The paper interprets policy-gradient algorithms using continuation theory.
problem Optimizing nonconvex functions in reinforcement learning.
method Formulates policy optimization as optimization by continuation, interprets policy-gradient algorithms as implicitly optimizing deterministic policies.
result Exploration in policy-gradient algorithms is seen as computing a continuation of the return of the policy.
Faster policy learning via continuous-time gradients.
problem Efficiently estimating policy gradients for continuous-time systems.
method Approximating continuous-time gradients directly, using adaptive discretization.
result More efficient policy gradient estimator leads to faster learning.
Develops DPG methods for continuous-time RL with deterministic policies.
problem High variance and slow convergence in stochastic policy RL methods.
method Derives continuous-time policy gradient formula and proposes CT-DDPG algorithm.
result CT-DDPG achieves superior stability and faster convergence in continuous-time RL.
Naive investors make riskier choices than optimal strategies in continuous-time finance.
problem Continuous-time Markowitz portfolio selection with naive reoptimization.
method Analytical derivation of naive policies from discretely naive policies.
result Naive policies are always riskier and less efficient than equilibrium policies.
We study the problem of policy evaluation and learning from batched contextual bandit data when treatments are continuous, going beyond previous work on discrete treatments. Previous work for discrete treatment/action spaces focuses on inverse probability weighting (IPW) and doubly robust (DR) methods that use a reject…
Paper tackles RL with continuous actions and unmeasured confounders.
problem Offline policy learning with continuous actions and unmeasured confounders.
method Developed a novel identification result and a minimax estimator for nonparametric policy value estimation.
result Introduced a policy-gradient-based algorithm to identify the optimal policy.
We propose a method for tackling catastrophic forgetting in deep reinforcement learning that is \textit{agnostic} to the timescale of changes in the distribution of experiences, does not require knowledge of task boundaries, and can adapt in \textit{continuously} changing environments. In our \textit{policy consolidati…
Study improves policy search in continuous control by using heavy-tailed distributions.
problem Challenges in continuous space policy search due to non-convexity and myopic-farsighted incentives.
method Introduced heavy-tailed policy parameterizations and analyzed convergence rates and stability.
result Convergence rate to stationarity depends on policy's tail index and exploration tolerance.
We identify a fundamental problem in policy gradient-based methods in continuous control. As policy gradient methods require the agent's underlying probability distribution, they limit policy representation to parametric distribution classes. We show that optimizing over such sets results in local movement in the actio…
Study shows certainty equivalent policy minimizes regret in continuous-time systems.
problem Minimizing regret in continuous-time stochastic linear-quadratic systems.
method Theoretical analysis of randomized certainty equivalent policy.
result Establishes square-root of time regret bounds and linear scaling with parameters.
New method optimizes decision-making in uncertain environments.
problem Optimal decision-making under partial observability.
method Nested sequential Monte Carlo algorithm for continuous POMDPs.
result Demonstrated effectiveness on continuous POMDP benchmarks.
Solving goal-oriented tasks is an important but challenging problem in reinforcement learning (RL). For such tasks, the rewards are often sparse, making it difficult to learn a policy effectively. To tackle this difficulty, we propose a new approach called Policy Continuation with Hindsight Inverse Dynamics (PCHID). Th…
Efficient deep policy gradient method for continuous-time control problems.
problem Optimal control in continuous time with fine time discretization.
method Multi-scale deep policy gradient method with varying time discretization.
result Targeted efficiency in computational resources achieved through multi-scale approach.
Entropy regularization improves policy optimization in reinforcement learning.
problem Improving policy optimization in reinforcement learning.
method Entropy regularization is introduced to soften the greedy policy towards a more diverse softmax policy, leading to a continuously parameterized algorithm that interpolates between policy gradient and Q-learning.
result An intermediate algorithm can improve performance in reinforcement learning.
Study policy gradient and actor-critic methods for continuous-time reinforcement learning.
problem Continuous-time reinforcement learning with policy gradient and actor-critic approaches.
method Regularized exploratory formulation, martingale approach, simultaneous policy and value function updates.
result Proposed two types of actor-critic algorithms for online and offline learning.
The paper tackles counterfactual learning for stochastic policies with continuous actions.
problem Learning stochastic policies with continuous actions from logged data.
method Introduces a joint kernel embedding of contexts and actions to model continuous actions, and uses proximal point algorithms and smooth estimators for optimization.
result Demonstrates the benefits of using proximal point algorithms and smooth estimators for counterfactual learning.
