Pharmacokinetics of a novel formulation of ivermectin after administration to goats

By Andrew Y. Ng, Stuart J. Russell. 2000

1,115 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Medical & Health AI

Citations in the corpus

Artificial Intelligence, Values, and Alignment (2020, background): “From a practical point of view, any attempt to infer a reward or utility function from observed behaviour encounters the problem of what Ng and Russell (2000) term ‘degeneracy’: at any moment, there is a large set of reward functions for which an observed behaviour is optimal.”
Concrete Problems in AI Safety (2016, background): “ess divorced from interaction with the environment. This could involve either learning a reward function from samples of state-reward pairs, or from trajectories, as in inverse reinforcement learning [107, 51]. However, this forfeits the ability to further learn the reward function after the pretrai”
Research Priorities for Robust and Beneficial Artificial Intelligence (2015, background): “For example, inverse reinforcement learning may offer a viable approach, in which a system infers the preferences of another rational or nearly rational actor by observing its behavior (Russell 1998; Ng and Russell 2000).”
Implicit Behavioral Cloning (2021, background): “One route is by collecting on-policy data of the learned policy, and potentially either labeling with rewards to perform on-policy reinforcement learning (RL) [54, 55, 56] or labeling actions by an expert [2].”
Hyperparameter Selection for Imitation Learning (2021, methodology): “This approach, also known as Inverse Reinforcement Learning (IRL), tries to recover this unknown reward function and use an RL algorithm to train a policy to maximize it (Russell, 1998; Ng et al., 2000; Ziebart et al., 2008).”
End-to-End Robotic Reinforcement Learning without Reward Engineering (2019, background): “Data-driven approaches for reward specification [29, 1, 11, 17, 12, 48, 33, 9, 27, 4, 18] seek to overcome this issue, but typically require demonstration data to acquire rewards.”
Learning from Physical Human Corrections, One Feature at a Time (2018, methodology): “Similar to prior IRL work [13, 16, 18, 24], the robot’s objective function (here a cost function) is parametrized by θ ∈ RF , which weights the importance of these features along the entire trajectory:”
Cooperative Inverse Reinforcement Learning (2016, background): “Ng & Russell (2000) deﬁne inverse reinforcement learning (IRL) as follows: “ Given measurements of an [actor]’s behavior over time. . . . Determine the reward function being optimized.””
Concrete Problems for Autonomous Vehicle Safety: Advantages of Bayesian Deep Learning (2017, background): “IRL is the problem of inferring another agent’s latent reward function from observations [Ng and Russell, 2000].”
Task-Embedded Control Networks for Few-Shot Imitation Learning (2018, background): “ behavioural cloning, in which the agent learns a mapping from observations to actions given demonstrations, in a supervised learning manner [12,13]. Another focus is inverse reinforcement 2 learning [14], where an agent attempts to estimate a reward function that describes the given demonstrations ”
Inverse Optimal Control with Linearly-Solvable MDPs (2010, methodology): “Indeed in generic MDPs there is a continuum of cost and value functions for which a given control law is optimal (Ng & Russell, 2000).”
Pretraining Deep Actor-Critic Reinforcement Learning Algorithms With Expert Demonstrations (2018, methodology): “Here the expert policy π ∗ is different from that of IRL [Ng and Russell, 2000], imitation learning [Ho and Ermon, 2016b] or LfD [Hester et al. , 2017], since π ∗ here is not the optimum policy of the MDPs.”
Penalizing Side Effects using Stepwise Relative Reachability (2018, background): “…to specifying a side effects penalty is to teach the agent to avoid side effects through human oversight, such as inverse reinforcement learning [Ng and Russell, 2000; Hadfield-Menell et al., 2016], demonstrations [Abbeel and Ng, 2004], or human feedback [Christiano et al., 2017; Saunders et…”
Preference-learning based Inverse Reinforcement Learning for Dialog Control (2012, background): “To set an appropriate reward function automatically, inverse reinforcement learning (IRL) has been proposed [5, 6] and is adopted for dialog control [7, 8].”