pith. sign in

arxiv: 1907.02606 · v1 · pith:QCGJGWIGnew · submitted 2019-07-04 · 💻 cs.IR · cs.AI· cs.CL

A Road-map Towards Explainable Question Answering A Solution for Information Pollution

Saeedeh Shekarpour , Faisal Alshargi This is my paper

Pith reviewed 2026-05-25 08:39 UTC · model grok-4.3

classification 💻 cs.IR cs.AIcs.CL
keywords explainable question answeringinformation pollutionAI transparencyquestion answering systemsweb information quality
0
0 comments X

The pith

Explainable Question Answering systems can alleviate web information pollution by providing transparency into their reasoning and allowing users to validate key information aspects.

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

The paper is a position paper proposing Explainable Question Answering (XQA) as a solution to information pollution on the Web. It notes that current QA systems are black boxes that do not show their learning or reasoning steps. XQA would expose details on provenance, validity, context, circulation, interpretation, and feedbacks so users can check the information. The paper identifies five guiding questions for developing XQA: defining it, explaining its necessity and timing, representing explanations, and evaluating the systems.

Core claim

The Explainable Question Answering (XQA) system can alleviate the pain of information pollution where it provides transparency to the underlying computational model and exposes an interface enabling the end-user to access and validate provenance, validity, context, circulation, interpretation, and feedbacks of information.

What carries the argument

The XQA interface for accessing and validating multiple aspects of information including provenance and validity.

If this is right

  • XQA makes the details of reasoning in QA systems visible to users.
  • Users can validate the source, context, and other properties of answers.
  • Addressing XQA requires defining its core concepts and challenges.
  • XQA is needed to tackle increasing information pollution on the Web.

Where Pith is reading between the lines

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

  • XQA principles might apply to other AI applications facing trust issues.
  • New methods for representing explanations could emerge from this roadmap.
  • Evaluation of XQA would likely involve user studies on information validation success.

Load-bearing premise

That exposing explanations in QA systems will enable end-users to effectively validate information and reduce the effects of information pollution.

What would settle it

An empirical study in which users of an XQA system fail to detect or avoid polluted information more effectively than users of standard QA systems.

Figures

Figures reproduced from arXiv: 1907.02606 by Faisal Alshargi, Saeedeh Shekarpour.

Figure 1
Figure 1. Figure 1: The existing QA systems are a black box which do not provide any explanation for their inference. ataset for Diverse, Explainable QtiAi [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: The explainable question answering exposes explainable models and explainable interface; then [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: The features of explainable interface for [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 4
Figure 4. Figure 4: The life cycle of information on the Web. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
read the original abstract

The increasing rate of information pollution on the Web requires novel solutions to tackle that. Question Answering (QA) interfaces are simplified and user-friendly interfaces to access information on the Web. However, similar to other AI applications, they are black boxes which do not manifest the details of the learning or reasoning steps for augmenting an answer. The Explainable Question Answering (XQA) system can alleviate the pain of information pollution where it provides transparency to the underlying computational model and exposes an interface enabling the end-user to access and validate provenance, validity, context, circulation, interpretation, and feedbacks of information. This position paper sheds light on the core concepts, expectations, and challenges in favor of the following questions (i) What is an XQA system?, (ii) Why do we need XQA?, (iii) When do we need XQA? (iv) How to represent the explanations? (iv) How to evaluate XQA systems?

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript is a position paper that introduces the concept of Explainable Question Answering (XQA) as a potential solution to information pollution on the Web. It argues that by providing transparency to the underlying models and enabling users to validate information aspects including provenance, validity, context, circulation, interpretation, and feedbacks, XQA can help mitigate the issues. The paper poses and aims to shed light on questions regarding the definition, necessity, timing, representation of explanations, and evaluation of XQA systems.

