Long-context Non-factoid Question Answering in Indic Languages

Ponnurangam Kumaraguru; Rajiv Ratn Shah; Ritwik Mishra

REVIEW 3 major objections 4 minor 10 references

Long-context Non-factoid Question Answering in Indic Languages

T0 review · 3 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Shrinking the context with an answer-paragraph retriever improves non-factoid QA in four Indic languages.

desk verdict The headline gains (4% semantic, 47% token-level) are real but measured only on examples where the full-context baseline fit in memory; the paper is a solid, transparent empirical study whose central claim is plausible, but the headline numbers need a fuller comparison before being taken at face value. read the letter →

arxiv 2504.13615 v2 pith:QCXYBCYW submitted 2025-04-18 cs.CL

Ritwik Mishra , Rajiv Ratn Shah , Ponnurangam Kumaraguru This is my paper · ORCID

classification cs.CL

keywords non-factoidquestionansweringlong-contextQAanswerparagraphselectionIndiclanguagescontextshorteningmultilingualretrievallow-resourceNLPexplainable

verification ladder T0 review T1 audit T2 compute T3 formal

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 paper sets out to establish that the main obstacle to non-factoid question answering in low-resource languages is the length of the context, not the generative ability of the LLM. It claims that a trained Answer Paragraph Selection (APS) model, which scores each paragraph against the question and keeps the top five, yields better answers than feeding the full long context to the model. Across Gemma-2b, Gemma-7b and Llama-3.1-8b without fine-tuning, the shortened-context pipeline improves average semantic similarity by 4% and token-level overlap scores by 47%, while also reducing memory and latency. The same pipeline with fine-tuning gives about 2% improvement on both metric families. The paper also uses LIME and SHAP to claim that the APS model's confident decisions rest on nearly all tokens in the chosen paragraph, whereas low-confidence decisions rest on only a few.

What carries the argument

The central object is the Answer Paragraph Selection (APS) model, a cross-encoder that reads a question and a paragraph together and outputs a relevance score. The paper uses a fine-tuned XLM-R based checkpoint to score every paragraph in the long context, ranks them, and keeps the top five as the LLM's context. Around this core sit four retrieval variants: APS alone (A1), verbalized OpenIE triples re-ranked by APS (A2), coreference-linked paragraphs re-ranked by APS (A3), and coreference clusters of verbalized triples re-ranked by APS (A4). The evaluation side is carried by STS-MuTe, a new semantic metric that averages cosine similarities from three multilingual embedding models with BERTScore, and by standard ROUGE variants for token-level comparison.

What would settle it

Run the baseline B and the A1 pipeline on every one of the 1,100 sampled questions on hardware that avoids out-of-memory errors and compare average STS-MuTe and ROUGE; if the gains shrink, reverse, or vanish on the previously excluded examples, the paper's central claim fails. A second check is to compare B and A1 on the translated Natural-QA subset with contexts above 8,000 tokens, where the paper reports only topline results for A1.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that a multilingual paragraph-scoring model can serve as a question-specific context compressor for non-factoid QA in Hindi, Tamil, Telugu, and Urdu. The A1 pipeline, built from a fine-tuned XLM-R based APS model, outperforms the full-context baseline B on the common test examples for all three LLMs, with a 4% average gain in STS-MuTe (a composite of BERTScore, USE, LaBSE, and LASER similarities) and a 47% average gain in ROUGE scores when the LLMs are not fine-tuned; fine-tuning shrinks the gain to about 2%. The token-level gain is large but the semantic gain is modest, and the paper attributes the semantic gain to the retriever's ability to find the paragraphs that actually answer the question. In an ablation, fine-tuned APS beats both BM25 and an embedding-based vectorstore retriever, and 75% of its top-k paragraphs differ from BM25's, supporting the claim that the selection is semantic rather than lexical.

