Research map / Vision-language and multimodal / Visual question answering
Large-Scale Answerer in Questioner's Mind for Visual Dialog Question Generation
Read the paperCheck 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 |
|---|---|---|---|
| Replace SL-Q/RL-QA question-and-guessing pipeline with AQM+ information-gain question selection and posterior guessing | GuessWhich, delta setting · GuessWhich | PMR: 95.45% -> 98.25% | Helped |
| Use AQM+ with independently trained answer model (indA) instead of SL-Q | GuessWhich, delta setting · GuessWhich | PMR: 95.45 -> 97.17 | Helped |
| Use AQM+ with answer model trained from interaction (depA) instead of RL-QA | GuessWhich, delta setting · GuessWhich | PMR: 95.69 -> 98.25 | Helped |
| Use AQM+ posterior guessing with SL-Q generated questions | GuessWhich, delta setting · GuessWhich | PMR: 96.37 | Helped |
| Use a fixed random candidate answer set instead of generating candidate answers at each turn (ablation) | AQM+ indA, GuessWhich non-delta · GuessWhich | PMR: 94.64% -> 92.78% | Hurt |
| Increase question, answer, and class candidate subset size from K=20 to K=40 (ablation) | AQM+ indA, GuessWhich non-delta · GuessWhich | PMR: 94.64% -> 94.79% | Helped |
| Increase question, answer, and class candidate subset size from K=20 to K=40 (ablation) | AQM+ depA, GuessWhich non-delta · GuessWhich | PMR: 97.44% -> 97.77% | Helped |
| Remove dialog history from the answer model (ablation) | AQM+ indA, GuessWhich non-delta · GuessWhich | PMR: 0.22% | Hurt |
| Remove dialog history from the answer model (ablation) | AQM+ depA, GuessWhich non-delta · GuessWhich | PMR: 0.56% | Hurt |
| Remove dialog history from the answer model (ablation) | AQM+ indA, GuessWhich delta · GuessWhich | PMR: 0.46% | Hurt |
| Remove dialog history from the answer model (ablation) | AQM+ depA, GuessWhich delta · GuessWhich | PMR: 0.21% | Hurt |
| Use AQM+ information-gain question selection with the SL-Q guesser (ablation) | SL-Q guesser · GuessWhich | PMR | No clear effect |
Closest papers by meaning
- Answerer in Questioner's Mind: Information Theoretic Approach to Goal-Oriented Visual Dialog (NeurIPS 2018 · similarity 0.92)
- Two Can Play This Game: Visual Dialog With Discriminative Question Generation and Answering (CVPR 2018 · similarity 0.88)
- Divide-and-Conquer: Tree-structured Strategy with Answer Distribution Estimator for Goal-Oriented Visual Dialogue (AAAI 2025 · similarity 0.87)
- Rethinking Generative Large Language Model Evaluation for Semantic Comprehension (ICML 2024 · similarity 0.86)
- MindGYM: What Matters in Question Synthesis for Thinking-Centric Fine-Tuning? (NeurIPS 2025 · similarity 0.86)
- ChatQA: Surpassing GPT-4 on Conversational QA and RAG (NeurIPS 2024 · similarity 0.85)
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.