Open-Domain Question Answering Goes Conversational via Question Rewriting
Open accessAbstract
Raviteja Anantha, Svitlana Vakulenko, Zhucheng Tu, Shayne Longpre, Stephen Pulman, Srinivas Chappidi. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Retrieval accuracy in long-term dialogue frameworks can be compromised because retrieval models are generally trained on semantic similarity rather than such dialogues, and real-world dialogues often feature co-references and missing content.
“The accuracy of retrieval can be compromised, as the retrieval model is generally trained on tasks focusing on semantic similarity rather than specifically on such dialogues. Additionally, real-world dialogues often feature co-references and missing content ( i.e., anaphora) (Anantha et al., 2021), which further complicate the retrieval process (Mallen et al., 2023; Gao et al., 2023b; Liu et al.,”
From Evaluating Very Long-Term Conversational Memory of LLM Agents (2024) · cites this paper as Anantha et al., 2021
The accuracy of retrieval in long-term dialogue frameworks can be compromised because retrieval models are generally trained on semantic similarity rather than such dialogues.
“The accuracy of retrieval can be compromised, as the retrieval model is generally trained on tasks focusing on semantic similarity rather than specifically on such dialogues.”
From Evaluating Very Long-Term Conversational Memory of LLM Agents (2024) · cites this paper as Anantha et al., 2021
Real-world dialogues often feature co-references and missing content (anaphora), which complicate retrieval.
“Additionally, real-world dialogues often feature co-references and missing content ( i.e., anaphora) (Anantha et al., 2021), which further complicate the retrieval process (Mallen et al., 2023; Gao et al., 2023b; Liu et al., 2023)”
From Evaluating Very Long-Term Conversational Memory of LLM Agents (2024) · cites this paper as Anantha et al., 2021
The citation lists haven’t been fetched yet
References 24 · Cited by 130
- Authors
- Anantha, R., Vakulenko, S., Tu, Z., Longpre, S., Pulman, S., & Chappidi, S.
- Affiliations
- Apple (Israel) (IL)
- University of Amsterdam (NL)
- Venue
- Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
- Published
- 2021
- Pages
- 520-534
- OpenAlex
- W3092307876
- Cited by
- 130
- References
- 24
- Citations by year
- 2026: 17 · 2025: 17 · 2024: 19 · 2023: 38 · 2022: 25 · 2021: 13 · 2020: 1
- FWCI
- 10.97
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Keywords
- open-domain question answering, question rewriting, conversational question answering, question generation, question answering systems
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
