Contrastive learning as goal-conditioned reinforcement learning
In reinforcement learning (RL), it is easier to solve a task if given a good representation.
While deep RL should automatically acquire such good representations, prior work often …
While deep RL should automatically acquire such good representations, prior work often …
Hiql: Offline goal-conditioned rl with latent states as actions
Unsupervised pre-training has recently become the bedrock for computer vision and natural
language processing. In reinforcement learning (RL), goal-conditioned RL can potentially …
language processing. In reinforcement learning (RL), goal-conditioned RL can potentially …
Compositional generalization from first principles
T Wiedemer, P Mayilvahanan… - Advances in Neural …, 2024 - proceedings.neurips.cc
Leveraging the compositional nature of our world to expedite learning and facilitate
generalization is a hallmark of human perception. In machine learning, on the other hand …
generalization is a hallmark of human perception. In machine learning, on the other hand …
What is essential for unseen goal generalization of offline goal-conditioned rl?
Offline goal-conditioned RL (GCRL) offers a way to train general-purpose agents from fully
offline datasets. In addition to being conservative within the dataset, the generalization …
offline datasets. In addition to being conservative within the dataset, the generalization …
[PDF][PDF] Structure in reinforcement learning: A survey and open problems
Reinforcement Learning (RL), bolstered by the expressive capabilities of Deep Neural
Networks (DNNs) for function approximation, has demonstrated considerable success in …
Networks (DNNs) for function approximation, has demonstrated considerable success in …
Learning to extrapolate: A transductive approach
Machine learning systems, especially with overparameterized deep neural networks, can
generalize to novel test instances drawn from the same distribution as the training data …
generalize to novel test instances drawn from the same distribution as the training data …
Drone Landing and Reinforcement Learning: State-of-art, Challenges and Opportunities
J Amendola, LR Cenkeramaddi… - IEEE Open Journal of …, 2024 - ieeexplore.ieee.org
Unmanned aerial vehicles, and special multirotor drones, have shown great relevance in a
plethora of missions that require high affordance, field of view, and precision. Their limited …
plethora of missions that require high affordance, field of view, and precision. Their limited …
Constrained gpi for zero-shot transfer in reinforcement learning
For zero-shot transfer in reinforcement learning where the reward function varies between
different tasks, the successor features framework has been one of the popular approaches …
different tasks, the successor features framework has been one of the popular approaches …
Metric Residual Network for Sample Efficient Goal-Conditioned Reinforcement Learning
Goal-conditioned reinforcement learning (GCRL) has a wide range of potential real-world
applications, including manipulation and navigation problems in robotics. Especially in such …
applications, including manipulation and navigation problems in robotics. Especially in such …
Structure in Deep Reinforcement Learning: A Survey and Open Problems
Reinforcement Learning (RL), bolstered by the expressive capabilities of Deep Neural
Networks (DNNs) for function approximation, has demonstrated considerable success in …
Networks (DNNs) for function approximation, has demonstrated considerable success in …