השוואת שיטות
סקרו את השיטות שבחרתם זו לצד זו; שורות שבהן יש הבדל מודגשות.
| למידת חיזוק מפוקחת באופן חלש× | למידת חיזוק בפיקוח-עצמי× | |
|---|---|---|
| תחום | למידה עמוקה | למידה עמוקה |
| משפחה | Machine learning | Machine learning |
| שנת המקור≠ | 2010s–present | 2020 |
| הוגה השיטה≠ | Multiple contributors; reward-learning framing: Christiano et al. (2017) | Laskin, M.; Srinivas, A.; Abbeel, P. (and contemporaries) |
| סוג≠ | Reinforcement learning with imperfect or partial reward supervision | Self-supervised auxiliary-task learning for RL |
| מקור מכונן≠ | Sutton, R. S. & Barto, A. G. (2018). Reinforcement Learning: An Introduction (2nd ed.). MIT Press. ISBN: 978-0-262-03924-6 | Laskin, M., Srinivas, A., & Abbeel, P. (2020). CURL: Contrastive Unsupervised Representations for Reinforcement Learning. Proceedings of the 37th International Conference on Machine Learning (ICML), PMLR 119, 5639–5650. link ↗ |
| כינויים | WSRL, weak-reward RL, imperfect-reward reinforcement learning, reward-impoverished RL | SSL-RL, self-supervised RL, representation-based reinforcement learning, auxiliary-task RL |
| קשורות≠ | 3 | 4 |
| תקציר≠ | Weakly supervised reinforcement learning (WSRL) trains agents in environments where the reward signal is imperfect, sparse, delayed, or only partially informative — unlike dense fully-supervised RL. The agent must learn effective policies despite incomplete feedback, using auxiliary signals, reward modeling, or preference learning to compensate for the weak supervision. | Self-supervised Reinforcement Learning (SSL-RL) augments standard RL training with self-supervised auxiliary objectives — such as contrastive, predictive, or data-augmentation-based tasks — applied to the agent's own experience. These objectives improve the quality of learned representations without requiring extra human labels, enabling faster convergence and better sample efficiency, especially in high-dimensional observation spaces like raw pixels. |
| ScholarGateמערך נתונים ↗ |
|
|