In machine translation (MT), one of the challenging tasks is to translate the proper nouns and technical terms from the source language to the target language while preserving the phonetic equivalent of original term. Machine transliteration, an essential part of MT systems, plays a vital role in handling proper nouns and technical terms. In this paper, a hybrid attention-based encoder–decoder machine transliteration system is proposed for the low-resource English to the Assamese language. In this work, the proposed machine transliteration system is integrated with the previously published hybrid attention-based encoder–decoder neural MT model to improve the translation quality of English to the Assamese language. The proposed integrated MT system demonstrated good results across various performance metrics such as BLEU, sacreBLEU, METEOR, chrF, RIBES, and TER for English to Assamese translation. Additionally, human evaluation was also conducted to assess translation quality. The proposed integrated MT system was compared with two existing systems: the Bing translation service model and the Samanantar Indic translation model.

A network of linked consumer electronics devices that use Social IoT principles to create socially-like connections is known as the Consumer Electronics enabled Social Internet of Things (CE-SIoT). CE-SIoT provides smooth information exchange, customized user experiences, and astute decision-making across linked devices. Social multimedia contains sensitive data and are posted and or exchanged for communication of object-to-object or object-to-human over the SIoT systems. However, SIoT systems continue to have challenges with social multimedia authentication, data security, integrity, and nonrepudiation. In this article, we propose ParticleStego to secure social multimedia and covert communication in SIoT. The ParticleStego employs Adaptive Particle Swarm Optimization (APSO) for selecting optimal pixel locations of the cover image to conceal a confidential message. The approach balances embedding capacity, structural-similarity index, and entropy through a multi-objective fitness function. APSO dynamically adjusts parameters for efficient convergence and ensures minimal perceptual distortion. Embedding is performed using bit-substitution approach, while the secret message is extracted from the least two bits of each optimal pixel of the stego image. This method ensures robust, adaptive, high-capacity, intelligent, and imperceptible steganography with an average 48.3733 dB PSNR and 0.9926 SSIM for hiding 155728 bits inside a 256×256×3 RGB image with lower entropy difference. Our robust ParticleStego is useful to secure sensitive information, such as personal data, copyright information, ownership details, covert communication, licensing terms, digital signatures or watermarks, etc., over advanced consumer electronics enabled SIoT systems. We have made the source code publicly available at https://github.com/somcse/ParticleStego.