Research map / Reinforcement learning / Visual and representation RL
Sidekick Policy Learning for Active Visual Exploration
Santhosh K. Ramakrishnan, Kristen Grauman
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 |
|---|---|---|---|
| add reward-based sidekick | ltla (REINFORCE) · SUN360 | avg MSE x1000: 24.94 -> 23.44 | Improved |
| add reward-based sidekick | ltla (REINFORCE) · ModelNet Hard | avg MSE x1000: 6.30 -> 5.80 | Improved |
| add demonstration-based sidekick | ltla (REINFORCE) · SUN360 | avg MSE x1000: 24.94 -> 24.24 | Improved |
| add demonstration-based sidekick | ltla (REINFORCE) · ModelNet Hard | avg MSE x1000: 6.30 -> 6.32 | No clear effect |
| add reward sidekick to actor-critic | asymm-ac · SUN360 | avg MSE x1000: 23.74 -> 23.36 | Improved |
| add reward sidekick to actor-critic | asymm-ac · ModelNet Hard | avg MSE x1000: 6.24 -> 5.75 | Improved |
| add demonstration sidekick to actor-critic | asymm-ac · ModelNet Hard | avg MSE x1000: 6.24 -> 6.13 | Improved |
| replace behavior cloning of an expert with reward sidekick | expert-clone · ModelNet Hard | avg MSE x1000: 6.41 -> 5.80 | Improved |
| replace behavior cloning of an expert with reward sidekick | expert-clone · SUN360 | adv MSE x1000: 28.50 -> 28.54 | No clear effect |
| replace full-observability test-time demo policy with demo sidekick | demo-actions · SUN360 | avg MSE x1000: 26.12 -> 24.24 | Improved |
| replace full-observability test-time demo policy with demo sidekick | demo-actions · ModelNet Hard | avg MSE x1000: 5.82 -> 6.32 | Hurt |
| use actor-critic instead of REINFORCE with reward sidekick | ours(rew) · ModelNet Hard | avg MSE x1000: 5.80 -> 5.75 | Improved |
Closest papers by meaning
- AdaGlimpse: Active Visual Exploration with Arbitrary Glimpse Position and Scale (ECCV 2024 · similarity 0.89)
- Learning to Look Around: Intelligently Exploring Unseen Environments for Unknown Tasks (CVPR 2018 · similarity 0.89)
- Learning Exploration Policies for Navigation (ICLR 2019 · similarity 0.88)
- Learning Active Camera for Multi-Object Navigation (NeurIPS 2022 · similarity 0.88)
- Glimpse-Attend-and-Explore: Self-Attention for Active Visual Exploration (ICCV 2021 · similarity 0.87)
- Learning with a Mole: Transferable latent spatial representations for navigation without reconstruction (ICLR 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.
