PGS uses neural networks to improve policies online without search trees.
problem Limited scalability of Monte Carlo Tree Search (MCTS) for high branching factor games.
method Adapts a neural network simulation policy via policy gradient updates, avoiding search trees.
result PGS achieves comparable performance to MCTS and defeats strong Hex agents.
Trust-region methods and natural gradients are equivalent in certain policy search scenarios.
problem Improving policy search methods in continuous control tasks.
method Introducing compatible policy search (COPOS) that uses natural parameterization and compatible value function approximation to control entropy loss.
result COPOS yields state-of-the-art results in challenging tasks and reduces entropy loss.
Bayesian approach for policy search in stochastic domains.
problem Policy search in stochastic domains.
method Nested probabilistic programs, Lightweight Metropolis-Hastings (LMH) adaptation.
result Similar quality policies learned with simpler algorithm.
Bayesian optimization for contextual policy search improves robot learning.
problem Scalable robot learning with limited data.
method Factored contextual representation using target and environment contexts.
result Experience can be generalized over target contexts, leading to faster learning and better generalization.
New method learns search policies by inspecting and improving past roll-outs.
problem Learning good search policies for complex combinatorial spaces.
method Retrospective imitation learning, improving policy through past roll-outs.
result Policy can iteratively scale up to larger problems.
Beam search policies learned via imitation learning.
problem Beam search policies are not explicitly learned by models during training.
method Developed a meta-algorithm for learning beam search policies using imitation learning.
result Showed no-regret guarantees for learning beam search policies.
Enhances RL performance with a population-guided parallel learning scheme.
problem Improving off-policy reinforcement learning performance.
method Population-guided parallel learning scheme with shared experience replay buffer and soft policy update.
result Monotone improvement of the expected cumulative return proved theoretically and demonstrated in practice.
CF-GPS learns policies from logged data by considering counterfactual outcomes.
problem Learning policies from limited real experience in complex environments.
method Assumes logged real experience and models counterfactual outcomes. Uses structural causal models for evaluation.
result Improves policy evaluation and search results on a grid-world task.
New approach learns walk and trot gaits from simulated quadruped using strategic exploration.
problem Learning symmetric gaits (walk and trot) from high-dimensional action spaces.
method Introduced symmetry properties into initial covariance of Gaussian search distribution for strategic exploration. Used episode-based likelihood ratio policy gradient and relative entropy policy search.
result Significant performance enhancement in learning walk and trot gaits compared to random gaits.
Direct contextual policy search methods learn to improve policy parameters and simultaneously generalize these parameters to different context or task variables. However, learning from high-dimensional context variables, such as camera images, is still a prominent problem in many real-world tasks. A naive application o…
EffAcTS uses active learning to select model parameters for robust policy search.
problem Learning robust policies that perform well in unseen environments.
method EffAcTS is an active learning framework that selects model parameters using Linear Bandits to minimize data collection.
result Efficiency gains in sample collection and improved policy performance.
Methods for learning to search for structured prediction typically imitate a reference policy, with existing theoretical guarantees demonstrating low regret compared to that reference. This is unsatisfactory in many applications where the reference policy is suboptimal and the goal of learning is to improve upon it. Ca…
Paper shows MCTS approximates policy optimization, proposing an improved variant.
problem Improving AI performance through better MCTS algorithms.
method Shows MCTS approximates policy optimization problem, proposes a new algorithm.
result Proposed algorithm reliably outperforms original AlphaZero in multiple domains.
Simple policy search outperforms advanced learnable test-time augmentation techniques.
problem Improving predictive performance through test-time data augmentation.
method Greedy policy search (GPS) for learning test-time augmentation policies.
result Augmentation policies learned with GPS achieve superior predictive performance and robustness.
Paper proposes MCTSPO for better reinforcement learning policy optimization.
problem Local optima and saddle points in gradient-based methods and poor initialization in gradient-free methods.
method Monte-Carlo tree search combined with gradient-free optimization.
result Improved performance on reinforcement learning tasks with deceptive or sparse reward functions.
