AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

The Abduction of Sherlock Holmes: A Dataset for Visual Abductive Reasoning

Jack Hessel, Jena D. Hwang, Jae Sung Park, Rowan Zellers, Chandra Bhagavatula, Anna Rohrbach, Kate Saenko, Yejin Choi

ECCV 2022 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 37 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add multitask clue learningCLIP RN50x64 · Sherlock retrievalP@1 (im->txt): 31.8 -> 33.4Improved
add multitask clue learningCLIP RN50x64 · Sherlock comparison (val)human accuracy: 25.1 -> 26.6Improved
scale CLIP backbone from ViT-B/16 to RN50x64CLIP · Sherlock comparison (val)human accuracy: 20.1 -> 25.1Improved
use CLIP instead of UNITER BaseCLIP ViT-B/16 · Sherlock retrievalP@1 (im->txt): 19.8 -> 30.6Improved
pass the region as input (vs not passing region) (ablation)CLIP ViT-B/16 · Sherlock retrievalP@1 (im->txt): 18 -> 31Improved
keep surrounding context rather than only the region (No Context ablation) (ablation)CLIP ViT-B/16 · Sherlockretrieval and comparisonImproved
keep region content rather than only context (Only Context ablation) (ablation)CLIP ViT-B/16 · Sherlockretrieval and comparisonImproved
train on inferences rather than clues only (ablation)CLIP ViT-B/16 · Sherlockretrieval and comparisonImproved
use two-pass widescreen processing instead of center cropping or pad-and-resize (ablation)CLIP ViT-B/16 · Sherlock retrievalP@1Improved
use two-pass widescreen processing instead of center cropping or pad-and-resize (ablation)CLIP ViT-B/16 · Sherlock comparisonhuman accuracyNo clear effect
train on clues rather than inferences, test on cluesCLIP RN50x64 · Sherlock clue retrievalP@1: 43.4 -> 54.3Improved
add clue-to-inference text-only oracle versus image+text modelT5-Large oracle · Sherlock comparison (val)human accuracy: 26.6 -> 22.8Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy