Pith. sign in

REVIEW 4 major objections 4 minor 130 references

This survey claims that the state of the art in conversational question answering can be organized into three interacting components—history selection, question understanding, and answer prediction—plus five machine-learning techniques, fiv

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

A review that categorizes ConvQA components, techniques, models, and datasets, with no new experimental result.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection A usable orientation survey of ConvQA with a coherent taxonomy, but the LLM comparison table contains concrete factual errors that need fixing before the paper can serve as a reference. the 4 major comments →

arxiv 2509.05716 v1 pith:I3KRGWX4 submitted 2025-09-06 cs.CL cs.AI

A Survey of the State-of-the-Art in Conversational Question Answering Systems

Manoj Madushanka Perera , Adnan Mahmood , Kasun Eranda Wijethilake , Fahmida Islam , Maryam Tahermazandarani , Quan Z. Sheng This is my paper
classification cs.CL cs.AI
keywords Conversational AIConversational Question AnsweringLarge Language ModelsConversational History ManagementNatural Language ProcessingMachine LearningDatasets
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This survey tries to establish a structured map of conversational question answering (ConvQA) as of 2025. It argues that every ConvQA system can be described by three interacting components—history selection, question understanding, and answer prediction—and that recent progress is driven by five machine-learning techniques applied to large language models and evaluated on a small set of benchmarks. If the map is accurate, it gives researchers a shared vocabulary for comparing systems and for locating unsolved problems, especially cross-domain adaptability, multimodal input, personalization, real-time adaptation, dynamic history management, and handling ambiguity and uncertainty. The survey's value therefore rests on whether its selection and classification of the literature are representative.

Core claim

The paper's central claim is that the current state of the art in ConvQA can be usefully organized as a taxonomy: four strategies for history selection (K-turn, immediate turn, entire history, dynamic), five for question understanding (rewriting, reformulation, named entity recognition, semantic parsing, attention), and four for answer prediction (retrieval-based, generative, retrieval-augmented generation, knowledge graph-based). On top of that architecture, the survey sorts recent work by machine-learning technique—reinforcement learning, knowledge distillation, contrastive learning, active learning, and transfer learning—and profiles the LLMs and datasets that anchor the field. It does no

What carries the argument

The carrying structure is the survey's three-component taxonomy—history selection, question understanding, answer prediction—shown as the system architecture and instantiated in Table 1. The taxonomy is what lets the paper place roughly thirty recent systems into comparable slots, and the same tripartite frame organizes the later sections on machine learning, LLMs, datasets, and open problems.

Load-bearing premise

The survey's claim to cover the state of the art depends on its selection of papers being representative and on its taxonomy categories matching what the cited systems actually do.

What would settle it

Re-run the paper's stated literature-selection protocol—keywords, venues, citation counts—for the same period and check whether Tables 1 and 2 capture every qualifying work with accurate category assignments. Finding a substantial cluster of qualifying ConvQA work that fits no category, or that is systematically misclassified, would settle the comprehensiveness claim against the survey.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • A new ConvQA system can be located in the taxonomy by stating its history-selection, question-understanding, and answer-prediction strategies, which makes comparisons across papers more direct.
  • No single history-selection strategy dominates: static K-turn selection works well on CoQA and QuAC, while dynamic and reinforcement-learning-based selection is the current frontier for long or noisy conversations.
  • The LLM comparison table (parameter count, context length, training tokens, open-source status) gives practitioners a quick filter for choosing a model for ConvQA work.
  • The six open research directions act as a checklist for future work, with dynamic conversational history management and ambiguity handling identified as the main barriers to real-world deployment.
  • Because datasets differ in whether they require question rewriting, topic switching, or open-domain retrieval, benchmark choice changes what ability a ConvQA system must demonstrate.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A testable extension the survey leaves implicit: hold the dataset fixed and sweep the history-selection strategy across LLMs to quantify which component matters most as models grow.
  • The taxonomy's boundary between question rewriting, reformulation, and structured representations is likely to blur in practice, since several cited systems combine them; a method-level re-coding would test the taxonomy's stability.
  • The paper's multimodal direction could be operationalized by extending a topic-switching dataset like TopiOCQA with image- or audio-bearing turns, creating a benchmark that does not exist in Table 4.
  • Given how fast the LLM landscape changes, the model tables will age quickly even if the three-component taxonomy persists; the survey's durable contribution is the architecture, not the model list.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. This manuscript is a survey of conversational question answering (ConvQA). It proposes a three-part taxonomy—history selection (k-turn, immediate-turn, entire-history, dynamic), question understanding (rewriting, reformulation, NER, semantic parsing, attention), and answer prediction (retrieval-based, generative, RAG, knowledge-graph)—and uses this taxonomy to organize a review of machine learning techniques (RL, KD, contrastive learning, active learning, transfer learning), large language models (RoBERTa, GPT-4, Gemini 2.0 Flash, Mistral 7B, LLaMA 3), and datasets (CoQA, QuAC, SQuAD 2.0, CANARD, QReCC, TopiOCQA). The paper claims to provide a comprehensive, reference-quality map of the field and identifies open research directions.