Load-bearing premise

The comparison that carries the paper's headline numbers was run only on the subset of a 1,100-question sample where the full-context baseline did not run out of memory, so the load-bearing premise is that those completed examples represent the full test distribution.

Editorial extensions

If this is right

If the gain is real, long-context QA systems for low-resource languages can be improved and made cheaper at the same time by inserting a paragraph-selection retriever before generation.
The result implies that semantic metrics and token-level metrics move together in this setting, so the improvement is not just a paraphrase effect.
The explainability finding suggests the APS score can be used as a confidence signal: low-scoring paragraphs are likely to be irrelevant, so the retriever can safely discard them.
Because verbalizing OIE triples did not help, triple extraction is probably not a useful context-shortening step for non-factoid Indic QA, despite its success in knowledge-grounded English tasks.
The 75% disagreement with BM25 indicates that lexical retrieval alone is insufficient for non-factoid Indic QA, supporting the need for learned semantic retrieval.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: the A1 pipeline should transfer to other low-resource languages with a cross-lingual encoder, because it depends on no OIE or coreference tools; the paper only tests Hindi, Tamil, Telugu, and Urdu.
Editorial inference: the true improvement over the baseline is probably closer to the 4% semantic gain than the 47% token-level gain, because token-level metrics penalize legitimate paraphrase in LLM answers.
Editorial inference: a direct stress test is to run the baseline on the OOM examples with larger GPUs; if the baseline wins there, the central comparison is an artifact of memory limits rather than a property of long contexts.
Editorial inference: using the APS score as a reward model for preference tuning, which the paper names as future work, is a testable next step: it would show whether retriever confidence aligns with downstream answer quality.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Let me give you the short version first. This is a legitimate empirical paper about context-shortening for non-factoid QA in Hindi, Tamil, Telugu, and Urdu. The best part is the retriever-based pipeline (APS) is tested against a full-context baseline, with ablations against BM25 and dense retrieval, compute footprint reporting, and a genuinely candid limitations section. The STS-MuTe metric is a reasonable addition, and the released checkpoints for gemma-2b/7b and llama-3.1 are reusable. If you work on multilingual QA, there is real value here.

The central claim—that APS-based shortening beats feeding the whole context—is plausible and the direction is consistent across three LLMs. But the headline numbers are softer than they look. The baseline B OOMs on many test examples, and the A1-vs-B comparison is restricted to the examples where B fit in memory. Those are exactly the shorter contexts. The paper gives the average token counts for the OOM examples (1238, 1189, 1069 per model) but not B's performance on them, so the reported 4% and 47% gains are conditional on hardware. On top of that, there are no error bars or significance tests, and some per-model gains are tiny (gemma-7b STS-MuTe goes from 0.60 to 0.61). The contamination check is a point in their favor, but it only reports how much B drops (13%, 2.5%, 22%) after excluding pretraining-era data; they don't give the A1-vs-B comparison on that filtered set, so we can't rule out that part of the relative gain is contamination-driven.

None of this kills the paper. The evaluation is transparent about what was done, and the limitations section reads like the authors know where the weak points are. The heavy reliance on their own prior tools (APS, coref, MuNfQuAD) is a dependence rather than a flaw, since those resources are published and the held-out split is proper. But the headline claim should be treated as provisional.

Send it to review. The right reviewers will ask for the full-non-OOM comparison, significance tests, and the contamination-filtered A1-vs-B table. This is the kind of paper that gets noticeably better with one solid revision round, and the released artifacts make it worth the referee time.

Referee Report

3 major / 4 minor

Summary. This paper studies context-shortening for non-factoid question answering in Hindi, Tamil, Telugu, and Urdu. It builds on the authors' MuNfQuAD dataset and compares four retrieval-based pipelines (APS, OIE+APS, coref+APS, OIE+coref+APS) against a baseline that feeds the full long context to Gemma-2B, Gemma-7B, and Llama-3.1-8B, in both zero-shot and LoRA fine-tuned settings. The authors report that the APS-only pipeline (A1) improves the average semantic score (STS-MuTe) by about 4% and ROUGE by about 47% over the full-context baseline, while also reducing GPU memory and inference time. They also introduce a new semantic similarity metric, report an LLM-as-a-judge evaluation, and use LIME and SHAP to analyze the APS model's rationales. Code, checkpoints, and resources are released. The headline comparison is limited to a 1100-question subset of the 6k+ test set and, within that subset, to the examples on which the full-context baseline did not run out of memory.

Significance. If the reported improvement is robust, this is a useful practical contribution: it shows that a fine-tuned paragraph selector can reduce long contexts for non-factoid QA in low-resource languages while matching or exceeding full-context LLM performance at substantially lower computational cost. The paper is transparent in several important ways: the APS training data is excluded from evaluation (a proper held-out split), ablations against BM25 and dense retrieval are included, computational footprints are reported, and a contamination analysis is attempted. These strengths make the central claim worth taking seriously. The main weakness is that the central A1-vs-B comparison is restricted to the subset where the baseline did not OOM, and no uncertainty quantification is provided; the headline numbers are therefore conditional on the hardware-restricted subset rather than unconditional claims about the test distribution.

major comments (3)

[§5, Tables 1 and 3] The central claim that A1 outperforms B is evaluated only on the test examples common to both, because B OOMs on many long examples. The paper states: 'The resource intensive nature of B led to Out of Memory (OOM) errors on many test examples' and therefore B and A1 were compared on the 'test examples common to both' among the 1100. Since the OOM examples have average context lengths of 1238, 1189, and 1069 tokens by model (reported in §5), the comparison excludes exactly the longest contexts, where context-shortening is arguably most relevant. To support the abstract's 4% and 47% claims, the authors should either run B on the full 1100-item subset (e.g., with chunking, a longer-context model, or reduced batch size) or provide evidence that the common subset is representative in terms of context length, language, and question category distributions. As written, the reported gains are estimates conditional on the non-OOM subset.
[§8, contamination analysis] The contamination filter shows that B's zero-shot scores drop by 13%, 2.5%, and 22% when test data that were publicly available before each model release are excluded, but the paper does not report A1's filtered scores or the A1-vs-B comparison on the filtered examples. This matters because the relative gain could change if contamination inflates B and A1 differently. The filtered sets are small (78, 197, and 18 examples), but the manuscript should still report the filtered comparison for at least STS-MuTe and ROUGE, with the small sample sizes acknowledged, so that readers can see whether the relative improvement survives the filter.
[§5, Table 1] Several per-model differences are small (e.g., gemma-7b-it STS-MuTe: B 0.60 vs A1 0.61; BERTScore 0.66 vs 0.69 for gemma-7b-it), and no confidence intervals, bootstrap estimates, or significance tests are provided. Since the headline is an average over three models and four languages, the 4% semantic improvement may be within run-to-run noise. Please report per-language, per-model results with bootstrap confidence intervals or pairwise significance tests on the common subset, or explicitly state that the observed differences are not tested for significance.

minor comments (4)

[§5, LLM-as-a-judge] The LLM-as-a-judge protocol should state whether the order of option1 and option2 is randomized across examples, and should report agreement with human judgments if available; the current text only gives preference percentages.
[§3 and §4, data splits] Please clarify the relation between the 'over 40K' question-answer pairs after excluding the APS training data, the 'more than 29K' training questions, and the 'more than 6k' test questions; these numbers do not obviously sum, and a precise split description would help reproducibility.
[§2 and References] There are minor typos (e.g., 'The A1 approach does not only shows better performance' in §5) and duplicate bibliography entries (Kolluru et al. 2022a/2022b and Ribeiro et al. 2016a/2016b); please clean these before publication.
[§6, Figure 6] The text describing Figure 6 gives a concrete example of 7 out of 10 tokens above a threshold, but the figure caption does not explain the bucketing clearly; please make the caption self-contained and state which languages and how many examples the 8.5k question-paragraph rationale computation covers.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central A1-vs-B comparison is an empirical, held-out evaluation that could have favored either pipeline.

