseen domains and unseen languages. Additionally, we also perform a human post-hoc evaluation on unseen languages. The source code and dataset is present at https://github.com/Atulkmrsingh/lexgen.

ABSTRACT Ensuring data confidentiality and integrity in dynamic cloud storage environments is a growing challenge, particularly in the face of key exposure threats. Traditional key management schemes, which rely on periodic updates, introduce significant vulnerabilities due to long windows of exposure between key rotations. Also, it often requires high computational overhead from re‐encrypting entire datasets during key updates and frequently depends on third‐party auditors for integrity verification, which can compromise privacy. However, a major research gap remains in developing a scalable, efficient, and auditor‐free key management protocol that can adapt in real time to evolving cloud access patterns. In this paper, we propose a novel Dynamic Event‐Driven Key Regeneration System that leverages Elliptic Curve Diffie‐Hellman (ECDH) for secure key exchange and Advanced Encryption Standard—Galois/Counter Mode (AES‐GCM) for combined encryption and integrity verification. Unlike conventional time‐ or session‐based strategies, the proposed design uses statistically adaptive thresholding derived from real‐time file access patterns to enable on‐demand key regeneration and selective re‐encryption, drastically reducing computational overhead. By re‐encrypting only affected files, the system is optimized for large‐scale, multi‐tenant cloud environments. Furthermore, the proposed approach eliminates the need for external auditors, as integrity verification is performed internally via cryptographic mechanisms, ensuring both privacy and security. Experimental results show that the proposed system achieves an average key generation time of 2.3 ms, encryption latency of just 0.21 s for 100 MB files, and key regeneration times as low as 0.0012–0.0350 s, outperforming existing approaches by up to 80% in computational efficiency. The system scales efficiently in multi‐tenant environments, maintaining low overhead with up to 100 users and providing near‐linear performance even with 1000 concurrent encryption operations. These results demonstrate that the proposed adaptive, event‐driven system offers enhanced protection against key exposure while maintaining low overhead, making it a viable and secure solution for modern cloud infrastructures.

ABSTRACT Cloud computing offers revolutionary data storage and management features through advancements in elastic scaling and affordable usage rates. However, the widespread adoption of cloud computing solutions has created significant security challenges, including attacks from internal staff, unauthorized system access, confidentiality breaches, and difficulties in meeting sector‐specific compliance requirements. This research technically investigates cloud storage security through a multilevel evaluation process, assessing vulnerabilities in physical infrastructure, application systems, and user interfaces, as well as implementing the principles of confidentiality, integrity, and availability (CIA). The study hypothesizes that traditional security solutions based solely on encryption are inadequate for modern cloud security requirements. Therefore, a comprehensive security system is needed—one that incorporates access regulations with cryptographic protocols and continuous intrusion monitoring to fully protect data assets. To validate this hypothesis, a systematic review was conducted in which authors evaluated existing studies on cloud system vulnerabilities, attack methods, and defensive strategies. The study assesses key security risks within the Infrastructure as a Service (IaaS), Software as a Service (SaaS), and Platform as a Service (PaaS) cloud service models. The analysis evaluates cloud security mechanisms by examining encryption algorithms and access control models to determine their performance in cloud computing environments. Emerging technologies such as software‐defined networking (SDN) and network function virtualization (NFV) are examined to determine their effectiveness in improving cloud security and their associated security challenges. Findings show that, while encryption and existing security tools remain essential, they are insufficient to provide complete protection against modern threats such as API vulnerabilities, unauthorized access, and SDN controller attacks. This research concludes that protection strategies must incorporate multiple layers of security, integrating a continuous monitoring environment along with adaptive user authentication and policy compliance management to effectively address current threat vectors.

We release Saamayik, a dataset of around 53,000 parallel English-Sanskrit sentences, written in contemporary prose. Sanskrit is a classical language still in sustenance and has a rich documented heritage. However, due to the limited availability of digitized content, it still remains a low-resource language. Existing Sanskrit corpora, whether monolingual or bilingual, have predominantly focused on poetry and offer limited coverage of contemporary written materials. Saamayik is curated from a diverse range of domains, including language instruction material, textual teaching pedagogy, and online tutorials, among others. It stands out as a unique resource that specifically caters to the contemporary usage of Sanskrit, with a primary emphasis on prose writing. Translation models trained on our dataset demonstrate statistically significant improvements when translating out-of-domain contemporary corpora, outperforming models trained on older classical-era poetry datasets. Finally, we also release benchmark models by adapting four multilingual pre-trained models, three of them have not been previously exposed to Sanskrit for translating between English and Sanskrit while one of them is multi-lingual pre-trained translation model including English and Sanskrit. The dataset and source code can be found at https://github.com/ayushbits/saamayik.