Significance. If accurate, this survey would be a useful structured entry point to ConvQA, with a coherent taxonomy, helpful figures, and recent coverage through 2025. The paper is an expository contribution, not a derivation, so its value rests on the accuracy and representativeness of its classifications and comparison tables. The taxonomy itself is defensible, and the manuscript makes its scope and contributions explicit. However, the reliability of the LLM and dataset comparisons is load-bearing for the 'comprehensive analysis' claim, and the factual inconsistencies identified below must be resolved before the survey can serve as a dependable reference.

major comments (4)
  1. [Table 3, GPT-4 row] The GPT-4 row reports 1.76T parameters and 13T training tokens, citing [97], the GPT-4 Technical Report. That report explicitly does not disclose GPT-4's parameter count or training-token count; these figures are unofficial public estimates, not facts from the cited source. Since Contribution 3 (§1.2) promises a 'dedicated analysis of LLMs' in terms of parameter size, context length, training tokens, and training data, Table 3 is the empirical backbone of that contribution. Please replace such entries with 'not disclosed' or explicitly mark them as unofficial estimates with a separate citation, and likewise verify the 'Training Dataset' column for proprietary models.
  2. [Table 3, other rows] Other rows in Table 3 need the same precision pass. 'RoBERTa [95] 355M' conflates RoBERTa-base (125M) and RoBERTa-large (355M). 'BERT [114] 137B' training tokens is not a figure reported in the BERT paper. 'PaLM-2 [113] Open Source = Yes' is inaccurate: while PaLM-2 was made available via APIs, its checkpoints were not open-sourced. Also, the column headed 'Token Size' is actually the context-window length; rename it. These issues matter because practitioners may extract Table 3 as a standalone quick reference.
  3. [Section 1.1] The paper-selection strategy is not reproducible. The text lists databases and journal/conference venues, but gives no search strings, date ranges, numbers of papers retrieved/screened/included, or explicit inclusion/exclusion criteria. 'Citation count' and 'significant impact' are subjective ranking criteria. For a survey whose central claim is comprehensiveness, this omission weakens confidence that the taxonomy and tables are representative. Add a systematic protocol or temper the 'comprehensive' claim accordingly.
  4. [Section 5.3 and Table 4] The dataset comparison is internally inconsistent. Section 5.3 states that SQuAD 2.0 combines '100,000+ questions from the original dataset with 50,000+ unanswerable questions,' which implies at least 150K questions, but Table 4 lists 129K. If 129K refers only to the train split, that should be stated explicitly. Since the survey promises a 'comprehensive analysis of key ConvQA datasets,' factual consistency in Table 4 is essential.
minor comments (4)
  1. [Table 1] The row/column alignment in Table 1 is hard to verify from the typeset version, and some classifications appear ambiguous. Please ensure each reference's check marks match the descriptions in Sections 2.1–2.3, and add a legend or notes to clarify overlapping categories such as question rewriting vs. question reformulation.
  2. [Section 4.3] Claims about Gemini 2.0 Flash's 'unified tokenization and cross-modal attention' should cite the primary technical/model documentation rather than only a comparative preprint [101].
  3. [References] Several web citations ([106], [110], [112]) lack consistent access dates or version identifiers. A uniform citation style would improve verifiability.
  4. [Section 5.4] CANARD is described as having 40K question-answer pairs and 5K conversations; this is fine, but add a sentence clarifying that it is derived from QuAC and is a rewriting benchmark, since that context is useful for readers.

Circularity Check

0 steps flagged

No circular derivation; survey is an independent literature exposition with only non-load-bearing self-citations.

full rationale

This paper is a survey, so its 'derivation chain' is a literature organization and summarization task rather than a mathematical derivation. The central claims ('comprehensive exploration', taxonomy of history selection/question understanding/answer prediction, ML techniques, LLMs, datasets) are supported by a cited corpus and the survey's own descriptive text, not by fitting parameters or deriving predictions from assumptions. The several self-citations ([23], [29], [41], [44], [46], [47]) are used as examples of prior approaches or as sources for descriptive distinctions such as hard vs. soft history selection; however, no benchmark result, equation, or prediction is forced by these citations, and the survey invokes no self-cited uniqueness theorem and does not forbid alternatives. The skeptical concerns about Table 3 — e.g., GPT-4's 1.76T parameters and 13T training tokens being unofficial estimates not disclosed in the cited GPT-4 technical report, and BERT's '137B' training-token figure not appearing in the BERT paper — are factual/correctness risks for the survey's reliability, not circularity, because they do not reduce the survey's claims to its own inputs. No specific circular reduction can be exhibited, so the appropriate circularity score is 0.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

This survey contributes no free parameters or invented entities. Its load-bearing assumptions are about literature coverage and classification fidelity, plus the accuracy of its model and dataset tables.

axioms (3)
  • domain assumption The selected literature is representative of the state of the art in ConvQA.
    Section 1.1 uses Google Scholar and citation counts without a systematic protocol, so coverage may be biased toward highly cited works.
  • domain assumption The taxonomy (history selection, question understanding, answer prediction; plus the five ML techniques) is a faithful way to describe all surveyed systems.
    The survey imposes this structure on heterogeneous papers; some papers may not fit cleanly into a single category, as seen in Table 1 where [43] appears in multiple history-selection columns.
  • domain assumption The factual specifications in Table 3 are accurate.
    The table lists parameter counts and training tokens for models; some entries, such as GPT-4's 1.76T parameters, are not supported by the cited technical report [97].

