Research map / Image generation / Image restoration
Pragmatic Image Compression for Human-in-the-Loop Decision-Making
Siddharth Reddy, Anca Dragan, Sergey Levine
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 |
|---|---|---|---|
| Train compression to match user actions on original and compressed inputs | PICO · MNIST digit identification | user action agreement | Improved |
| Train compression to match user actions on original and compressed inputs | PICO · CelebA eye-cover verification | user action agreement | Improved |
| Train compression to match user actions on original and compressed inputs | PICO · LSUN Car shopping | user action agreement | Improved |
| Adapt the compression model to the user's task | PICO · CelebA head-cover verification | preservation of task-relevant attribute | Improved |
| Adapt the compression model to the user's task | PICO · LSUN Car color survey | preservation of task-relevant attribute | Improved |
| Use PICO instead of non-adaptive compression at 85 bits per step | Car Racing · Car Racing | road progress: f(1, 11) = 176.32, p < .0001 | Improved |
| Use PICO instead of non-adaptive compression | Car Racing user study · Car Racing | Likert rating: keep the car on the road: 3.45 -> 5.64 | Improved |
| Use PICO instead of non-adaptive compression | Car Racing user study · Car Racing | Likert rating: anticipate steering consequences: 3.82 -> 5.36 | Improved |
| Use PICO instead of non-adaptive compression | Car Racing user study · Car Racing | Likert rating: tell when car was about to go off road: 3.55 -> 5.36 | Improved |
| Use PICO instead of non-adaptive compression | Car Racing user study · Car Racing | Likert rating: tell when to steer: 4.09 -> 5.73 | Improved |
| Use PICO instead of non-adaptive compression | Car Racing user study · Car Racing | Likert rating: determine car position: 4.00 -> 5.82 | Improved |
Closest papers by meaning
- David and Goliath: Small One-step Model Beats Large Diffusion with Score Post-training (ICML 2025 · similarity 0.84)
- Unified Coding for Both Human Perception and Generalized Machine Analytics with CLIP Supervision (AAAI 2025 · similarity 0.84)
- Test-Time Fine-Tuning of Image Compression Models for Multi-Task Adaptability (CVPR 2025 · similarity 0.83)
- Towards Unified Human Perception and Machine Understanding: Token Flow Guided Compression Framework (CVPR 2026 · similarity 0.83)
- CatVTON: Concatenation Is All You Need for Virtual Try-On with Diffusion Models (ICLR 2025 · similarity 0.83)
- Vision Model Pre-training on Interleaved Image-Text Data via Latent Compression Learning (NeurIPS 2024 · similarity 0.82)
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.
