Cause-Conditioned Multi-Task Learning for Answerable Question Suggestion in MRC

Hadiseh Moradisani, Fattane Zarrinkalam, Julien Serbanescu, Zeinab Noorian
Proceedings of the The 39th Canadian Conference on Artificial Intelligence, PMLR 318:248-259, 2026.

Abstract

Machine Reading Comprehension (MRC) systems struggle when user questions are unanswerable given the passage: most simply output “no answer”, leaving users without guidance on how to recover useful information. We introduce a \textit{cause-conditioned multi-task learning (MTL)} framework that turns failure into follow-up by jointly (1) classifying an input as answerable or as one of six fine-grained unanswerability causes (Entity Swap, Number Swap, Antonym, Negation, Mutual Exclusion, No Information), and (2) generating a revised, context-grounded answerable question conditioned on the predicted cause label and an extracted guidance sentence. Using an ensemble of strong readers plus LLMs-as-judges, we apply majority voting to test whether rewrites become answerable. A human study further assesses fluency, relevance, and usefulness. Our cause-conditioning MTL framework yields better recovery from unanswerable inputs and earns strong human ratings, advancing user-supportive, failure-aware MRC.

Cite this Paper

BibTeX
@InProceedings{pmlr-v318-moradisani26a, title = {Cause-Conditioned Multi-Task Learning for Answerable Question Suggestion in MRC}, author = {Moradisani, Hadiseh and Zarrinkalam, Fattane and Serbanescu, Julien and Noorian, Zeinab}, booktitle = {Proceedings of the The 39th Canadian Conference on Artificial Intelligence}, pages = {248--259}, year = {2026}, editor = {Bouzar-Benlabiod, Lydia and Leung, Carson}, volume = {318}, series = {Proceedings of Machine Learning Research}, month = {25--29 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v318/main/assets/moradisani26a/moradisani26a.pdf}, url = {https://proceedings.mlr.press/v318/moradisani26a.html}, abstract = {Machine Reading Comprehension (MRC) systems struggle when user questions are unanswerable given the passage: most simply output “no answer”, leaving users without guidance on how to recover useful information. We introduce a \textit{cause-conditioned multi-task learning (MTL)} framework that turns failure into follow-up by jointly (1) classifying an input as answerable or as one of six fine-grained unanswerability causes (Entity Swap, Number Swap, Antonym, Negation, Mutual Exclusion, No Information), and (2) generating a revised, context-grounded answerable question conditioned on the predicted cause label and an extracted guidance sentence. Using an ensemble of strong readers plus LLMs-as-judges, we apply majority voting to test whether rewrites become answerable. A human study further assesses fluency, relevance, and usefulness. Our cause-conditioning MTL framework yields better recovery from unanswerable inputs and earns strong human ratings, advancing user-supportive, failure-aware MRC.} }
Endnote
%0 Conference Paper %T Cause-Conditioned Multi-Task Learning for Answerable Question Suggestion in MRC %A Hadiseh Moradisani %A Fattane Zarrinkalam %A Julien Serbanescu %A Zeinab Noorian %B Proceedings of the The 39th Canadian Conference on Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Lydia Bouzar-Benlabiod %E Carson Leung %F pmlr-v318-moradisani26a %I PMLR %P 248--259 %U https://proceedings.mlr.press/v318/moradisani26a.html %V 318 %X Machine Reading Comprehension (MRC) systems struggle when user questions are unanswerable given the passage: most simply output “no answer”, leaving users without guidance on how to recover useful information. We introduce a \textit{cause-conditioned multi-task learning (MTL)} framework that turns failure into follow-up by jointly (1) classifying an input as answerable or as one of six fine-grained unanswerability causes (Entity Swap, Number Swap, Antonym, Negation, Mutual Exclusion, No Information), and (2) generating a revised, context-grounded answerable question conditioned on the predicted cause label and an extracted guidance sentence. Using an ensemble of strong readers plus LLMs-as-judges, we apply majority voting to test whether rewrites become answerable. A human study further assesses fluency, relevance, and usefulness. Our cause-conditioning MTL framework yields better recovery from unanswerable inputs and earns strong human ratings, advancing user-supportive, failure-aware MRC.
APA
Moradisani, H., Zarrinkalam, F., Serbanescu, J. & Noorian, Z.. (2026). Cause-Conditioned Multi-Task Learning for Answerable Question Suggestion in MRC. Proceedings of the The 39th Canadian Conference on Artificial Intelligence, in Proceedings of Machine Learning Research 318:248-259 Available from https://proceedings.mlr.press/v318/moradisani26a.html.

Related Material