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of A Survey of the State-of-the-Art in Conversational Question Answering Systems." pith.science (2026). https://pith.science/paper/I3KRGWX4

@misc{pith2026250905716,
  author       = {Pith},
  title        = {Pith review of: A Survey of the State-of-the-Art in Conversational Question Answering Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I3KRGWX4}},
  note         = {Machine review of arXiv:2509.05716}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Conversational Question Answering (ConvQA) systems have emerged as a pivotal area within Natural Language Processing (NLP) by driving advancements that enable machines to engage in dynamic and context-aware conversations. These capabilities are increasingly being applied across various domains, i.e., customer support, education, legal, and healthcare where maintaining a coherent and relevant conversation is essential. Building on recent advancements, this survey provides a comprehensive analysis of the state-of-the-art in ConvQA. This survey begins by examining the core components of ConvQA systems, i.e., history selection, question understanding, and answer prediction, highlighting their interplay in ensuring coherence and relevance in multi-turn conversations. It further investigates the use of advanced machine learning techniques, including but not limited to, reinforcement learning, contrastive learning, and transfer learning to improve ConvQA accuracy and efficiency. The pivotal role of large language models, i.e., RoBERTa, GPT-4, Gemini 2.0 Flash, Mistral 7B, and LLaMA 3, is also explored, thereby showcasing their impact through data scalability and architectural advancements. Additionally, this survey presents a comprehensive analysis of key ConvQA datasets and concludes by outlining open research directions. Overall, this work offers a comprehensive overview of the ConvQA landscape and provides valuable insights to guide future advancements in the field.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

130 extracted references · 61 canonical work pages · 16 internal anchors

  1. [1]

    Frontiers in Computer Science6, 1486581 (2025)

    Orynbay, L., Bekmanova, G., Yergesh, B., Omarbekova, A., Sairanbekova, A., Sharipbay, A.: The Role of Cognitive Computing in NLP. Frontiers in Computer Science6, 1486581 (2025)

    2025

  2. [2]

    Language Resources and Evaluation, 1–28 (2024)

    Frenda, S., Abercrombie, G., Basile, V., Pedrani, A., Panizzon, R., Cignarella, A.T., Marco, C., Bernardi, D.: Perspectivist Approaches To Natural Language Processing: A Survey. Language Resources and Evaluation, 1–28 (2024)

    2024

  3. [3]

    In: Proceedings of the 3rd International Confer- ence on Intelligent Communication Technologies and Virtual Mobile Networks, ICICV 2021, pp

    Nagarhalli, T.P., Vaze, V., Rana, N.K.: Impact of Machine Learning in Natural Language Processing: A Review. In: Proceedings of the 3rd International Confer- ence on Intelligent Communication Technologies and Virtual Mobile Networks, ICICV 2021, pp. 1529–1534. Institute of Electrical and Electronics Engineers Inc., Online (2021) 32

    2021

  4. [4]

    The Routledge Handbook of Translation and Technology, 419–436 (2019)

    Melby, A.K.: Future of Machine Translation: Musings on Weaver’s Memo. The Routledge Handbook of Translation and Technology, 419–436 (2019)

    2019

  5. [5]

    In: Proceedings of 2020 IEEE 3rd Interna- tional Conference of Safe Production and Informatization, IICSPI 2020, pp

    Jiang, K., Lu, X.: Natural Language Processing and Its Applications in Machine Translation: A Diachronic Review. In: Proceedings of 2020 IEEE 3rd Interna- tional Conference of Safe Production and Informatization, IICSPI 2020, pp. 210–214. Institute of Electrical and Electronics Engineers Inc., Online (2020)

    2020

  6. [6]

    (eds.) Prolog: Past, Present, and Future, pp

    Gupta, G., Salazar, E., Shakerin, F., Arias, J., Varanasi, S.C., Basu, K., Wang, H., Li, F., Erbatur, S., Padalkar, P., Rajasekharan, A., Zeng, Y., Carro, M.: In: Warren, D.S., Dahl, V., Eiter, T., Hermenegildo, M.V., Kowalski, R., Rossi, F. (eds.) Prolog: Past, Present, and Future, pp. 48–61. Springer, Cham (2023)

    2023

  7. [7]

    Johnson, M.L.: SHRDLU, Procedures, Mini-world, pp. 113–122. Palgrave Macmillan UK, London (1988)

    1988

  8. [8]

    Philosoph- ical Transactions of the Royal Society A381(2251), 20220049 (2023)

    Wahlster, W.: Understanding computational dialogue understanding. Philosoph- ical Transactions of the Royal Society A381(2251), 20220049 (2023)

    2023

  9. [9]

    Expert Systems with Applications209, 118221 (2022)