New framework for policy gradient methods in continuous time reinforcement learning.
problem Addressing policy gradient methods for continuous time reinforcement learning.
method Control randomisation technique to derive policy gradient representation for various Markovian control problems.
result Demonstrated application to optimal switching problems in the energy sector.
This paper proposes a novel deep reinforcement learning architecture that was inspired by previous tree structured architectures which were only useable in discrete action spaces. Policy Prediction Network offers a way to improve sample complexity and performance on continuous control problems in exchange for extra com…
RANDPOL uses randomized networks for efficient reinforcement learning in continuous state and action MDPs.
problem Efficient reinforcement learning in environments with continuous state and action spaces.
method RANDPOL uses randomized function approximation to represent policy and value functions, providing finite time guarantees and improved numerical performance.
result RANDPOL achieves better numerical performance and provides finite time guarantees compared to deep neural network based algorithms.
Paper solves Bayesian bandit problem with continuous-time limit and approximate policy.
problem Finding optimal policy in Bayesian bandit problems with large horizons.
method Reformulates Bayesian bandit problem as continuous Hamilton-Jacobi-Bellman (HJB) equation and proposes approximate Bayes-optimal policy.
result Approximate Bayes-optimal policy for large horizons with constant computational cost.
Study policy gradient for large-agent mean-field control and game in continuous time.
problem Optimal policy learning for large number of agents in continuous-time mean-field systems.
method Policy gradient method applied to linear-quadratic mean-field control and game models.
result Policy gradient converges to optimal solution at a linear rate for both mean-field control and game.
Second-order estimator improves continuous-time policy evaluation.
problem Estimating value surfaces from discrete data with time-inhomogeneous dynamics.
method Moment-matching coefficients for high-order generator regression.
result Second-order estimator consistently outperforms Bellman baseline.
Unified framework for policy improvement in RL with benefits in data efficiency and computation.
problem Improving data efficiency and computation in reinforcement learning for continuous control.
method Local, regularized policy improvement with tree search for continuous action spaces.
result Improves data efficiency and reduces wall-clock time in high-dimensional domains.
The (contextual) multi-armed bandit problem (MAB) provides a formalization of sequential decision-making which has many applications. However, validly evaluating MAB policies is challenging; we either resort to simulations which inherently include debatable assumptions, or we resort to expensive field trials. Recently …
Modyn automates continuous ML model training on growing datasets.
problem Continuous model retraining is costly and impractical.
method Data-centric ML platform with policies for continuous training.
result Modyn enables high throughput training with sample-level data selection.
In the context of learning deterministic policies in continuous domains, we revisit an approach, which was first proposed in Continuous Actor Critic Learning Automaton (CACLA) and later extended in Neural Fitted Actor Critic (NFAC). This approach is based on a policy update different from that of deterministic policy g…
This paper improves reinforcement learning policies in a scalable way.
problem Ensuring monotonic policy improvement in entropy-regularized RL.
method Derives an entropy-aware lower bound and proposes a novel RL algorithm.
result Demonstrates effectiveness in continuous-state tasks using a linear function approximator.
Paper proposes interpretable RL policies from a mixture of experts.
problem Making RL policies transparent and understandable in real-world applications.
method Policy iteration scheme with interpretable experts and prototypical states.
result Proposed algorithm learns policies comparable to neural networks but more interpretable.
We observe that several existing policy gradient methods (such as vanilla policy gradient, PPO, A2C) may suffer from overly large gradients when the current policy is close to deterministic (even in some very simple environments), leading to an unstable training process. To address this issue, we propose a new method, …
Many continuous control tasks have bounded action spaces. When policy gradient methods are applied to such tasks, out-of-bound actions need to be clipped before execution, while policies are usually optimized as if the actions are not clipped. We propose a policy gradient estimator that exploits the knowledge of action…
Reinforcement learning algorithms rely on exploration to discover new behaviors, which is typically achieved by following a stochastic policy. In continuous control tasks, policies with a Gaussian distribution have been widely adopted. Gaussian exploration however does not result in smooth trajectories that generally c…
Continuous-time Q-learning theory developed for reinforcement learning.
problem Continuous-time reinforcement learning challenges.
method Entropy-regularized, exploratory diffusion process formulation; first-order approximation of Q-function; martingale conditions.
result Developed a q-learning theory independent of time discretization.
