{"id":{"repo_id":"trento","oai_identifier":"oai:iris.unitn.it:11572/483952"},"canonical_url":"https://search.dev.ndltd.org/etd/trento/oai:iris.unitn.it:11572/483952","repository":{"repo_id":"trento","name":"Università degli Studi di Trento","base_url":"https://iris.unitn.it/oai/request"},"display":{"title":"Models and Application of Question Retrieval for Natural Language Processing","abstract":"This thesis investigates the role of question understanding in Question Answering systems, developing methods that exploit question semantic equivalence at progressively larger scales: from individual question pairs, through equivalence clusters, to entire datasets. The first part addresses question retrieval at scale. We introduce QUADRo, a retrieval framework operating over millions of question-answer pairs, and the Question Ranking Corpus (QRC), a large-scale resource with answer-aware annotations and challenging hard negatives. We demonstrate that incorporating answers during retrieval substantially improves accuracy, as answers serve as a semantic bridge between questions that share little lexical overlap but seek the same information. To reduce annotation costs, we develop Question Ranking Pre-training (QRP), a self-supervised method that learns question equivalence patterns without labeled data, achieving significant improvements while reducing model variance by over 50\\%. The second part extends pairwise equivalence to question clusters. We analyze coherence in Large Language Models, finding that a substantial portion of question clusters exhibit incoherent behavior: models answer some phrasings correctly while failing on semantically equivalent alternatives. This reveals that understanding failures, not just knowledge gaps, limit LLM performance. We introduce Question-Augmented Generation (q-RAG), which supplements prompts with retrieved similar questions, improving accuracy by up to 9 percentage points and coherence by up to 28 points. We further show that q-RAG's benefits can be distilled into model parameters through Direct Preference Optimization (DPO) and Supervised Fine-Tuning, producing standalone models with improved coherence that surpass the inference-time approach. For retrieval systems, we apply clusters to train models for consistency: the Coherence Ranking Loss improves ranking coherence by up to 30\\% while simultaneously improving relevance. The third part lifts equivalence to the dataset level. We introduce dataset declassification, a framework that replaces proprietary questions with semantically equivalent public alternatives, enabling dataset sharing without exposing sensitive content. Models trained on fully declassified data match baseline performance (WikiQA $\\Delta \\approx 0$, TrecQA $|\\Delta| \\leq 1.2$ points), and test set declassification preserves evaluation validity when high-quality mappings exist ($|\\Delta| \\leq 2$ on standard benchmarks), enabling the release of ``shadow benchmarks'' for evaluation integrity. We identify boundary conditions through experiments on adversarially-constructed benchmarks. Together, these contributions show that question semantic equivalence, systematically exploited at multiple scales, enables substantial improvements to QA system accuracy, consistency, and evaluation integrity.","abstract_html":"This thesis investigates the role of question understanding in Question Answering systems, developing methods that exploit question semantic equivalence at progressively larger scales: from individual question pairs, through equivalence clusters, to entire datasets. The first part addresses question retrieval at scale. We introduce QUADRo, a retrieval framework operating over millions of question-answer pairs, and the Question Ranking Corpus (QRC), a large-scale resource with answer-aware annotations and challenging hard negatives. We demonstrate that incorporating answers during retrieval substantially improves accuracy, as answers serve as a semantic bridge between questions that share little lexical overlap but seek the same information. To reduce annotation costs, we develop Question Ranking Pre-training (QRP), a self-supervised method that learns question equivalence patterns without labeled data, achieving significant improvements while reducing model variance by over 50\\%. The second part extends pairwise equivalence to question clusters. We analyze coherence in Large Language Models, finding that a substantial portion of question clusters exhibit incoherent behavior: models answer some phrasings correctly while failing on semantically equivalent alternatives. This reveals that understanding failures, not just knowledge gaps, limit LLM performance. We introduce Question-Augmented Generation (q-RAG), which supplements prompts with retrieved similar questions, improving accuracy by up to 9 percentage points and coherence by up to 28 points. We further show that q-RAG's benefits can be distilled into model parameters through Direct Preference Optimization (DPO) and Supervised