    Shao, Z., Zhao, R., Yuan, S., Ding, M., Wang, Y.: Tracing the Evolution of AI in the Past Decade and Forecasting the Emerging Trends. Expert Systems with Applications209, 118221 (2022)

    2022

  10. [10]

    Engineering6(3), 275–290 (2020)

    Zhou, M., Duan, N., Liu, S., Shum, H.-Y.: Progress in Neural NLP: Modeling, Learning, and Reasoning. Engineering6(3), 275–290 (2020)

    2020

  11. [11]

    ACM Transactions on Asian and Low-Resource Language Information Processing

    Rai, A., Borah, S.: Tokenization and Stemming of Limbu Language. ACM Transactions on Asian and Low-Resource Language Information Processing. (2025)

    2025

  12. [12]

    Natural Language Processing Journal10, 100132 (2025)

    Azmi, N.S.A.B.N., Ptaszynski, M., Masui, F., Eronen, J., Nowakowski, K.: Token and Part-of-speech Fusion for Pretraining of Transformers with Application in Automatic Cyberbullying Detection. Natural Language Processing Journal10, 100132 (2025)

    2025

  13. [13]

    SN Computer Science6(2), 106 (2025)

    Bharathi Mohan, G., Prasanna Kumar, R., Krishna Jayanth, K., Doss, S.: Telugu Language Analysis with XLM-RoBERTa: Enhancing Parts of Speech Tagging for Effective Natural Language Processing. SN Computer Science6(2), 106 (2025)

    2025

  14. [14]

    IEEE Access13, 30444–30468 (2025)

    Albladi, A., Islam, M., Seals, C.: Sentiment Analysis of Twitter Data Using NLP Models: A Comprehensive Review. IEEE Access13, 30444–30468 (2025)

    2025

  15. [15]

    SN Computer Science6(1), 1–18 (2025)

    Chabukswar, A., Shenoy, P.D., Venugopal, K.: Detecting Misinformation in COVID-19 Content: A Machine Learning and Deep Learning Approach with Word Embeddings. SN Computer Science6(1), 1–18 (2025)

    2025

  16. [16]

    Human Trans- lation: A Comparative Study of Translation Quality

    Haseeb, M., Akbar, M., Abbasi, W.S.: Machine Translation vs. Human Trans- lation: A Comparative Study of Translation Quality. Social Science Review Archives3(1), 885–894 (2025)

    2025

  17. [17]

    Artificial Intelligence Review58(5), 127 (2025)

    Jolfaei, S.A., Mohebi, A.: A Review on Persian Question Answering Systems: From Traditional to Modern Approaches. Artificial Intelligence Review58(5), 127 (2025)

    2025

  18. [18]

    https://doi.org/10

    Rayhan, A., Kinzler, R., Rayhan, R.: Natural Language Processing: Transform- ing How Machines Understand Human Language (2023). https://doi.org/10. 13140/RG.2.2.34900.99200

    work page arXiv 2023

  19. [19]

    International Journal of Computational Intelligence Systems18(1), 21 (2025)

    Rana, M.R.R., Nawaz, A., Rehman, S.U., Abid, M.A., Garayevi, M., Kajanov´ a, 33 J.: BERT-BiGRU-Senti-GCN: An Advanced NLP Framework for Analyzing Customer Sentiments in e-Commerce. International Journal of Computational Intelligence Systems18(1), 21 (2025)

    2025

  20. [20]

    RAMQA: A Unified Framework for Retrieval-Augmented Multi-Modal Question Answering

    Bai, Y., Grant, C.E., Wang, D.Z.: RAMQA: A Unified Framework for Retrieval- augmented Multi-modal Question Answering (2025). https://arxiv.org/abs/ 2501.13297

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  21. [21]

    Knowledge and Information Systems65(4), 1399–1485 (2023)

    Abdel-Nabi, H., Awajan, A., Ali, M.Z.: Deep Learning-based Question Answer- ing: A Survey. Knowledge and Information Systems65(4), 1399–1485 (2023)

    2023

  22. [22]

    SDNet: Contextualized Attention-based Deep Network for Conversational Question Answering

    Zhu, C., Zeng, M., Huang, X.: SDNet: Contextualized Attention-based Deep Network for Conversational Question Answering (2019). https://arxiv.org/abs/ 1812.03593

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  23. [23]

    Knowledge and Information Systems64(12), 3151–3195 (2022)

    Zaib, M., Zhang, W., Sheng, Q., Mahmood, A., Zhang, Y.: Conversational question answering. Knowledge and Information Systems64(12), 3151–3195 (2022)

    2022

  24. [24]

    In: Proceedings of the 17th ACM International Conference on Web Search and Data Mining

    Abbasiantaeb, Z., Yuan, Y., Kanoulas, E., Aliannejadi, M.: Let the LLMs Talk: Simulating Human-to-human Conversational QA via Zero-Shot LLM-to-LLM Interactions. In: Proceedings of the 17th ACM International Conference on Web Search and Data Mining. WSDM ’24, pp. 8–17. Association for Computing Machinery, Merida, Mexico (2024)

    2024

  25. [25]

    In: The 41st International ACM SIGIR Conference on Research & Development in Informa- tion Retrieval

