Research map / Learning theory and algorithms / Prediction and decision losses
Hyperparameter Optimization for Tracking With Continuous Deep Q-Learning
Xingping Dong, Jianbing Shen, Wenguan Wang, Yu Liu, Ling Shao, Fatih Porikli
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 |
|---|---|---|---|
| Optimize Siam-py hyperparameters dynamically for each frame | Siam-py · OTB-2013 OPE | AUC: 0.597 to 0.629 | Improved |
| Optimize Siam-py hyperparameters dynamically | Siam-py · VOT-2015 | accuracy: 0.540 -> 0.578 | Improved |
| Optimize Siam-py hyperparameters dynamically | Siam-py · VOT-2015 | robustness: 1.366 -> 1.578 | Hurt |
| Optimize Siam-py hyperparameters dynamically | Siam-py · OTB-2013 | FPS: 74 FPS to 69 FPS | Hurt |
| Pre-train the action-predict network with supervised learning on original hyperparameters (ablation) | NAF for tracking hyperparameter optimization · OTB-2013 | mean IoU | Improved |
| Add the staged heuristic: pre-train the action network, learn value and matrix networks with the action network fixed, then train all networks (ablation) | NAF for tracking hyperparameter optimization · OTB-2013 | mean IoU | Improved |
| Use the original NAF training without the supervised action-network initialization (ablation) | NAF for tracking hyperparameter optimization · OTB-2013 | mean IoU | Hurt |
| Use dynamic hyperparameter optimization | Siam-py · OTB-100 | precision: 1.5% | Improved |
| Use dynamic hyperparameter optimization | Siam-py · OTB-100 | AUC: 3.3% | Improved |
| Use dynamic hyperparameter optimization | Siam-py · OTB-50 | AUC: 7.4% | Improved |
| Use dynamic hyperparameter optimization | Siam-py · OTB-100 attributes: SV, OCC, MB, FM, IPR, OPR | overlap success | Improved |
Closest papers by meaning
- Dynamical Hyperparameter Optimization via Deep Reinforcement Learning in Tracking (TPAMI 2019 · similarity 0.95)
- Real-time 'Actor-Critic' Tracking (ECCV 2018 · similarity 0.85)
- Continuous Deep Q-Learning with Model-based Acceleration (ICML 2016 · similarity 0.84)
- QLABGrad: A Hyperparameter-Free and Convergence-Guaranteed Scheme for Deep Learning (AAAI 2024 · similarity 0.81)
- Learning from Suboptimal Data in Continuous Control via Auto-Regressive Soft Q-Network (ICML 2025 · similarity 0.81)
- Adaptive $Q$-Network: On-the-fly Target Selection for Deep Reinforcement Learning (ICLR 2025 · similarity 0.81)
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.