Survey on algorithms for quick robot learning.
problem Efficiently learn robot controllers with limited data.
method Leverage prior knowledge and data-driven models.
result Combining prior knowledge and surrogate models improves learning.
Introduces GAMPS for better model-based policy learning.
problem Misspecified model classes lead to poor policy estimates.
method Exploits current policy to learn approximate transition model, focusing on relevant parts of the environment.
result Empirically validated GAMPS on benchmark domains, demonstrating improved properties.
Efficient neural architecture search by sampling structure and operations.
problem Efficiently searching for optimal neural architectures.
method Decouples structure and operation search, using reinforcement learning with policy vectors.
result Significantly improved efficiency compared to traditional methods.
New methods use vector search and nearest-neighbor matching for policy learning in causal inference.
problem Learning optimal policies in causal inference with limited data.
method RAG-based policy learning with vector search and nearest-neighbor matching.
result The methods bound the within-candidate choice regret and evaluate the one-step method directly as a policy.
Early stopping method saves up to 75% computation time in policy search tasks.
problem Lengthy evaluation times in optimization problems, especially in robotics.
method A generalized early stopping criterion that only uses objective value at each time step.
result The method saves up to 75% computation time compared to no stopping.
Bayesian optimization improves policy search in reinforcement learning.
problem Finding optimal policies with high variance estimates from random samples.
method Develops an algorithm combining Bayesian optimization and policy gradients.
result Improves sample complexity and reduces variance in empirical evaluations.
This paper improves self-play learning in games by manipulating experience distributions.
problem Improving self-play learning in games through better experience sampling.
method Three approaches: weighted sampling, Prioritized Experience Replay, and diversifying trajectories.
result Major improvements in early training performance in some games, minor improvements overall.
New algorithm scales model-based policy search for robotics to high-dimensional systems.
problem Inefficient scaling of model-based policy search algorithms in high-dimensional state/action spaces.
method Introduces parameterized black-box priors to scale up model learning and improve robustness to prior inaccuracies.
result Significantly more data-efficient than previous algorithms, learning gaits in 16-30 seconds.
Transformer learns to search through reinforcement learning, mimicking DFS.
problem Understanding how transformers learn search capabilities in RL.
method Two-head transformer, depth-wise curriculum, discounted returns.
result Transformer policy generalizes depth and prioritizes high-probability branches.
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.
Contextual policy search allows adapting robotic movement primitives to different situations. For instance, a locomotion primitive might be adapted to different terrain inclinations or desired walking speeds. Such an adaptation is often achievable by modifying a small number of hyperparameters. However, learning, when …
Genie optimizes search marketplaces by estimating policy impacts without risky experiments.
problem Optimizing search marketplaces with frequent policy changes and limited randomized experiments.
method Genie uses an open box simulation engine and click calibration model to estimate KPI impacts.
result Genie outperforms existing approaches in optimizing Bing Ads Marketplace.
Paper develops efficient nonmyopic active search methods for drug and materials discovery.
problem Efficiently identifying many members of a class in high-throughput screening.
method Bayesian decision framework, nonmyopic approximations, efficient computation.
result Proposed policy outperforms sequential and batch settings in drug and materials discovery.
Improves sample efficiency in evolutionary policy search methods.
problem Worse sample efficiency in evolutionary policy search methods.
method Importance mixing to reuse more samples.
result Improves sample efficiency but still less than deep reinforcement learning.
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.
New method learns better branching policies for MILP problems.
problem Improving branch and bound search for solving MILP problems.
method Imitates strong branching rule with parameterized state of B&B search tree.
result Generalized policies outperform current state-of-the-art.
ProMP improves Meta-RL by better handling credit assignment.
problem Poor credit assignment in Meta-RL leads to suboptimal performance.
method Develops a novel meta-learning algorithm controlling pre-adaptation and adapted policies.
result Significantly improves sample-efficiency and asymptotic performance.
Greedy AutoAugment improves accuracy with less computation.
problem Finding effective data augmentation policies to cover the search space.
method Greedy approach to reduce the number of trials from exponential to linear growth.
result Greedy AutoAugment increases accuracy by 360 times with fewer resources.