full rationale

The paper's headline claim is that context-shortening with an Answer Paragraph Selection retriever (A1) improves LLM non-factoid QA over a full-context baseline (B) on MuNfQuAD. This is an empirical comparison on held-out test questions: B is fed the whole context, A1 receives the top-5 APS paragraphs, and both are scored with ROUGE and the paper's own STS-MuTe. The outcome is not forced by construction; B could have won, and indeed Table 1 shows B at or above A1 on some individual metrics (e.g., gemma-7b-it USE/LASER). The fine-tuned APS model, MuNfQuAD, Transmucores, and IndIE come from the authors' prior work, but they are used as components rather than as premises that entail the conclusion; the paper even ablates A1 against LangChain and BM25 retrievers, providing internal evidence for the APS retriever's contribution. STS-MuTe is an evaluation metric (a mean of existing embedding similarities and BERTScore), not a predicted quantity that is defined in terms of the systems' outputs, so using it to measure improvement is not a definitional reduction. The restriction of A1-vs-B to a 1100-question subset and to test examples where B did not OOM is a coverage limitation that affects generalizability and statistical strength, but it is not a circularity. Consequently, no self-definitional, fitted-input, or self-citation-chain reduction is present.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The paper's central claim rests on an empirical pipeline, not a derivation, so no fitted constants are introduced. The free parameters are hand-chosen hyperparameters (top-k, LoRA rank/alpha, temperature, context threshold). The axioms are domain assumptions about dataset quality, tool transfer, and evaluation reliability, including the use of GPT-4o as a judge and NLLB translations.

free parameters (5)

APS top-k = 5
Number of paragraphs selected by the APS retriever in approach A1; set by hand without ablation. The central claim depends on this choice.
OIE triple top-k = 10
Number of verbalized triples selected in approaches A2 and A4.
LoRA rank and alpha = 32
Hyperparameters for parameter-efficient fine-tuning of LLMs, stated in Section 4.1.
Sampling temperature = 0.001
Low temperature used for LLM inference to ensure reproducibility.
Long-context threshold = 512 tokens
Adopted definition of long context, justified by BERT/RoBERTa input limits and the dataset average length.

assumptions (5)

domain assumption MuNfQuAD silver answers are reliable enough to evaluate generated answers.
The paper uses silver answers as ground truth for fine-tuning and evaluation; if the silver answers are noisy, the metric scores may not reflect true quality. Section 3.
domain assumption The fine-tuned APS model from prior work generalizes to held-out MuNfQuAD splits.
The retriever is trained on a split of the same dataset used for evaluation; the assumption is that this transfer is valid. Section 3.
ad hoc to paper Gen2OIE checkpoints transfer zero-shot from Hindi/Telugu to Urdu/Tamil.
Stated in Section 8: Gen2OIE was assessed on Hindi and Telugu only, and used zero-shot for Urdu and Tamil based on linguistic relatedness.
domain assumption GPT-4o as a judge provides reliable qualitative evaluation.
LLM-as-a-judge is used without human validation; prior work shows judges can be biased. Section 5.
ad hoc to paper NLLB 1.3B translations of NaturalQA questions are adequate for Indic evaluation.
A subset of NaturalQA was automatically translated to six Indic languages; translation quality is not manually checked. Section 5.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Long-context Non-factoid Question Answering in Indic Languages." pith.science (2026). https://pith.science/paper/QCXYBCYW