    Gao, J., Galley, M., Li, L.: Neural Approaches to Conversational AI. In: The 41st International ACM SIGIR Conference on Research & Development in Informa- tion Retrieval. SIGIR ’18, pp. 1371–1374. Association for Computing Machinery, Melbourne, Australia (2018)

    2018

  26. [26]

    In: Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval

    Qu, C., Yang, L., Qiu, M., Croft, W.B., Zhang, Y., Iyyer, M.: Bert with his- tory answer embedding for conversational question answering. In: Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. SIGIR’19, pp. 1133–1136. Association for Computing Machinery, Paris, France (2019)

    2019

  27. [27]

    ACM Comput

    Biancofiore, G.M., Deldjoo, Y., Noia, T.D., Di Sciascio, E., Narducci, F.: Inter- active Question Answering Systems: Literature Review. ACM Comput. Surv. 56(9) (2024)

    2024

  28. [28]

    IEEE Computational Intelligence Magazine13, 55–75 (2018)

    Young, T., Hazarika, D., Poria, S., Cambria, E.: Recent trends in deep learn- ing based natural language processing [review article]. IEEE Computational Intelligence Magazine13, 55–75 (2018)

    2018

  29. [29]

    Springer, Singapore (2021)

    Zaib, M., Tran, D.H., Sagar, S., Mahmood, A., Zhang, W.E., Sheng, Q.Z.: BERT-CoQAC: BERT-based Conversational Question Answering in Context. Springer, Singapore (2021)

    2021

  30. [30]

    In: Barzilay, R., Kan, M.-Y

    Tian, Z., Yan, R., Mou, L., Song, Y., Feng, Y., Zhao, D.: How to Make Con- text More Useful? An Empirical Study on Context-Aware Neural Conversational Models. In: Barzilay, R., Kan, M.-Y. (eds.) Proceedings of the 55th Annual Meet- ing of the Association for Computational Linguistics, pp. 231–236. Association for Computational Linguistics, Vancouver, Cana...

    2017

  31. [31]

    In: European Conference on Information Retrieval, pp

    Raposo, G., Ribeiro, R., Martins, B., Coheur, L.: Question Rewriting? Assess- ing Its Importance For Conversational Question Answering. In: European Conference on Information Retrieval, pp. 199–206 (2022). Springer

    2022

  32. [32]

    https://arxiv.org/ abs/2401.10225

    Liu, Z., Ping, W., Roy, R., Xu, P., Lee, C., Shoeybi, M., Catanzaro, B.: ChatQA: 34 Surpassing GPT-4 on Conversational QA and RAG (2024). https://arxiv.org/ abs/2401.10225

    Pith/arXiv arXiv 2024

  33. [33]

    In: Proceedings of the 14th ACM International Conference on Web Search and Data Mining, pp

    Vakulenko, S., Longpre, S., Tu, Z., Anantha, R.: Question rewriting for con- versational question answering. In: Proceedings of the 14th ACM International Conference on Web Search and Data Mining, pp. 355–363 (2021)

    2021

  34. [34]

    In: Proceedings of the First Workshop on NLP for Conversational AI

    Ohsugi, Y., Saito, I., Nishida, K., Asano, H., Tomita, J.: A Simple But Effective Method To Incorporate Multi-turn Context With BERT For Conversational Machine comprehension. In: Proceedings of the First Workshop on NLP for Conversational AI. Association for Computational Linguistics, Florence, Italy (2019)

    2019

  35. [35]

    In: Merlo, P., Tiedemann, J., Tsarfaty, R

    Kacupaj, E., Plepi, J., Singh, K., Thakkar, H., Lehmann, J., Maleshkova, M.: Conversational Question Answering Over Knowledge Graphs with Transformer and Graph Attention Networks. In: Merlo, P., Tiedemann, J., Tsarfaty, R. (eds.) Proceedings of the 16th Conference of the European Chapter of the Associa- tion for Computational Linguistics, pp. 850–862. Ass...

    2021

  36. [36]

    In: Vlachos, A., Augenstein, I

    Perez-Beltrachini, L., Jain, P., Monti, E., Lapata, M.: Semantic Parsing for Conversational Question Answering over Knowledge Graphs. In: Vlachos, A., Augenstein, I. (eds.) Proceedings of the 17th Conference of the European Chap- ter of the Association for Computational Linguistics, pp. 2507–2522. Association for Computational Linguistics, Dubrovnik, Croa...

