The rapid growth of Decentraland has crafted a new frontier for digital asset investment. However, navigating and understanding this platform pose a significant challenge, especially for nonexpert users. The key obstacles include understanding market trends and price forecasting of the virtual assets, interpreting social media sentiment, and analyzing the impact of various economic factors. Furthermore, there is a lack of comprehensible, user-friendly tools to guide both the novice and experienced users in this complex, fast-evolving markets. To address this research gap, this article introducesDecentraBot, a domain specific conversational AI framework designed specifically for guiding users in making informed decisions on possible land investment and trading opportunities within Decentraland. By leveraging a systematic approach, which integrates fine-tuning methods, predictive and forecasting models, real-time data utilization, and novel prompt engineering techniques; we enable the open-domain large language models (LLMs) to transform into a domain-specific assistant capable of handling conversational, explanatory, and analytical queries effectively. To support users lacking prompting expertise, we introduce What-Next prompting; an efficient strategy that aims to minimize interaction barriers, making AI-based conversations more accessible and inclusive. Extensive experimental evaluation of DecentraBot over our proposed comprehensive fine-tune datasets demonstrates a significant performance improvement in terms of both quality and relevance of the responses to a broad spectrum of domain-specific queries. To the best of our knowledge, this is the first initiative of its kind in the realm of user-centric AI assistance within the metaverse ecosystem.

Soumitra Ghosh, Gopendra Vikram Singh, Shambhavi Shambhavi, Sabarna Choudhury, Asif Ekbal. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

The primary objective of this study is to enhance the recognition performance ofzero-shot children'sautomatic speech recognition (ASR) task. In such a setup, statistical models are trained on adults' speech while the test data is from child speakers. Due to the stark differences in the speech data from adult and children, poorer recognition performances are obtained. To mitigate the ill-effects of acoustic mismatch between the training and the test data, we have first resorted to data augmentation wherein the attributes of adults' training speech were modified in such way that those become acoustically similar to children's speech. Data augmentation helped us build a better baseline ASR system. In order to further enhance the recognition performance, we have studied the impact of concatenating glottal activity parameters with the Mel-frequency cepstral coefficients (MFCC). The explored glottal activity parameters are strength of excitation (SoE), normalized auto-correlation peak strength (NAPS), higher-order statistics (HOS), jitter and shimmer. However, shimmer was found to be unsuitable for thezero-shot children'sASR task. Therefore, only SoE, NAPS, HOS and jitter parameters were included into the training and this, in turn, yielded significantly improved recognition performance. A relative reduction in word error rate by$24.5\%$over the baseline was obtained on including glottal activity parameters. Similarly, the relative reduction in character error rate was noted to be$30\%$over the baseline. Finally, in order to further build confidence in the idea of modeling glottal activity parameters, we have extended the study tolow-resource children'sASR task employing end-to-end architecture. Even in this case, significant reductions in error rates were noted.