@misc{pith2026250413615,
  author       = {Pith},
  title        = {Pith review of: Long-context Non-factoid Question Answering in Indic Languages},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QCXYBCYW}},
  note         = {Machine review of arXiv:2504.13615}
}

read the original abstract

Question Answering (QA) tasks, which involve extracting answers from a given context, are relatively straightforward for modern Large Language Models (LLMs) when the context is short. However, long contexts pose challenges due to the quadratic complexity of the self-attention mechanism. This challenge is compounded in Indic languages, which are often low-resource. This study explores context-shortening techniques, including Open Information Extraction (OIE), coreference resolution, Answer Paragraph Selection (APS), and their combinations, to improve QA performance. Compared to the baseline of unshortened (long) contexts, our experiments on four Indic languages (Hindi, Tamil, Telugu, and Urdu) demonstrate that context-shortening techniques yield an average improvement of 4\% in semantic scores and 47\% in token-level scores when evaluated on three popular LLMs without fine-tuning. Furthermore, with fine-tuning, we achieve an average increase of 2\% in both semantic and token-level scores. Additionally, context-shortening reduces computational overhead. Explainability techniques like LIME and SHAP reveal that when the APS model confidently identifies the paragraph containing the answer, nearly all tokens within the selected text receive high relevance scores. However, the study also highlights the limitations of LLM-based QA systems in addressing non-factoid questions, particularly those requiring reasoning or debate. Moreover, verbalizing OIE-generated triples does not enhance system performance. These findings emphasize the potential of context-shortening techniques to improve the efficiency and effectiveness of LLM-based QA systems, especially for low-resource languages. The source code and resources are available at https://github.com/ritwikmishra/IndicGenQA.

Figures

Figures reproduced from arXiv: 2504.13615 by the authors.

**Figure 2.** Illustration of the four approaches used to con [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Prompt used to finetune LLMs in the QA pipeline. The {context} will be a long-context for QA pipeline with B whereas it will be a short-context for QA pipeline with A1-A4. During the inference stage, {answer} was omitted from the prompt. is used in the QA pipeline. We used 2 billion and 7 billion parameter model of instruction finetuned Gemma (Team et al., 2024b). We also used 8 billion parameter model of instructio… view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Prompt used to query ChatGPT model to perform qualitative analysis. experiments on a subset of test set showed that base gpt-4o-mini from (Hurst et al., 2024) is able to outperform all QAf inetuned on all the quantitative evaluation metrics. We used the prompt shown in…

**Figure 5.** Figure 5: Rationales generated by LIME. The figures on the left correspond to the Hindi text from the test set, while those on the right display the English-translated text. Tokens with high relevance values (depicted in brighter red) indicate their significant contribution to t…

**Figure 6.** Figure 6: Consider an example to interpret the figure above: an APS model predicts a logit of 0.35 for a text containing 10 tokens, where 7 tokens exhibit a relevance value exceeding a specified threshold (e.g., 0.5). In this case, the logit bucket would fall within the range of…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

10 extracted references · 4 canonical work pages

[8]

Multilingual Non-Factoid Question Answering with Answer Paragraph Selection

Indie: A multilingual open information ex- traction tool for indic languages. InFindings of the Association for Computational Linguistics: IJCNLP- AACL 2023 (Findings), pages 312–326. Ritwik Mishra, Sreeram V ennam, Rajiv Ratn Shah, and Ponnurangam Kumaraguru. 2024b. Multilingual non-factoid question answering with silver answers. arXiv preprint arXiv:240...

work page Pith review arXiv 2023
[10]

In2024 Conference on AI, Sci- ence, Engineering, and Technology (AIxSET), pages 276–289

Llm-based localization in the context of low- resource languages. In2024 Conference on AI, Sci- ence, Engineering, and Technology (AIxSET), pages 276–289. IEEE. David Uthus, Santiago Ontanon, Joshua Ainslie, and Mandy Guo. 2023. mlongt5: A multilingual and efﬁcient text-to-text transformer for longer sequences. InThe 2023 Conference on Empirical Methods i...

work page 2023
[2004]

Kabir Ahuja, Harshita Diddee, Rishav Hada, Milli- cent Ochieng, Krithika Ramesh, Prachi Jain, Ak- shay Nambi, Tanuja Ganu, Sameer Segal, Mohamed Ahmed, and 1 others

