Research map / Vision-language and multimodal / Visual question answering
Game of Sketches: Deep Recurrent Models of Pictionary-Style Word Guessing
Ravi Kiran Sarvadevabhatla, Shiv Surya, Trisha Mittal, R. Venkatesh Babu
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 |
|---|---|---|---|
| Use a unified LSTM sequence model instead of the two-phase architecture | Pictionary-style word guessing · GuessWord-160 test | average sequence-level accuracy (k=1): 46.33 -> 62.04 | Improved |
| Use a unified LSTM sequence model instead of the two-phase architecture | Pictionary-style word guessing · GuessWord-160 test | average sequence-level accuracy (k=3): 52.08 -> 69.35 | Improved |
| Use a unified LSTM sequence model instead of the two-phase architecture | Pictionary-style word guessing · GuessWord-160 test | average sequence-level accuracy (k=5): 54.46 -> 71.11 | Improved |
| Add recurrent temporal modeling to the CNN-only predictor | M3 CNN sketch features · GuessWord-160 test | average sequence-level accuracy (k=1): 43.61 -> 62.04 | Improved |
| Add recurrent temporal modeling to the CNN-only predictor | M3 CNN sketch features · GuessWord-160 test | average sequence-level accuracy (k=3): 51.54 -> 69.35 | Improved |
| Add recurrent temporal modeling to the CNN-only predictor | M3 CNN sketch features · GuessWord-160 test | average sequence-level accuracy (k=5): 54.18 -> 71.11 | Improved |
| Use 128 LSTM hidden units (ablation) | Guess-word sequence model · GuessWord-160 validation | average sequence-level accuracy (k=1): 52.77 -> 54.13 | Improved |
| Use 256 LSTM hidden units (ablation) | Guess-word sequence model · GuessWord-160 validation | average sequence-level accuracy (k=1): 52.77 -> 55.03 | Improved |
| Use 512 LSTM hidden units (ablation) | Guess-word sequence model · GuessWord-160 validation | average sequence-level accuracy (k=1): 52.77 -> 55.35 | Improved |
| Add 160-way auxiliary object classification to the Phase I CNN (ablation) | Phase I transition predictor · GuessWord-160 test | transition location accuracy (window width 1): 17.37 -> 20.45 | Improved |
| Use transition-weighted sequence loss rather than default sequence loss (ablation) | Phase I LSTM transition predictor · GuessWord-160 test | transition location accuracy (window width 3): 39.25 -> 41.48 | Improved |
| Use modified ranking loss rather than transition-weighted loss (ablation) | Phase I LSTM transition predictor · GuessWord-160 test | transition location accuracy (window width 1): 19.20 -> 18.87 | Hurt |
Closest papers by meaning
- SketchAgent: Language-Driven Sequential Sketch Generation (CVPR 2025 · similarity 0.87)
- COLUMBUS: Evaluating COgnitive Lateral Understanding Through Multiple-Choice reBUSes (AAAI 2025 · similarity 0.86)
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.86)
- SketchThinker-R1: Towards Efficient Sketch-Style Reasoning in Large Multimodal Models (ICLR 2026 · similarity 0.86)
- ArtWhisperer: A Dataset for Characterizing Human-AI Interactions in Artistic Creations (ICML 2024 · similarity 0.86)
- Prompt Sketching for Large Language Models (ICML 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.
