Research map / Vision-language and multimodal / Visual question answering
GuessWhat?! Visual Object Discovery Through Multi-Modal Dialogue
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| include the question, object category, and spatial features in the oracle | Oracle MLP · GuessWhat?! oracle test | classification error: 50.9% -> 21.5% | Helped |
| add crop features to Question + Category + Spatial (ablation) | Oracle MLP · GuessWhat?! oracle test | classification error: 21.5% -> 22.1% | Hurt |
| add full-image features to Question + Category + Spatial (ablation) | Oracle MLP · GuessWhat?! oracle test | classification error: 21.5% -> 23.5% | Hurt |
| remove spatial features from Question + Category + Spatial + Crop (ablation) | Oracle MLP · GuessWhat?! oracle test | classification error: 22.1% -> 24.7% | Hurt |
| remove category features from Question + Category + Spatial + Crop (ablation) | Oracle MLP · GuessWhat?! oracle test | classification error: 22.1% -> 26.2% | Hurt |
| add category and spatial features to the question | Oracle MLP · GuessWhat?! oracle test | classification error: 41.2% -> 21.5% | Helped |
| add VGG image features (ablation) | LSTM guesser · GuessWhat?! guesser test | classification error: 38.7% -> 39.5% | Hurt |
| add VGG image features (ablation) | HRED guesser · GuessWhat?! guesser test | classification error: 39.0% -> 39.6% | Hurt |
| use a hierarchical HRED dialogue encoder instead of a flat LSTM (ablation) | Guesser · GuessWhat?! guesser test | classification error: 38.7% -> 39.0% | Hurt |
| train the question generator with oracle answers rather than ground-truth answers (ablation) | QGen · GuessWhat?! question generation test | guesser classification error: 53.2% -> 66.0% | Hurt |
| generate questions instead of using human-generated dialogue | QGen+GT · GuessWhat?! question generation test | guesser classification error: 38.7% -> 53.2% | Hurt |
Closest papers by meaning
- Learning Better Visual Dialog Agents With Pretrained Visual-Linguistic Representation (CVPR 2021 · similarity 0.91)
- Parallel Attention: A Unified Framework for Visual Object Discovery Through Dialogs and Queries (CVPR 2018 · similarity 0.89)
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.87)
- Divide-and-Conquer: Tree-structured Strategy with Answer Distribution Estimator for Goal-Oriented Visual Dialogue (AAAI 2025 · similarity 0.86)
- DialogueVPR: Towards Conversational Visual Place Recognition (CVPR 2026 · similarity 0.86)
- Exploring the Effectiveness of Object-Centric Representations in Visual Question Answering: Comparative Insights with Foundation Models (ICLR 2025 · similarity 0.86)
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.