Uporedite metode

	Višestruko modalno mađioničarsko učenje	Samostalno nadgledano pojačano učenje
Oblast	Duboko učenje	Duboko učenje
Porodica	Machine learning	Machine learning
Godina nastanka	2015–2022	2020
Tvorac	Multiple contributors (DeepMind, OpenAI, Google Brain, 2010s–2020s)	Laskin, M.; Srinivas, A.; Abbeel, P. (and contemporaries)
Tip	Multimodal deep RL agent	Self-supervised auxiliary-task learning for RL
Temeljni izvor	Reed, S., Zolna, K., Parisotto, E., Colmenarejo, S. G., Novikov, A., Barth-Maron, G., ... & de Freitas, N. (2022). A Generalist Agent. Transactions on Machine Learning Research.	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.
Drugi nazivi	Multimodal RL, Multi-Sensory Reinforcement Learning, Vision-Language RL, Multi-Input RL	SSL-RL, self-supervised RL, representation-based reinforcement learning, auxiliary-task RL
Srodne	6	4
Sažetak	Multimodal Reinforcement Learning trains agents to make sequential decisions by perceiving and integrating multiple input modalities — such as raw pixels, language instructions, audio, and proprioceptive sensors — simultaneously. Rather than acting on a single data stream, the agent fuses heterogeneous signals into a unified state representation and learns a policy through environmental reward feedback.	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.