    2023

  37. [37]

    In: Pro- ceedings of the 17th ACM International Conference on Web Search and Data Mining

    Kaiser, M., Saha Roy, R., Weikum, G.: Robust Training for Conversational Question Answering Models with Reinforced Reformulation Generation. In: Pro- ceedings of the 17th ACM International Conference on Web Search and Data Mining. WSDM ’24, pp. 322–331. Association for Computing Machinery, Merida, Mexico (2024)

    2024

  38. [38]

    Few-shot Policy (de)composition in Conversational Question Answering

    Erwin, K., Axelrod, G., Chang, M., Fokoue, A., Crouse, M., Dan, S., Gao, T., Uceda-Sosa, R., Makondo, N., Khan, N., et al.: Few-shot Policy (de) composition in Conversational Question Answering. arXiv preprint arXiv:2501.11335 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  39. [39]

    (eds.) Proceedings of the Third Work- shop on Insights from Negative Results in NLP, pp

    Ishii, E., Xu, Y., Cahyawijaya, S., Wilie, B.: Can Question Rewriting Help Conversational Question Answering? In: Tafreshi, S., Sedoc, J., Rogers, A., Drozd, A., Rumshisky, A., Akula, A. (eds.) Proceedings of the Third Work- shop on Insights from Negative Results in NLP, pp. 94–99. Association for Computational Linguistics, Dublin, Ireland (2022)

    2022

  40. [40]

    In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval

    Ye, L., Lei, Z., Yin, J., Chen, Q., Zhou, J., He, L.: Boosting Conversational Ques- tion Answering with Fine-grained Retrieval-augmentation and Self-check. In: Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. SIGIR ’24, pp. 2301–2305. Association for Computing Machinery, Washington DC, USA (2024)

    2024

  41. [41]

    In: Advanced Data Mining and Applications, pp

    Perera, M.M., Mahmood, A., Wijethilake, K.E., Sheng, Q.Z.: Towards Adap- tive Context Management for Intelligent Conversational Question Answering. In: Advanced Data Mining and Applications, pp. 360–375. Springer, Singapore (2025)

    2025

  42. [42]

    In: Zong, C., Xia, F., Li, W., Navigli, R

    Kim, G., Kim, H., Park, J., Kang, J.: Learn to Resolve Conversational Dependency: A Consistency Training Framework for Conversational Question Answering. In: Zong, C., Xia, F., Li, W., Navigli, R. (eds.) Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language P...

    2021

  43. [43]

    Preference-based Learning with Retrieval Augmented Generation for Conversational Question Answering

    Kaiser, M., Weikum, G.: Preference-based Learning with Retrieval Aug- mented Generation for Conversational Question Answering. arXiv preprint arXiv:2503.22303 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  44. [44]

    In: 2023 International Joint Conference on Neural Net- works (IJCNN)

    Zaib, M., Sheng, Q., Zhang, W., Mahmood, A.: Keeping the questions conversa- tional: Using structured representations to resolve dependency in conversational question answering. In: 2023 International Joint Conference on Neural Net- works (IJCNN). Institute of Electrical and Electronics Engineers (IEEE), Online (2023). 2023 International Joint Conference ...

    2023

  45. [45]

    Pro- ceedings of the AAAI Conference on Artificial Intelligence35(15), 13718–13726 (2021)

    Qiu, M., Huang, X., Chen, C., Ji, F., Qu, C., Wei, W., Huang, J., Zhang, Y.: Reinforced History Backtracking for Conversational Question Answering. Pro- ceedings of the AAAI Conference on Artificial Intelligence35(15), 13718–13726 (2021)

    2021

  46. [46]

    In: Lecture Notes in Computer Science, vol

    Zaib, M., Zhang, W.E., Sheng, Q.Z., Sagar, S., Mahmood, A., Zhang, Y.: Learn- ing to Select the Relevant History Turns in Conversational Question Answering. In: Lecture Notes in Computer Science, vol. 14306 LNCS, pp. 334–348. Springer, Berlin, Heidelberg (2023)

    2023

  47. [47]

    In: International Conference on Web Information Systems Engineering, pp

    Zaib, M., Sheng, Q.Z., Zhang, W.E., Alhazmi, E., Mahmood, A.: Learning contrastive representations for dense passage retrieval in open-domain conver- sational question answering. In: International Conference on Web Information Systems Engineering, pp. 3–13 (2024). Springer

    2024

  48. [48]

    https://arxiv.org/abs/2407.21712

    Wang, X., Sen, P., Li, R., Yilmaz, E.: Adaptive Retrieval-augmented Generation for Conversational Systems (2024). https://arxiv.org/abs/2407.21712

    Pith/arXiv arXiv 2024

  49. [49]

    SIGIR ’24, pp

    Kostric, I., Balog, K.: A Surprisingly Simple yet Effective Multi-query Rewrit- ing Method for Conversational Passage Retrieval. SIGIR ’24, pp. 2271–2275. Association for Computing Machinery, New York, NY, USA (2024)

    2024

  50. [50]

    In: 2024 IEEE 18th International Conference on Semantic Computing (ICSC), pp

    Rashid, M.S., Meem, J.A., Hristidis, V.: NORMY: Non-Uniform History Model- ing for Open Retrieval Conversational Question Answering. In: 2024 IEEE 18th International Conference on Semantic Computing (ICSC), pp. 101–109 (2024). IEEE

    2024

  51. [51]

    In: Ku, L.-W., Martins, A., Srikumar, V

    Liu, L., Hill, B., Du, B., Wang, F., Tong, H.: Conversational Question Answering with Language Models Generated Reformulations over Knowledge Graph. In: Ku, L.-W., Martins, A., Srikumar, V. (eds.) Findings of the Association for Com- putational Linguistics ACL 2024, pp. 839–850. Association for Computational Linguistics, Bangkok, Thailand (2024)

    2024

  52. [52]