FiDi-RL combines deep RL with FiDi policy search for efficient continuous control.
problem Challenges in continuous control tasks using reinforcement learning.
method Combines Deep Deterministic Policy Gradients (DDPG) with Augment Random Search (ARS) using FiDi policy search.
result Improves data efficiency and stability of ARS.
Combines evolutionary and RL methods for better policy search.
problem Low sample efficiency in evolutionary methods and high sensitivity in RL methods.
method Combines CEM and td3 algorithms for policy search.
result cem-rl offers a better trade-off between performance and sample efficiency.
A new algorithm solves sparse reward tasks efficiently in robotics.
problem Sparse or misleading rewards in reinforcement learning.
method Multi-objective model-based policy optimization with three objectives.
result Multi-DEX solves sparse reward scenarios in fewer episodes than existing methods.
The paper proposes a new policy training objective to reduce exploration in self-play.
problem Training policies to mimic MCTS search behavior can lead to excessive exploration.
method Derive a policy gradient expression using MCTS value estimates to reduce exploration.
result Empirically evaluated policies show reduced exploration compared to MCTS-based training.
AutoAugment learns optimal data augmentation policies automatically.
problem Improving image classifier accuracy through better data augmentation.
method AutoAugment uses a search algorithm to find the best augmentation policies in a defined search space.
result AutoAugment achieves state-of-the-art accuracy on multiple datasets.
The paper proposes an interpretable off-policy learning algorithm for medical treatments.
problem Lack of interpretable methods for personalized treatment decisions from observational data.
method Hyperbox search approach for interpretable policies in disjunctive normal form.
result The proposed algorithm outperforms state-of-the-art methods in terms of regret and is rated highly interpretable by clinical experts.
Scarce data is a major challenge to scaling robot learning to truly complex tasks, as we need to generalize locally learned policies over different "contexts". Bayesian optimization approaches to contextual policy search (CPS) offer data-efficient policy learning that generalize over a context space. We propose to impr…
Optimizes Thompson sampling policies using policy gradient methods.
problem Improving Thompson sampling in bandit problems.
method Applies policy gradient algorithms to optimize Thompson sampling policies.
result Direct policy search on Thompson sampling improves performance.
HPSDE learns hierarchical policies from multi-modal rewards.
problem Learning optimal policies from a multi-modal reward function.
method Density estimation with return-weighted importance sampling for identifying modes.
result HPSDE successfully learns option policies corresponding to modes of the return function.
A new imitation learning method uses random search for simple policies, outperforming complex models.
problem Complexity and reward dependency issues in imitation learning.
method Derivative-free optimization with simple linear policies and random search.
result The proposed method achieves competitive performance on MuJoCo locomotion tasks without a direct reward signal.
The paper proposes an iterative approach to batch reinforcement learning for safer and more informative data collection.
problem Learning policies that are too rigid and do not adapt to new data.
method Safe diversified model-based policy search in an iterative batch reinforcement learning framework.
result Improved learned policies through continuous data collection and adaptation.
RL teaches agents to make decisions through trial and error.
problem Sequential decision-making problems in dynamic environments.
method Value-based and policy search approaches, including actor-critic methods.
result RL can learn good policies from numeric reward feedback.
Paper uses CMAB to improve NAS efficiency and accuracy.
problem Improving efficiency and accuracy of NAS for DNNs.
method Formulated NAS as CMAB, used Nested Monte-Carlo Search.
result Discovered cell structure achieves comparable accuracy to state-of-the-art, 20x faster.
Paper proposes a new evaluation method for NAS search phase.
problem NAS search phase effectiveness not well evaluated.
method Compare NAS solutions with random selection; evaluate weight sharing strategy.
result State-of-the-art NAS algorithms perform similarly to random selection.
New algorithms improve reinforcement learning with multi-step greedy policies.
problem Difficulty in monotonic policy improvement with soft-policy updates.
method Formulated and analyzed online and approximate algorithms using multi-step greedy operators.
result Guaranteed monotonic policy improvement with sufficiently large update stepsize.