Significance. If the proposed ideas are pursued, XQA could enhance trust in QA interfaces within information retrieval. The paper correctly identifies the black-box nature of current QA systems as a problem. However, without any concrete examples, formal definitions, or preliminary results, the significance is limited to conceptual framing rather than a substantive contribution.

major comments (2)
  1. [Abstract] Abstract: The claim that the XQA system 'can alleviate the pain of information pollution' rests on the untested premise that user access to explanations will enable effective validation and reduce pollution; no supporting analysis or evidence is provided anywhere in the manuscript.
  2. [Abstract] Abstract: Although the paper is described as a road-map, it does not provide any specific guidance, architecture, or methodological steps for implementing or developing XQA systems, only identifying the questions to be considered.
minor comments (2)
  1. [Abstract] The list of questions contains a duplicate '(iv)' label.
  2. The title lacks proper punctuation or separation between 'Explainable Question Answering' and 'A Solution for Information Pollution'.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our position paper. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that the XQA system 'can alleviate the pain of information pollution' rests on the untested premise that user access to explanations will enable effective validation and reduce pollution; no supporting analysis or evidence is provided anywhere in the manuscript.

    Authors: As this is a position paper whose goal is to frame open research questions rather than to present validated results, we did not include empirical analysis. We agree that the original phrasing is too definitive for a conceptual proposal. We will revise the abstract to replace 'can alleviate' with 'has the potential to help alleviate' and add a brief clause noting that the benefits remain to be demonstrated through future work. revision: partial

  2. Referee: [Abstract] Abstract: Although the paper is described as a road-map, it does not provide any specific guidance, architecture, or methodological steps for implementing or developing XQA systems, only identifying the questions to be considered.

    Authors: The manuscript is explicitly framed as a position paper whose contribution is to articulate the core questions that must be resolved before concrete XQA systems can be built. In our view, a road-map at this stage consists of defining those questions (definition, necessity, timing, representation, and evaluation) rather than prescribing architectures. We therefore do not plan to add implementation details, as doing so would change the paper's stated purpose. revision: no

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

This is a position paper that defines XQA conceptually and poses open questions about its need, representation, and evaluation without any equations, formal derivations, fitted parameters, or empirical claims. The central assertion that XQA can alleviate information pollution is presented as a forward-looking motivation rather than a result derived from prior results or self-citations within the paper. No load-bearing steps exist that reduce to inputs by construction, satisfying the criteria for a score of 0.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a conceptual position paper with no formal derivations, data fits, or new entities; the ledger is empty.

pith-pipeline@v0.9.0 · 5697 in / 1091 out tokens · 32834 ms · 2026-05-25T08:39:53.342709+00:00 · methodology

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

17 extracted references · 17 canonical work pages · 1 internal anchor

  1. [1]

    HCqa: Hybrid and Complex Question Answering on Textual Corpus and Knowledge Graph

    Somayeh Asadifar, Mohsen Kahani, and Saeedeh Shekarpour. Hcqa: Hybrid and com- plex question answering on textual corpus and knowledgegraph. CoRR,abs/1811.10986,2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  2. [2]

    Ester: effi- cient search on text, entities, and relations

    Holger Bast, Alexandru Chitea, Fabian Suchanek, and Ingmar Weber. Ester: effi- cient search on text, entities, and relations. In Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, pages 671–678. ACM, 2007

    work page 2007

  3. [3]

    Rise of the racist robots – how ai is learning all our worst im- pulses

    Stephen Buranyi. Rise of the racist robots – how ai is learning all our worst im- pulses. https://www.theguardian. com/inequality/2017/aug/08/ rise-of-the-racist-robots-how-ai-is-learning-all-our-worst-impulses . Accessed: 2017-8-8

    work page 2017

  4. [4]

    Explainable artificial intelli- gence (xai)

    David Gunning. Explainable artificial intelli- gence (xai). Defense Advanced Research Projects Agency (DARPA), nd Web (2017)

    work page 2017

  5. [5]

    A postmodern pandora’s box: anti-vaccination misinformation on the inter- net

    Anna Kata. A postmodern pandora’s box: anti-vaccination misinformation on the inter- net. Vaccine, 28(7):1709–1716, 2010

    work page 2010

  6. [6]

    Qing Li, Jianlong Fu, Dongfei Yu, Tao Mei, and JieboLuo.In Proceedingsofthe2018Conferenceon EmpiricalMethodsinNaturalLanguageProcessing, Brussels, Belgium, 2018, pages 1338–1346, 2018

    work page 2018

  7. [7]

    Computer says no: why mak- ing ais fair, accountable and transparent is crucial

    Ian Sample. Computer says no: why mak- ing ais fair, accountable and transparent is crucial. https://www.theguardian. com/science/2017/nov/05/ computer-says-no-why-making-ais-fair-accountable-and-transparent-is-crucial . Accessed: 2017-11-05

    work page 2017

  8. [8]