    Neural Computing and Applications36(16), 8995–9022 (2024)

    Hu, Z., Hou, W., Liu, X.: Deep Learning For Named Entity Recognition: A Survey. Neural Computing and Applications36(16), 8995–9022 (2024)

    2024

  53. [53]

    IEEE Transactions on Emerging Topics in Computational Intelligence8(3), 2640–2653 (2024)

    Liu, Z., He, J., Gong, T., Weng, H., Wang, F.L., Liu, H., Hao, T.: Improv- ing Topic Tracing with a Textual Reader for Conversational Knowledge Based Question Answering. IEEE Transactions on Emerging Topics in Computational Intelligence8(3), 2640–2653 (2024)

    2024

  54. [54]

    Symmetry16(9) (2024)

    Jiang, P., Cai, X.: A Survey of Semantic Parsing Techniques. Symmetry16(9) (2024)

    2024

  55. [55]

    Journal of Computer and Communications12(10), 1–13 (2024) 36

    Wu, Z.: Large Language Model Based Semantic Parsing for Intelligent Database Query Engine. Journal of Computer and Communications12(10), 1–13 (2024) 36

    2024

  56. [56]

    Evaluating Large Language Models in Semantic Parsing for Conversational Question Answering over Knowledge Graphs

    Schneider, P., Klettner, M., Jokinen, K., Simperl, E., Matthes, F.: Evaluat- ing Large Language Models in Semantic Parsing for Conversational Question Answering over Knowledge Graphs (2024). https://arxiv.org/abs/2401.01711

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  57. [57]

    In: Proceedings of the 10th SIGHAN Workshop on Chinese Language Processing (SIGHAN- 10), pp

    Bai, Z., Wang, B., Liang, B., Xu, R.: Auto-ACE: An Automatic Answer Correct- ness Evaluation Method for Conversational Question Answering. In: Proceedings of the 10th SIGHAN Workshop on Chinese Language Processing (SIGHAN- 10), pp. 80–87. Association for Computational Linguistics, Bangkok, Thailand (2024)

    2024

  58. [58]

    IEEE Transactions on Computational Social Systems11(2), 1888–1906 (2024)

    Ahmed, M., Khan, H.U., Munir, E.U.: Conversational ai: An explication of few-shot learning problem in transformers-based chatbot systems. IEEE Transactions on Computational Social Systems11(2), 1888–1906 (2024)

    1906

  59. [59]

    https://arxiv.org/abs/2405.17822

    Pan, Z., Luo, H., Li, M., Liu, H.: Conv-CoA: Improving Open-domain Question Answering in Large Language Models via Conversational Chain-of-action (2024). https://arxiv.org/abs/2405.17822

    Pith/arXiv arXiv 2024

  60. [60]

    Soft Computing25(11), 7341–7378 (2021)

    Ghahramani, F., Tahayori, H., Visconti, A.: Effects of central tendency mea- sures on term weighting in textual information retrieval. Soft Computing25(11), 7341–7378 (2021)

    work page 2021

  61. [61]

    In: Rogers, A., Boyd-Graber, J., Okazaki, N

    Jeong, S., Baek, J., Hwang, S.J., Park, J.: Phrase Retrieval for Open Domain Conversational Question Answering with Conversational Dependency Modeling via Contrastive Learning. In: Rogers, A., Boyd-Graber, J., Okazaki, N. (eds.) Findings of the Association for Computational Linguistics: ACL 2023, pp. 6019–

    work page 2023

  62. [62]

    In: Che, W., Shutova, E

    Xia, F., Li, B., Weng, Y., He, S., Liu, K., Sun, B., Li, S., Zhao, J.: MedConQA: Medical conversational question answering system based on knowledge graphs. In: Che, W., Shutova, E. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pp. 148–158. Association for Computational Linguistics, A...

    work page 2022

  63. [63]

    Transactions of the Association for Computational Linguistics11, 1–17 (2023)

    Siriwardhana, S., Weerasekera, R., Wen, E., Kaluarachchi, T., Rana, R., Nanayakkara, S.: Improving the Domain Adaptation of Retrieval Augmented Generation (RAG) Models for Open Domain Question Answering. Transactions of the Association for Computational Linguistics11, 1–17 (2023)

    work page 2023

  64. [64]

    https://arxiv.org/abs/2501.12789

    Filice, S., Horowitz, G., Carmel, D., Karnin, Z., Lewin-Eytan, L., Maarek, Y.: Generating Diverse Q&A Benchmarks for RAG Evaluation with DataMorgana (2025). https://arxiv.org/abs/2501.12789

    Pith/arXiv arXiv 2025

  65. [65]

    In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management

    Kacupaj, E., Singh, K., Maleshkova, M., Lehmann, J.: Contrastive Representa- tion Learning for Conversational Question Answering over Knowledge Graphs. In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management. CIKM ’22, pp. 925–934. Association for Computing Machinery, Atlanta, GA, USA (2022)

    work page 2022

  66. [66]

    Computers & Indus- trial Engineering, 110856 (2025)