Learning toﬁnd answers to questions on the web.ACM Transactions on Internet Technology (TOIT), 4(2):129–162. Kabir Ahuja, Harshita Diddee, Rishav Hada, Milli- cent Ochieng, Krithika Ramesh, Prachi Jain, Ak- shay Nambi, Tanuja Ganu, Sameer Segal, Mohamed Ahmed, and 1 others. 2023. Mega: Multilingual evaluation of generative ai. InProceedings of the 2023 Co...

work page 2023
[2014]

InProceedings of the 20th ACM SIGKDD international conference on Knowl- edge discovery and data mining, pages 1156–1165

Open question answering over curated and ex- tracted knowledge bases. InProceedings of the 20th ACM SIGKDD international conference on Knowl- edge discovery and data mining, pages 1156–1165. Haytham M Fayek and Justin Johnson. 2020. Temporal reasoning via audio question answering.IEEE/ACM Transactions on Audio, Speech, and Language Pro- cessing, 28:2283–2...

work page 2020
[2015]

InProceedings of the F orum for Information Retrieval Evaluation, FIRE ’14, pages 48–53, New Y ork, NY , USA

Iiit-h system submission forﬁre2014 shared task on transliterated search . InProceedings of the F orum for Information Retrieval Evaluation, FIRE ’14, pages 48–53, New Y ork, NY , USA. ACM. V aleriia Bolotova, Vladislav Blinov, Falk Scholer, W Bruce Croft, and Mark Sanderson. 2022. A non- factoid question-answering taxonomy. InProceed- ings of the 45th In...

arXiv 2022
[2020]

In Proceedings of the AAAI conference on articial in- telligence, volume 34, pages 7780–7788

Tanda: Transfer and adapt pre-trained trans- former models for answer sentence selection. In Proceedings of the AAAI conference on articial in- telligence, volume 34, pages 7780–7788. Michael Glass, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Ankita Naik, Pengshan Cai, and Alﬁo Gliozzo. 2022. Re2G: Retrieve, rerank, generate . InProceedings of the 202...

arXiv 2022
[2021]

cloze procedure

Nlquad: A non-factoid long question answer- ing data set. InProceedings of the 16th Conference of the European Chapter of the Association for Com- putational Linguistics: Main V olume, pages 1245– 1255. Radu Soricut and Eric Brill. 2006. Automatic ques- tion answering using the web: Beyond the factoid. Information Retrieval, 9:191–206. Mihai Surdeanu, Mas...

arXiv 2006
[2023]

InProceedings of the Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacic Region, pages 11–20

Retrieving supporting evidence for generative question answering. InProceedings of the Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacic Region, pages 11–20. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radfor...

arXiv 2024

Show all 10 references

[2024]

InProceedings of the 18th Conference of the Euro- pean Chapter of the Association for Computational Linguistics: Student Research Workshop, pages 358– 368

Dynamic task-oriented dialogue: A compara- tive study of llama-2 and bert in slot value generation. InProceedings of the 18th Conference of the Euro- pean Chapter of the Association for Computational Linguistics: Student Research Workshop, pages 358– 368. Tianle Li, Ge Zhang, ...

2024 arXiv
[2025]

Avinash Anand, Raj Jaiswal, Abhishek Dharmadhikari, Atharva Marathe, Harsh Popat, Harshil Mital, Ash- win R Nair, Kritarth Prasad, Sidharth Kumar, Astha V erma, and 1 others

Evaluation of open and closed-source llms for low-resource language with zero-shot, few-shot, and chain-of-thought prompting.Natural Language Processing Journal, page 100124. Avinash Anand, Raj Jaiswal, Abhishek Dharmadhikari, Atharva Marathe, Harsh Popat, Harshil Mital, Ash- ...