Fine-Tuning, producing standalone models with improved coherence that surpass the inference-time approach. For retrieval systems, we apply clusters to train models for consistency: the Coherence Ranking Loss improves ranking coherence by up to 30\\% while simultaneously improving relevance. The third part lifts equivalence to the dataset level. We introduce dataset declassification, a framework that replaces proprietary questions with semantically equivalent public alternatives, enabling dataset sharing without exposing sensitive content. Models trained on fully declassified data match baseline performance (WikiQA $\\Delta \\approx 0$, TrecQA $|\\Delta| \\leq 1.2$ points), and test set declassification preserves evaluation validity when high-quality mappings exist ($|\\Delta| \\leq 2$ on standard benchmarks), enabling the release of ``shadow benchmarks'' for evaluation integrity. We identify boundary conditions through experiments on adversarially-constructed benchmarks. Together, these contributions show that question semantic equivalence, systematically exploited at multiple scales, enables substantial improvements to QA system accuracy, consistency, and evaluation integrity.","abstract_has_math":true,"creators":["Campese, Stefano"],"institution":"Università degli studi di Trento","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Moschitti, Alessandro"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-04-27","date_published":"2026-04-27","updated_at":"2026-07-24T05:04:28Z","subjects":["Question Retrieval, Semantic Equivalence, Database Question Answering, Question Ranking, Self-Supervised Pre-training, LLM Coherence, Retrieval-Augmented Generation, Retrieval Coherence, Dense Retrieval, Dataset Declassification, Privacy-Preserving NLP, Answer Sentence Selection"],"languages":["eng"],"rights":["info:eu-repo/semantics/openAccess","license:Tutti i diritti riservati (All rights reserved)","license uri:iris.PRI01"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/11572/483952","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Campese, Stefano","Moschitti, Alessandro"]},{"key":"dc:creator","label":"Author","values":["Campese, Stefano"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026-04-27"]},{"key":"dc:publisher","label":"Institution","values":["Università degli studi di Trento","place:TRENTO"]},{"key":"dc:relation","label":"Dc Relation","values":["firstpage:1","lastpage:203","numberofpages:203"]},{"key":"dc:type","label":"Dc Type","values":["info:eu-repo/semantics/doctoralThesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Question Retrieval, Semantic Equivalence, Database Question Answering, Question Ranking, Self-Supervised Pre-training, LLM Coherence, Retrieval-Augmented Generation, Retrieval Coherence, Dense Retrieval, Dataset Declassification, Privacy-Preserving NLP, Answer Sentence Selection"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["info:eu-repo/semantics/openAccess","license:Tutti i diritti riservati (All rights reserved)","license uri:iris.PRI01"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/11572/483952"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["This thesis investigates the role of question understanding in Question Answering systems, developing methods that exploit question semantic equivalence at progressively larger scales: from individual question pairs, through equivalence clusters, to entire datasets. The first part addresses question retrieval at scale. We introduce QUADRo, a retrieval framework operating over millions of question-answer pairs, and the Question Ranking Corpus (QRC), a large-scale resource with answer-aware annotations and challenging hard negatives. We demonstrate that incorporating answers during retrieval substantially improves accuracy, as answers serve as a semantic bridge between questions that share little lexical overlap but seek the same information. To reduce annotation costs, we develop Question Ranking Pre-training (QRP), a self-supervised method that learns question equivalence patterns without labeled data, achieving significant improvements while reducing model variance by over 50\\%. The second part extends pairwise equivalence to question clusters. We analyze coherence in Large Language Models, finding that a substantial portion of question clusters exhibit incoherent behavior: models answer some phrasings correctly while failing on semantically equivalent alternatives. This reveals that understanding failures, not just knowledge gaps, limit LLM performance. We introduce Question-Augmented Generation (q-RAG), which supplements prompts with retrieved similar questions, improving accuracy by up to 9 percentage points and coherence by up to 28 points. We further show that q-RAG's benefits can be distilled into model parameters through Direct Preference Optimization (DPO) and Supervised Fine-Tuning, producing standalone models with improved coherence that surpass the inference-time approach. For retrieval systems, we apply clusters to train models for consistency: the Coherence Ranking Loss improves ranking coherence by up to 30\\% while simultaneously improving relevance. The third part lifts equivalence to the dataset level. We introduce dataset declassification, a framework that replaces proprietary questions with semantically equivalent public alternatives, enabling dataset sharing without exposing sensitive content. Models trained on fully declassified data match baseline performance (WikiQA $\\Delta \\approx 0$, TrecQA $|\\Delta| \\leq 1.2$ points), and test set declassification preserves evaluation validity when high-quality mappings exist ($|\\Delta| \\leq 2$ on standard benchmarks), enabling the release of ``shadow benchmarks'' for evaluation integrity. We identify boundary conditions through experiments on adversarially-constructed benchmarks. Together, these contributions show that question semantic equivalence, systematically exploited at multiple scales, enables substantial improvements to QA system accuracy, consistency, and evaluation integrity."]},{"key":"dc:title","label":"Title","values":["Models and Application of Question Retrieval for Natural Language Processing"]}]}],"canonical_facts":{"dc:contributor":["Campese, Stefano","Moschitti, Alessandro"],"dc:creator":["Campese, Stefano"],"dc:date":["2026-04-27"],"dc:description":["This thesis investigates the role of question understanding in Question Answering systems, developing methods that exploit question semantic equivalence at progressively larger scales: from individual question pairs, through equivalence clusters, to entire datasets. The first part addresses question retrieval at scale. We introduce QUADRo, a retrieval framework operating over millions of question-answer pairs, and the Question Ranking Corpus (QRC), a large-scale resource with answer-aware annotations and challenging hard negatives. We demonstrate that incorporating answers during retrieval substantially improves accuracy, as answers serve as a semantic bridge between questions that share little lexical overlap but seek the same information. To reduce annotation costs, we develop Question Ranking Pre-training (QRP), a self-supervised method that learns question equivalence patterns without labeled data, achieving significant improvements while reducing model variance by over 50\\%. The second part extends pairwise equivalence to question clusters. We analyze coherence in Large Language Models, finding that a substantial portion of question clusters exhibit incoherent behavior: models answer some phrasings correctly while failing on semantically equivalent alternatives. This reveals that understanding failures, not just knowledge gaps, limit LLM performance. We introduce Question-Augmented Generation (q-RAG), which supplements prompts with retrieved similar questions, improving accuracy by up to 9 percentage points and coherence by up to 28 points. We further show that q-RAG's benefits can be distilled into model parameters through Direct Preference Optimization (DPO) and Supervised Fine-Tuning, producing standalone models with improved coherence that surpass the inference-time approach. For retrieval systems, we apply clusters to train models for consistency: the Coherence Ranking Loss improves ranking coherence by up to 30\\% while simultaneously improving relevance. The third part lifts equivalence to the dataset level. We introduce dataset declassification, a framework that replaces proprietary questions with semantically equivalent public alternatives, enabling dataset sharing without exposing sensitive content. Models trained on fully declassified data match baseline performance (WikiQA $\\Delta \\approx 0$, TrecQA $|\\Delta| \\leq 1.2$ points), and test set declassification preserves evaluation validity when high-quality mappings exist ($|\\Delta| \\leq 2$ on standard benchmarks), enabling the release of ``shadow benchmarks'' for evaluation integrity. We identify boundary conditions through experiments on adversarially-constructed benchmarks. Together, these contributions show that question semantic equivalence, systematically exploited at multiple scales, enables substantial improvements to QA system accuracy, consistency, and evaluation integrity."],"dc:identifier":["https://hdl.handle.net/11572/483952"],"dc:language":["eng"],"dc:publisher":["Università degli studi di Trento","place:TRENTO"],"dc:relation":["firstpage:1","lastpage:203","numberofpages:203"],"dc:rights":["info:eu-repo/semantics/openAccess","license:Tutti i diritti riservati (All rights reserved)","license uri:iris.PRI01"],"dc:subject":["Question Retrieval, Semantic Equivalence, Database Question Answering, Question Ranking, Self-Supervised Pre-training, LLM Coherence, Retrieval-Augmented Generation, Retrieval Coherence, Dense Retrieval, Dataset Declassification, Privacy-Preserving NLP, Answer Sentence Selection"],"dc:title":["Models and Application of Question Retrieval for Natural Language Processing"],"dc:type":["info:eu-repo/semantics/doctoralThesis"]},"updated_at":"2026-07-24T05:04:28Z"}