ReCAP adapts to dynamic financial markets by segmenting and combining policy vectors.
problem Inefficient traditional PM approaches in non-stationary financial markets.
method Integrates continual learning into PM, segmenting regimes and adapting policies.
result Consistently outperforms baselines in real-world financial datasets.
Continuous control imitation learning fails if expert actions are smooth.
problem Continuous control imitation learning fails if expert actions are smooth.
method Study of imitation learning in discrete-time, continuous state-and-action control systems.
result Any smooth, deterministic imitator policy suffers exponentially larger error than the expert.
Proposes a new method to estimate continuous treatment policies and match treatments effectively.
problem Current methods struggle with continuous treatment policies and complex matching.
method Formulates treatment effectiveness as a parametrizable model, using deep learning for optimization.
result Significant improvement in treatment effectiveness and matching efficiency.
Develops deep jump learning for continuous treatment OPE.
problem Estimating mean outcomes under new treatment rules using historical data from different rules.
method Adaptive deep discretization of continuous treatment space using deep learning and multi-scale change point detection.
result Validated method through theoretical results, simulations, and real application to Warfarin Dosing.
Improves RL algorithms with two techniques.
problem Enhance off-policy RL performance.
method Formulates RL as proximal point iteration; uses value functions for improved action value estimate.
result Significant performance improvement on RL benchmarks.
In multi-task reinforcement learning there are two main challenges: at training time, the ability to learn different policies with a single model; at test time, inferring which of those policies applying without an external signal. In the case of continual reinforcement learning a third challenge arises: learning tasks…
Study investigates key design choices in on-policy RL algorithms.
problem Lack of transparency in RL algorithm implementations.
method Implemented >50 design choices in a unified RL framework.
result Insights and recommendations for on-policy RL training.
We focus on the problem of teaching a robot to solve tasks presented sequentially, i.e., in a continual learning scenario. The robot should be able to solve all tasks it has encountered, without forgetting past tasks. We provide preliminary work on applying Reinforcement Learning to such setting, on 2D navigation tasks…
Evolutionary Strategies optimize hyper-parameters for off-policy learning.
problem Hyper-parameter sensitivity in off-policy learning.
method Application of Evolutionary Strategies for online hyper-parameter tuning.
result Our method outperforms state-of-the-art baselines.
Hybrid Policy Optimization tackles reinforcement learning in hybrid spaces, improving performance over PPO.
problem Credit assignment issues and biased gradients in hybrid discrete-continuous action spaces.
method Mixed gradient estimator combining pathwise and score-function gradients, reformulating problems in hybrid form.
result HPO substantially outperforms PPO on inventory control and switched systems, with performance gaps increasing with continuous action dimension.
PFPN uses particle filtering to improve character control in physics-based simulations.
problem Premature commitment to suboptimal actions in high-dimensional continuous control problems for articulated characters.
method Proposes a particle-based action policy using particle filtering to dynamically explore and discretize the action space.
result Demonstrates better imitation performance and robustness to external perturbations compared to Gaussian policies.
Paper proposes a policy-search algorithm to learn entropy-maximizing exploration policies in reward-free environments.
problem Reward-free learning in high-dimensional, continuous-control domains.
method Maximum Entropy POLicy optimization (MEPOL) algorithm that maximizes a non-parametric state entropy estimate.
result MEPOL learns a maximum-entropy exploration policy that facilitates learning various reward-based tasks.
It has long been assumed that high dimensional continuous control problems cannot be solved effectively by discretizing individual dimensions of the action space due to the exponentially large number of bins over which policies would have to be learned. In this paper, we draw inspiration from the recent success of sequ…
Deep Reinforcement Learning (DRL) algorithms for continuous action spaces are known to be brittle toward hyperparameters as well as \cut{being}sample inefficient. Soft Actor Critic (SAC) proposes an off-policy deep actor critic algorithm within the maximum entropy RL framework which offers greater stability and empiric…
Maximum entropy deep reinforcement learning (RL) methods have been demonstrated on a range of challenging continuous tasks. However, existing methods either suffer from severe instability when training on large off-policy data or cannot scale to tasks with very high state and action dimensionality such as 3D humanoid l…
Develops a new RL algorithm for medical treatment regimes.
problem Optimal dose determination in continuous action environments.
method Quasi-optimal learning algorithm for near-optimal actions.
result Guaranteed convergence and effectiveness in real applications.