Papers › Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering

Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering

25 Jul 2017CVPR 2018 6arXiv:1707.07998archive 2025-07-28

Peter Anderson, Xiaodong He, Chris Buehler, Damien Teney, Mark Johnson, Stephen Gould, Lei Zhang

Top-down visual attention mechanisms have been used extensively in image captioning and visual question answering (VQA) to enable deeper image understanding through fine-grained analysis and even multiple steps of reasoning. In this work, we propose a combined bottom-up and top-down attention mechanism that enables attention to be calculated at the level of objects and other salient image regions. This is the natural basis for attention to be considered. Within our approach, the bottom-up mechanism (based on Faster R-CNN) proposes image regions, each with an associated feature vector, while the top-down mechanism determines feature weightings. Applying this approach to image captioning, our results on the MSCOCO test server establish a new state-of-the-art for the task, achieving CIDEr / SPICE / BLEU-4 scores of 117.9, 21.5 and 36.9, respectively. Demonstrating the broad applicability of the method, applying the same approach to VQA we obtain first place in the 2017 VQA Challenge.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="1707.07998")

Code

Syntology Ran 9 of 9 code samples harvested from 3 repositories linked to this paper; 0 have no recorded run. Of those that ran: 1 ran · violated contract; 8 ran · our draft was wrong.

By repository: community (archive-listed): 8 samples from 3 repositories, 8 ran; 1 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

65 repositories listed; official and paper-mentioned ones first.

peteanderson80/bottom-up-attention officialmentioned on GitHub report
BierOne/VQA-AttReg mentioned on GitHubpytorch report
BigRedT/info-ground mentioned on GitHubpytorchNOASSERTION report
Cloud-CV/visual-chatbot mentioned on GitHubpytorch report
Dlut-lab-zmn/Image-Captioning-Attack mentioned on GitHubpytorch report
FJSam/SelfCritical_ImageCaptioning mentioned on GitHubpytorch report
FengSuSky/CCB-VQA mentioned on GitHubpytorch report
JHKim-snu/GVCCI mentioned on GitHubpytorch report
JHKim-snu/PGA mentioned on GitHubpytorch report
SJTU-VLG/ImageCaption-LBPF-qy mentioned on GitHubpytorch report
SatyamGaba/visual_question_answering mentioned on GitHubpytorch report
SatyamGaba/vqa mentioned on GitHubpytorch report
SpencerWhitehead/vqap2 mentioned on GitHubpytorch report
ThanThoai/Visual-Question-Answering_Vietnamese mentioned on GitHubpytorchApache-2.0 report
Wentong-DST/up-down-captioner mentioned on GitHubcaffe2 report
ZephyrZhuQi/ssbaseline mentioned on GitHubpytorchNOASSERTION report
Zhendong-Wang/arsm_image_captioning mentioned on GitHubpytorch report
airsplay/lxmert mentioned on GitHubpytorchMIT report
alab-nii/eval_vl_glue mentioned on GitHubjaxApache-2.0 report
allenai/pythia mentioned on GitHubpytorchNOASSERTION report
brandontrabucco/up_down_cell mentioned on GitHubtf report
chrisc36/bottom-up-attention-vqa mentioned on GitHubpytorch report
facebookresearch/mmf mentioned on GitHubpytorchNOASSERTION report
feifengwhu/question_attention mentioned on GitHubpytorch report
fuqianya/bottom-up-attention-paddle mentioned on GitHubpaddle report
hengyuan-hu/bottom-up-attention-vqa mentioned on GitHubpytorch report
hjjpku/dynamic_graph_caption mentioned on GitHubpytorch report
jackroos/pythia mentioned on GitHubpytorchNOASSERTION report
kangkang59812/GraphCaption mentioned on GitHubpytorch report
king-zark/self_critical mentioned on GitHubpytorch report
lauradhatt/Interesting-Reads mentioned on GitHub report
leaplabthu/pseudo-q mentioned on GitHubpytorchApache-2.0 report
longjj/Caffe-SGDR mentioned on GitHub report
marioskatsak/mapert mentioned on GitHubpytorch report
mazm13/Image-to-Tree.pytorch mentioned on GitHubpytorch report
meiqiguo/iccv2021-atypicalitydetection mentioned on GitHubpytorch report
mokhalid-dev/Attention-based-VQA-model mentioned on GitHubpytorch report
nocaps-org/updown-baseline mentioned on GitHubpytorchMIT report
peteanderson80/Up-Down-Captioner mentioned on GitHubcaffe2 report
rishavbb/bottom_up_vqa mentioned on GitHubtf report
ronghanghu/pythia mentioned on GitHubpytorchNOASSERTION report
ruotianluo/DiscCaptioning mentioned on GitHubpytorch report
ruotianluo/GoogleConceptualCaptioning mentioned on GitHubpytorch report
ruotianluo/ImageCaptioning.pytorch mentioned on GitHubpytorch report
ruotianluo/Transformer_Captioning mentioned on GitHubpytorch report
ruotianluo/neuraltalk2.pytorch mentioned on GitHubpytorch report
ruotianluo/self-critical.pytorch mentioned on GitHubpytorch report
sgondala/GoogleConceptualCaptioning mentioned on GitHubpytorch report
thilinicooray/Bottom-up-vqa mentioned on GitHubpytorch report
ukyh/switch_disc_caption mentioned on GitHubpytorch report
windweller/PragmaticVQA mentioned on GitHub report
xiaobai714/image_caption mentioned on GitHubpytorch report
xuewyang/Fashion_Captioning mentioned on GitHubpytorchNOASSERTION report
yangdsh/VQA-BUTD-demo mentioned on GitHub report
yanxinyan1/yxy mentioned on GitHubpytorch report
yifanh1/gqa_pytorch mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