    Keyword-driven SPARQL query generation leveraging background knowledge

    Saeedeh Shekarpour, Sören Auer, Axel- Cyrille Ngonga Ngomo, Daniel Gerber, Sebastian Hellmann, and Claus Stadler. Keyword-driven SPARQL query generation leveraging background knowledge. InProceed- ings of the 2011 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2011, Campus Scientifique de la Doua, Lyon, France, August 22-27, 2011, pages 203–...

    work page 2011

  9. [9]

    SINA: semantic interpretation of user queries for question answering on interlinked data.J

    Saeedeh Shekarpour, Edgard Marx, Axel- Cyrille Ngonga Ngomo, and Sören Auer. SINA: semantic interpretation of user queries for question answering on interlinked data.J. Web Semant., 30:39–51, 2015

    work page 2015

  10. [10]

    Question answering on interlinked data

    Saeedeh Shekarpour, Axel-Cyrille Ngonga Ngomo, and Sören Auer. Question answering on interlinked data. In22nd International World Wide Web Conference, WWW ’13, Rio de Janeiro, Brazil, May 13-17, 2013, pages 1145–1156, 2013

    work page 2013

  11. [11]

    Frankenstein: A platform enabling reuse of question answering components

    Kuldeep Singh, Andreas Both, Arun Sethu- pat Radhakrishna, and Saeedeh Shekarpour. Frankenstein: A platform enabling reuse of question answering components. In The Se- manticWeb-15thInternationalConference,ESWC 2018,Heraklion,Crete,Greece,June3-7,2018,Pro- ceedings, pages 624–638, 2018

    work page 2018

  12. [12]

    InPro- ceedingsofthe2018WorldWideWebConferenceon World Wide Web, WWW 2018, Lyon, France, April 23-27, 2018, pages 1247–1256, 2018

    Kuldeep Singh, Arun Sethupat Radhakrishna, Andreas Both, Saeedeh Shekarpour, Ioanna Ly- tra, Ricardo Usbeck, Akhilesh Vyas, Akmal Khikmatullaev, Dharmen Punjani, Christoph Lange, Maria-Esther Vidal, Jens Lehmann, and SörenAuer.Whyreinventthewheel: Let’sbuild question answering systems together. InPro- ceedingsofthe2018WorldWideWebConferenceon World Wide W...

    work page 2018

  13. [13]

    Template- based question answering over RDF data

    Christina Unger, Lorenz Bühmann, Jens Lehmann, Axel-Cyrille Ngonga Ngomo, 6 DanielGerber,andPhilippCimiano. Template- based question answering over RDF data. In Proceedingsofthe21stWorldWideWebConference 2012, WWW 2012, Lyon, France, April 16-20, 2012, pages 639–648, 2012

    work page 2012

  14. [14]

    Disinformation gets worse

    Claire Wardle. Disinformation gets worse. https://www.niemanlab.org/2017/12/ disinformation-gets-worse/. Accessed: 2018

    work page 2017

  15. [15]

    Infor- mation disorder: Toward an interdisciplinary framework for research and policymak- ing

    Claire Wardle and Hossein Derakhshan. Infor- mation disorder: Toward an interdisciplinary framework for research and policymak- ing. https://shorensteincenter.org/ information-disorder-framework-for-research-and-policymaking/ . Accessed: 2017-9-31

    work page 2017

  16. [16]

    Dynamic memory networks for visual andtextualquestionanswering

    Caiming Xiong, Stephen Merity, and Richard Socher. Dynamic memory networks for visual andtextualquestionanswering. In International conference on machine learning, pages 2397–2406, 2016

    work page 2016

  17. [17]

    Cohen, Ruslan Salakhutdi- nov,andChristopherD.Manning

    Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William W. Cohen, Ruslan Salakhutdi- nov,andChristopherD.Manning. Hotpotqa: A datasetfordiverse,explainablemulti-hopques- tion answering. InProceedings of the 2018 Con- ference on Empirical Methods in Natural Language Processing,Brussels,Belgium,October31-Novem- ber 4, 2018, pages 2369–2380, 2018. 7

    work page 2018