2019 arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [8]

Multilingual Non-Factoid Question Answering with Answer Paragraph Selection

Indie: A multilingual open information ex- traction tool for indic languages. InFindings of the Association for Computational Linguistics: IJCNLP- AACL 2023 (Findings), pages 312–326. Ritwik Mishra, Sreeram V ennam, Rajiv Ratn Shah, and Ponnurangam Kumaraguru. 2024b. Multilingual non-factoid question answering with silver answers. arXiv preprint arXiv:240...

work page Pith review arXiv 2023

[2] [10]

In2024 Conference on AI, Sci- ence, Engineering, and Technology (AIxSET), pages 276–289

Llm-based localization in the context of low- resource languages. In2024 Conference on AI, Sci- ence, Engineering, and Technology (AIxSET), pages 276–289. IEEE. David Uthus, Santiago Ontanon, Joshua Ainslie, and Mandy Guo. 2023. mlongt5: A multilingual and efﬁcient text-to-text transformer for longer sequences. InThe 2023 Conference on Empirical Methods i...

work page 2023

[3] [2004]

Kabir Ahuja, Harshita Diddee, Rishav Hada, Milli- cent Ochieng, Krithika Ramesh, Prachi Jain, Ak- shay Nambi, Tanuja Ganu, Sameer Segal, Mohamed Ahmed, and 1 others

Learning toﬁnd answers to questions on the web.ACM Transactions on Internet Technology (TOIT), 4(2):129–162. Kabir Ahuja, Harshita Diddee, Rishav Hada, Milli- cent Ochieng, Krithika Ramesh, Prachi Jain, Ak- shay Nambi, Tanuja Ganu, Sameer Segal, Mohamed Ahmed, and 1 others. 2023. Mega: Multilingual evaluation of generative ai. InProceedings of the 2023 Co...

work page 2023

[4] [2014]

InProceedings of the 20th ACM SIGKDD international conference on Knowl- edge discovery and data mining, pages 1156–1165

Open question answering over curated and ex- tracted knowledge bases. InProceedings of the 20th ACM SIGKDD international conference on Knowl- edge discovery and data mining, pages 1156–1165. Haytham M Fayek and Justin Johnson. 2020. Temporal reasoning via audio question answering.IEEE/ACM Transactions on Audio, Speech, and Language Pro- cessing, 28:2283–2...

work page 2020

[5] [2015]

InProceedings of the F orum for Information Retrieval Evaluation, FIRE ’14, pages 48–53, New Y ork, NY , USA

Iiit-h system submission forﬁre2014 shared task on transliterated search . InProceedings of the F orum for Information Retrieval Evaluation, FIRE ’14, pages 48–53, New Y ork, NY , USA. ACM. V aleriia Bolotova, Vladislav Blinov, Falk Scholer, W Bruce Croft, and Mark Sanderson. 2022. A non- factoid question-answering taxonomy. InProceed- ings of the 45th In...

arXiv 2022

[6] [2020]

In Proceedings of the AAAI conference on articial in- telligence, volume 34, pages 7780–7788

Tanda: Transfer and adapt pre-trained trans- former models for answer sentence selection. In Proceedings of the AAAI conference on articial in- telligence, volume 34, pages 7780–7788. Michael Glass, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Ankita Naik, Pengshan Cai, and Alﬁo Gliozzo. 2022. Re2G: Retrieve, rerank, generate . InProceedings of the 202...

arXiv 2022

[7] [2021]

cloze procedure

Nlquad: A non-factoid long question answer- ing data set. InProceedings of the 16th Conference of the European Chapter of the Association for Com- putational Linguistics: Main V olume, pages 1245– 1255. Radu Soricut and Eric Brill. 2006. Automatic ques- tion answering using the web: Beyond the factoid. Information Retrieval, 9:191–206. Mihai Surdeanu, Mas...

arXiv 2006

[8] [2023]

InProceedings of the Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacic Region, pages 11–20

Retrieving supporting evidence for generative question answering. InProceedings of the Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacic Region, pages 11–20. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radfor...

arXiv 2024