International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | A Survey of Retrieval-Augmented Generation over Heterogeneous, Multi-Source Knowledgebases |
|---|---|
| ABSTRACT | Retrieval-augmented generation (RAG) improves a language model’s output by conditioning it on passages retrieved from an external corpus, but most of this work assumes a single, roughly homogeneous document collection. Practical knowledgebases violate that assumption: websites, PDFs, video transcripts, and code repositories are extracted, chunked, cited, and queried in different ways, and one query may need to draw on several at once. This survey organizes the techniques required for grounded question answering over such heterogeneous, multi-source knowledgebases: dense retrieval and RAG foundations, hybrid lexical-dense reranking, chunking strategies for long and structured documents, modality-specific retrieval for web, video, and code, multi-document fusion, and confidence estimation with abstention. Rather than treating each technique in isolation, we examine how each is affected once a corpus is assembled from structurally different sources at query time. One system, SAGE, which ingests and answers questions over crawled websites, PDFs, video transcripts, and GitHub repositories through a shared pipeline, appears as a brief case study showing how the surveyed ideas instantiate in a working implementation. We close with open problems in cross-modal fusion, confidence calibration, incremental re-indexing, and evaluation that remain unresolved across the literature. |
| AUTHOR | SRUJAN ZANJAL, RIMANSHU SONULE, PRATIK BHOYAR, TANISHA SHINGNE, PROF. AYAZ KHAN B.Tech Student, Department of Computer Science and Engineering, St. Vincent Pallotti College of Engineering and Technology, Nagpur, Maharashtra, India Assistant Professor, Department of Computer Science and Engineering, St. Vincent Pallotti College of Engineering and Technology, Nagpur, Maharashtra, India |
| VOLUME | 188 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1409017 |
| pdf/17_A Survey of Retrieval-Augmented Generation over Heterogeneous, Multi-Source Knowledgebases.pdf | |
| KEYWORDS | |
| References |
1. P. Lewis et al., “Retrieval-augmented generation for knowledge-intensive NLP tasks,” in Advances in Neural Information Processing Systems (NeurIPS), 2020. arXiv:2005.11401. 2. Y. Gao et al., “Retrieval-augmented generation for large language models: A survey,” 2023. arXiv:2312.10997. 3. Z. Ji et al., “Survey of hallucination in natural language generation,” ACM Computing Surveys, vol. 55, no. 12, art. 248, 2023. 4. P.-S. Huang et al., “Learning deep structured semantic models for web search using clickthrough data,” in Proc. 22nd ACM Int. Conf. on Information and Knowledge Management (CIKM), 2013, pp. 2333–2338. 5. V. Karpukhin et al., “Dense passage retrieval for open-domain question answering,” in Proc. EMNLP, 2020, pp. 6769–6781. arXiv:2004.04906. 6. G. Izacard and E. Grave, “Leveraging passage retrieval with generative models for open domain question answering,” in Proc. EACL, 2021. arXiv:2007.01282. 7. N. Reimers and I. Gurevych, “Sentence-BERT: Sentence embeddings using Siamese BERT-networks,” in Proc. EMNLP-IJCNLP, 2019. arXiv:1908.10084. 8. W. Wang et al., “MiniLM: Deep self-attention distillation for task-agnostic compression of pre-trained transformers,” in NeurIPS, 2020. arXiv:2002.10957. 9. R. Nogueira and K. Cho, “Passage re-ranking with BERT,” 2019. arXiv:1901.04085. 10. O. Khattab and M. Zaharia, “ColBERT: Efficient and effective passage search via contextualized late interaction over BERT,” in Proc. 43rd Int. ACM SIGIR Conf., 2020. arXiv:2004.12832. 11. L. Gao, Z. Dai, and J. Callan, “Complement lexical retrieval model with semantic residual embeddings,” 2020. arXiv:2004.13969. 12. S. Robertson and H. Zaragoza, “The probabilistic relevance framework: BM25 and beyond,” Foundations and Trends in Information Retrieval, vol. 3, no. 4, pp. 333–389, 2009. 13. G. V. Cormack, C. L. A. Clarke, and S. Büttcher, “Reciprocal rank fusion outperforms Condorcet and individual rank learning methods,” in Proc. 32nd Int. ACM SIGIR Conf., 2009, pp. 758–759. 14. A. Jimeno-Yepes et al., “Financial report chunking for effective retrieval augmented generation,” 2024. arXiv:2402.05131. 15. M. Günther et al., “Late chunking: Contextual chunk embeddings using long-context embedding models,” 2024. arXiv:2409.04701. 16. C. Merola and J. Singh, “Reconstructing context: Evaluating advanced chunking strategies for retrieval-augmented generation,” in 2nd Workshop on Knowledge-Enhanced Information Retrieval, ECIR, 2025. arXiv:2504.19754. 17. P. Sarthi et al., “RAPTOR: Recursive abstractive processing for tree-organized retrieval,” in Proc. ICLR, 2024. arXiv:2401.18059. 18. J. Lei et al., “TVQA: Localized, compositional video question answering,” in Proc. EMNLP, 2018. arXiv:1809.01696. 19. C. Zhang et al., “A simple LLM framework for long-range video question-answering,” in Proc. EMNLP, 2024. arXiv:2312.17235. 20. H. Husain et al., “CodeSearchNet challenge: Evaluating the state of semantic code search,” 2019. arXiv:1909.09436. 21. F. Zhang et al., “RepoCoder: Repository-level code completion through iterative retrieval and generation,” in Proc. EMNLP, 2023. arXiv:2303.12570. 22. D. Wu et al., “Repoformer: Selective retrieval for repository-level code completion,” in Proc. ICML, 2024. arXiv:2403.10059. 23. D. S. Sachan et al., “End-to-end training of multi-document reader and retriever for open-domain question answering,” in NeurIPS, 2021. arXiv:2106.05346. 24. A. Kamath, R. Jia, and P. Liang, “Selective question answering under domain shift,” in Proc. ACL, 2020. arXiv:2006.09462. 25. S. Kadavath et al., “Language models (mostly) know what they know,” 2022. arXiv:2207.05221. 26. K. Shuster et al., “Retrieval augmentation reduces hallucination in conversation,” in Findings of EMNLP, 2021, pp. 3784–3803. arXiv:2104.07567. 27. A. Asai et al., “Self-RAG: Learning to retrieve, generate, and critique through self-reflection,” in Proc. ICLR, 2024. arXiv:2310.11511. 28. S.-Q. Yan et al., “Corrective retrieval augmented generation,” 2024. arXiv:2401.15884. 29. Google, “NotebookLM: How to try Google’s experimental AI-first notebook,” Google Blog, 2023. [Online]. Available: https://blog.google/technology/ai/notebooklm-google-ai/ 30. LangChain, Inc., “LangChain documentation.” [Online]. Available: https://docs.langchain.com/ 31. LlamaIndex, “LlamaIndex documentation.” [Online]. Available: https://developers.llamaindex.ai/python/framework/ 32. D. Edge et al., “From local to global: A graph RAG approach to query-focused summarization,” 2024. arXiv:2404.16130. |
