STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Advanced Database Management & Systems Original Research

VeriSci—AI-Based Multi-Modal Research Assistant

  1. Pradnya Bormane Department of Computer Engineering, AISSMS Institute of Information Technology
  2. Ayush Sangole Department of Computer Engineering, AISSMS Institute of Information Technology
  3. Arya Chunne Department of Computer Engineering, AISSMS Institute of Information Technology
  4. Apeksha Bhosale Department of Computer Engineering, AISSMS Institute of Information Technology

Abstract

The exponential growth of scientific literature is a major bottleneck for academic researchers who want to efficiently discover, assess, and synthesize relevant scholarly knowledge. Traditional methods of literature review are heavily reliant on manual keyword searching, human screening, and subjective data extraction, making them time-consuming, susceptible to cognitive bias, and less effective as the tidal wave of information continues to grow. To overcome these limitations, this study introduces VeriSci, an innovative artificial intelligence–based multi-modal research assistant engineered to optimize the efficiency dramatically, accuracy, and overall reliability of scholarly research workflows. The proposed framework seamlessly orchestrates Multi-Agent Systems (MAS) alongside advanced Retrieval-Augmented Generation (RAG) paradigms to completely automate systematic literature discovery, comparative analysis, cross-paper verification, and intelligent document summarization. VeriSci programmatically aggregates data from eight prominent academic repositories, applies a fuzzy-logic deduplication pipeline, builds interactive Neo4j knowledge graphs to expose citation dynamics, and leverages context-grounded large language models (LLMs) to synthesize precise, reliable answers. To minimize the persistent risk of LLM hallucinations, a robust multi-stage verification layer incorporates automated provenance tracking, citation network validation, and empirical claim checking. Experimental evaluation shows that VeriSci achieves a remarkable overall response accuracy of 87%, precision and recall of 85% and 89%, respectively, while maintaining a fast average response latency of 1.6 seconds. Importantly, as a direct comparison, VeriSci improves context-aware response relevance by 22% over baseline LLM implementations without retrieval augmentation. The empirical results demonstrate that this framework effectively addresses informational fragmentation, mitigates research blind spots, and offers a highly scalable infrastructure for next-generation automated scientific workflows.

Keywords

References (17)

  1. Gao Y, Xiong Y, Gao X, Jia K, Pan J, Bi Y, et al. Retrieval-augmented generation for large language models: a survey [Preprint]. arXiv:2312.10997; 2023. doi:10.48550/arXiv.2312.10997.
  2. Arslan M, Ghanem H, Munawar S, Cruz C. A survey on RAG with LLMs. Procedia Comput Sci. 2024;246:3781-3790. doi:10.1016/j.procs.2024.09.178.
  3. Cornelio C, Ito T, Cory-Wright R, Dash S, Horesh L. The need for verification in AI-driven scientific discovery [Preprint]. arXiv:2509.01398; 2025. doi:10.48550/arXiv.2509.01398.
  4. Myllyaho LS, Raatikainen M, Männistö T, Mikkonen T, Nurminen JK. Systematic literature review of validation methods for AI systems. J Syst Softw. 2021;181:111050. doi:10.1016/j.jss.2021.111050.
  5. Gridach M, Nanavati J, Abidine KZ, Mendes L, Mack C. Agentic AI for scientific discovery: a survey of progress, challenges, and future directions [Preprint]. arXiv:2503.08979; 2025.
  6. Su H, Chen R, Tang S, Yin Z, Zheng X, Li J, et al. Many heads are better than one: improved scientific idea generation by a LLM-based multi-agent system. In: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers); 2025 Jul; Vienna, Austria. Stroudsburg (PA): Association for Computational Linguistics; 2025. p. 28201-28240. doi:10.18653/v1/2025.acl-long.1368.
  7. Hughes L, Dwivedi YK, Malik T, Shawosh M, Albashrawi MA, Jeon I, et al. AI agents and agentic systems: a multi-expert analysis. J Comput Inf Syst. 2025;65(4):489-517. doi:10.1080/08874417.2025.2483832.
  8. Besrour I, He J, Schreieder T, Färber M. SQuAI: scientific question-answering with multi-agent retrieval-augmented generation. In: Proceedings of the 34th ACM International Conference on Information and Knowledge Management; 2025. p. 6603-6608. doi:10.1145/3746252.3761471.
  9. Amugongo LM, Mascheroni P, Brooks S, Doering S, Seidel J. Retrieval augmented generation for large language models in healthcare: a systematic review. PLOS Digit Health. 2025;4(6). doi:10.1371/journal.pdig.0000877. PMID:40498738.
  10. Papageorgiou G, Sarlis V, Maragoudakis M, Tjortjis C. Hybrid multi-agent GraphRAG for e-government: towards a trustworthy AI assistant. Appl Sci (Basel). 2025;15(11):6315. doi:10.3390/app15116315.
  11. Kharipova R, Khaydarov I, Akramova S, Lutfullaeva D, Saidov S, Erkinov A, et al. The role of artificial intelligence technologies in evaluating the veracity of scientific research. J Internet Serv Inf Secur. 2024;14(4):554-568. doi:10.58346/JISIS.2024.I4.035.
  12. Chander B, John C, Warrier L, Gopalakrishnan K. Toward trustworthy artificial intelligence (TAI) in the context of explainability and robustness. ACM Comput Surv. 2025;57(6):1-49. doi:10.1145/3675392.
  13. Du Y, Leibo JZ, Islam U, Willis R, Sunehag P. A review of cooperation in multi-agent learning [Preprint]. arXiv:2312.05162; 2023. doi:10.48550/arXiv.2312.05162.
  14. Litmaps: your literature review assistant [Online]. Litmaps; 2026. Available from: https://www.litmaps.com/
  15. ResearchPal [Online]. ResearchPal; 2026. Available from: https://researchpal.co/
  16. Paperpal for researchers: AI for research paper writing [Online]. Paperpal. Available from: https://paperpal.com/paperpal-for-researchers
  17. Chowdhury HMA, Fuller M, Oluwadare O. Robin: an advanced tool for comparative loop caller result analysis leveraging large language models. NAR Genom Bioinform. 2026;8(1). doi:10.1093/nargab/lqag009. PMID:41625146.
Support