Research map / Vision-language and multimodal / Visual question answering
Unified Questioner Transformer for Descriptive Question Generation in Goal-Oriented Visual Dialogue
Shoya Matsumori, Kosuke Shingyouchi, Yuki Abe, Yosuke Fukuchi, Komei Sugiura, Michita Imai
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 changed | Where | Measured | Effect |
|---|---|---|---|
| unify the Guesser and question generator in one transformer (vs Not Unified) (ablation) | UniQer · CLEVR Ask3 (new image) | task success ratio: 50.61 -> 84.10 | Improved |
| unify the Guesser and question generator in one transformer (vs Not Unified) (ablation) | UniQer · CLEVR Ask4 (new image) | task success ratio: 65.15 -> 83.47 | Improved |
| use transformer Guesser instead of MLP Guesser (vs Not Unified MLP Guesser) (ablation) | UniQer · CLEVR Ask3 (new image) | task success ratio: 69.43 -> 84.10 | Improved |
| use transformer Guesser instead of MLP Guesser (vs Not Unified MLP Guesser) (ablation) | UniQer · CLEVR Ask4 (new image) | task success ratio: 72.89 -> 83.47 | Improved |
| use transformer encoder/decoder instead of GRU/LSTM (vs Vanilla) (ablation) | UniQer · CLEVR Ask3 (new image) | task success ratio: 72.98 -> 84.10 | Improved |
| use transformer encoder/decoder instead of GRU/LSTM (vs Vanilla) (ablation) | UniQer · CLEVR Ask4 (new image) | task success ratio: 67.38 -> 83.47 | Improved |
| learn the OTM with RL instead of random actions (ablation) | UniQer · CLEVR Ask3 (new image) | task success ratio: 1.73 -> 84.10 | Improved |
| learn when to submit instead of forced stop at dialogue end (ablation) | UniQer · CLEVR Ask3 (new image) | task success ratio: 68.78 -> 84.10 | Improved |
| learn when to submit instead of forced stop at dialogue end (ablation) | UniQer · CLEVR Ask4 (new image) | task success ratio: 68.99 -> 83.47 | Improved |
| increase top-k from 5 to 7 (ablation) | UniQer · CLEVR Ask3 (new image) | task success ratio: 84.10 -> 82.66 | Hurt |
| increase top-k from 4 to 7 (ablation) | UniQer · CLEVR Ask4 (new image) | task success ratio: 83.47 -> 76.47 | Hurt |
| increase top-k from 4 to 5 (ablation) | UniQer · CLEVR Ask4 (new image) | task success ratio: 83.47 -> 81.20 | Hurt |
Closest papers by meaning
- Divide-and-Conquer: Tree-structured Strategy with Answer Distribution Estimator for Goal-Oriented Visual Dialogue (AAAI 2025 · similarity 0.90)
- Answerer in Questioner's Mind: Information Theoretic Approach to Goal-Oriented Visual Dialog (NeurIPS 2018 · similarity 0.89)
- Learning Better Visual Dialog Agents With Pretrained Visual-Linguistic Representation (CVPR 2021 · similarity 0.89)
- Automated Generation of Challenging Multiple-Choice Questions for Vision Language Model Evaluation (CVPR 2025 · similarity 0.87)
- ConVQG: Contrastive Visual Question Generation with Multimodal Guidance (AAAI 2024 · similarity 0.86)
- From Retrieval to Generation: A Simple and Unified Generative Model for End-to-End Task-Oriented Dialogue (AAAI 2024 · 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.