    Khadivi, M., Charter, T., Yaghoubi, M., Jalayer, M., Ahang, M., Shojaeinasab, A., Najjaran, H.: Deep Reinforcement Learning For Machine Scheduling: Methodology, The State-of-the-art, and Future Directions. Computers & Indus- trial Engineering, 110856 (2025)

    work page 2025

  67. [67]

    IEEE Transactions on Neural Networks and Learning Systems, 1–21 (2024) 37

    Cao, Y., Zhao, H., Cheng, Y., Shu, T., Chen, Y., Liu, G., Liang, G., Zhao, J., Yan, J., Li, Y.: Survey on Large Language Model-enhanced Reinforcement Learning: Concept, Taxonomy, and Methods. IEEE Transactions on Neural Networks and Learning Systems, 1–21 (2024) 37

    work page 2024

  68. [68]

    https://arxiv.org/abs/2405.11106

    Sun, C., Huang, S., Pompili, D.: LLM-based Multi-agent Reinforcement Learn- ing: Current and Future Directions (2024). https://arxiv.org/abs/2405.11106

    Pith/arXiv arXiv 2024

  69. [69]

    In: Li, Y., Lazaridou, A

    Chen, Z., Zhao, J., Fang, A., Fetahu, B., Rokhlenko, O., Malmasi, S.: Rein- forced Question Rewriting for Conversational Question Answering. In: Li, Y., Lazaridou, A. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pp. 357–370. Association for Computational Linguistics, Abu Dhabi, UAE (2022)

    work page 2022

  70. [70]

    Reinforcement Learning for Conversational Question Answering over Knowledge Graph

    Wu, M.: Reinforcement Learning for Conversational Question Answering over Knowledge Graph (2024). https://arxiv.org/abs/2401.08460

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  71. [71]

    Pattern Recognition159, 111095 (2025)

    Sun, T., Chen, H., Hu, G., Zhao, C.: Explainability-based knowledge distillation. Pattern Recognition159, 111095 (2025)

    work page 2025

  72. [72]

    DDK: Distilling Domain Knowledge for Efficient Large Language Models

    Liu, J., Zhang, C., Guo, J., Zhang, Y., Que, H., Deng, K., Bai, Z., Liu, J., Zhang, G., Wang, J., Wu, Y., Liu, C., Su, W., Wang, J., Qu, L., Zheng, B.: DDK: Distilling Domain Knowledge for Efficient Large Language Models (2024). https://arxiv.org/abs/2407.16154

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  73. [73]

    In: Carpuat, M., Marneffe, M.-C., Meza Ruiz, I.V

    You, C., Chen, N., Liu, F., Ge, S., Wu, X., Zou, Y.: End-to-end spoken conversa- tional question answering: Task, dataset and model. In: Carpuat, M., Marneffe, M.-C., Meza Ruiz, I.V. (eds.) Findings of the Association for Computational Linguistics: NAACL 2022. Association for Computational Linguistics, Seattle, United States (2022)

    work page 2022

  74. [74]

    https://arxiv.org/abs/1909.10772

    Ju, Y., Zhao, F., Chen, S., Zheng, B., Yang, X., Liu, Y.: Technical Report on Conversational Question Answering (2019). https://arxiv.org/abs/1909.10772

    Pith/arXiv arXiv 2019

  75. [75]

    Engineering Applications of Artificial Intelligence123, 106137 (2023)

    Boreshban, Y., Mirbostani, S.M., Ghassem-Sani, G., Mirroshandel, S.A., Amiri- parian, S.: Improving question answering performance using knowledge distilla- tion and active learning. Engineering Applications of Artificial Intelligence123, 106137 (2023)

    work page 2023

  76. [76]

    Computer Modeling in Engineering & Sciences

    Yin, L., Wang, L., Cai, Z., Lu, S., Wang, R., AlSanad, A., AlQahtani, S.A., Chen, X., Yin, Z., Li, X., et al.: Dpal-bert: A faster and lighter question answering model. Computer Modeling in Engineering & Sciences

    work page

  77. [77]

    In: Carpuat, M., Marneffe, M.-C., Meza Ruiz, I.V

    Caciularu, A., Dagan, I., Goldberger, J., Cohan, A.: Long context question answering via supervised contrastive learning. In: Carpuat, M., Marneffe, M.-C., Meza Ruiz, I.V. (eds.) Proceedings of the 2022 Conference of the North Ameri- can Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 2872–2879. Association for C...

    work page 2022

  78. [78]

    Computerized Medical Imaging and Graphics121, 102500 (2025)

    Xu, X., Wong, S.T.C.: Contrastive Learning in Brain Imaging. Computerized Medical Imaging and Graphics121, 102500 (2025)

    work page 2025

  79. [79]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Gao, X., Das, K.: Customizing Language Model Responses with Contrastive In-Context Learning. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, pp. 18039–18046 (2024)

    work page 2024

  80. [80]

    In: Goldberg, Y., Kozareva, Z., Zhang, Y

    Zhang, Z., Strubell, E., Hovy, E.: A Survey of Active Learning for Natural Lan- guage Processing. In: Goldberg, Y., Kozareva, Z., Zhang, Y. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pp. 6166–6190. Association for Computational Linguistics, Online (2022)

    work page 2022

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.