AI Paper Atlas

Research map / Benchmarks / SearchQA

SearchQA leaderboard

11 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on SearchQA. Best F1: 84.4 (SAIN SpanBERT, AAAI 2021).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

SearchQA · F1

11 papers, each with its best reported number for its own method; higher is better.

#F1 ↑MethodPaperTopic
184.4SAIN SpanBERT
SpanBERT
Semantics-Aware Inferential Network for Natural Language Understanding
AAAI 2021 · Table 2
Language, safety and interpretability
280.32Block-Skim Training
BERT-base
Block-Skim: Efficient Question Answering for Transformer
AAAI 2022 · Table: validation F1 and FLOPs
Attention and transformers
368.8MG-MRCMulti-Task Learning with Generative Adversarial Training for Multi-Passage Machine Reading Comprehension
AAAI 2020 · Table 2
GANs
468.7HAS-QA (MAX Ans. Span)HAS-QA: Hierarchical Answer Spans Model for Open-Domain Question Answering
AAAI 2019 · Table 2
Visual question answering
566.7WKLM + Ranking score
BERT-base
Pretrained Encyclopedia: Weakly Supervised Knowledge-Pretrained Language Model
ICLR 2020 · Table: Open-domain QA Results
CLIP and zero-shot
663.9DecaProp + CoDACompositional De-Attention Networks
NeurIPS 2019 · Table 1
Attention and transformers
763.2Full Re-Ranker
R3
Evidence Aggregation for Answer Re-Ranking in Open-Domain Question Answering
ICLR 2018 · Table 1
Recognition and long-tail
861.61Multi-step reasoner (DrQA)
DrQA
Multi-step Retriever-Reader Interaction for Scalable Open-domain Question Answering
ICLR 2019 · Table: Performance on test set
Retrieval and knowledge
953.4FHEFocused Hierarchical RNNs for Conditional Sequence Processing
ICML 2018 · Table, SearchQA results
Attention and transformers
1045.6AQA CNN
BiDAF
Ask the Right Questions: Active Question Reformulation with Reinforcement Learning
ICLR 2018 · Results table
Retrieval and knowledge
1 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.