9 samples harvested; 9 ran; 0 honoured the contract we drafted; 0 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · violated contract
8ran · our draft was wrong

Licence: 6 of the 9 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 repositories linked to this paper, official or community; each sample names its own and says which. Some samples are identical code Syntology first harvested from another repository; for those, this paper's copy is not located and its licence is not recorded. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

elementwise_logsumexp chrisc36/bottom-up-attention-vqa/vqa_debias_loss_functions.py community (archive-listed) ran · our draft was wrong fingerprinted GPL-3.0 (copyleft) · pointer only · 4bd23f3562cfbd1c · report
attention ruotianluo/Transformer_Captioning/models/TransformerModel.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 20bb0ff4a2d77a03 · report
convert_sigmoid_logits_to_binary_logprobs chrisc36/bottom-up-attention-vqa/vqa_debias_loss_functions.py community (archive-listed) ran · our draft was wrong fingerprinted GPL-3.0 (copyleft) · pointer only · 0e2c50fe59debfb7 · report
pack_wrapper kangkang59812/GraphCaption/models/AttModel.py community (archive-listed) ran · our draft was wrong MIT (permissive) · d2379d710eedc4ae · report
pad_unsort_packed_sequence kangkang59812/GraphCaption/models/AttModel.py community (archive-listed) ran · our draft was wrong MIT (permissive) · bfac58a04b6835f5 · report
renormalize_binary_logits chrisc36/bottom-up-attention-vqa/vqa_debias_loss_functions.py community (archive-listed) ran · our draft was wrong fingerprinted GPL-3.0 (copyleft) · pointer only · 21690a7e909f0f41 · report
sort_pack_padded_sequence kangkang59812/GraphCaption/models/AttModel.py community (archive-listed) ran · our draft was wrong fingerprinted MIT (permissive) · e56d2f9cbd9cb8b1 · report
subsequent_mask ruotianluo/Transformer_Captioning/models/TransformerModel.py community (archive-listed) ran · violated contract no licence file found · pointer only · dd03765c7da9ca51 · report
clones identical code first harvested elsewhere ran · our draft was wrong licence of this copy not recorded · a3722169bbc81569 · report

Tasks

Image CaptioningVisual Question AnsweringVisual Question Answering (VQA)

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Visual Question Answering (VQA) GQA Test2019 BottomUp Accuracy 49.74 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) GQA Test2019 BottomUp Binary 66.64 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) GQA Test2019 BottomUp Consistency 78.71 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) GQA Test2019 BottomUp Distribution 5.98 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) GQA Test2019 BottomUp Open 34.83 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) GQA Test2019 BottomUp Plausibility 84.57 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) GQA Test2019 BottomUp Validity 96.18 #107 of 127 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-std Up-Down overall 70.34 #28 of 38 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections