New method improves convergence of RL meta-learning.
problem Improving convergence in model-agnostic meta-reinforcement learning.
method Proposes Stochastic Gradient Meta-Reinforcement Learning (SG-MRL) to find ε-first-order stationary points. result Derives iteration and sample complexity for SG-MRL.
New complexity measure helps in agnostic reinforcement learning with or without access to MDP dynamics.
problem Understanding the number of rounds needed to learn an ε-suboptimal policy in unknown MDPs.
method Introducing spanning capacity as a new complexity measure and developing POPLER algorithm.
result There is a separation between generative and online access models for agnostic learnability.
New insights into RL with weak function approximation.
problem Statistical complexity of RL with function approximation in large state spaces.
method Agnostic policy learning framework, exploring environment access, coverage, and representational conditions.
result Characterization of fundamental performance bounds and statistical separations.
New framework promotes reproducible, domain-agnostic reinforcement learning algorithms.
problem Domain-specific, compute-resource-maximizing, and non-reproducible participant solutions in reinforcement learning competitions.
method Submission retraining, domain randomization, desemantization through domain obfuscation, and compute/environment-sample budget limitation.
result Participant submissions are reproducible, non-specific to the competition environment, and sample/resource efficient.
Proposes a new theoretical framework for PbRL that requires less human feedback.
problem Lack of theoretical work capturing practical PbRL frameworks.
method Introduces a reward-agnostic PbRL framework that acquires exploratory trajectories before human feedback.
result Demonstrates improved sample complexity for learning optimal policies in linear and low-rank MDPs.
Study agnostic RL in large state spaces with weak function approximation.
problem Statistical intractability of agnostic policy learning in various environments.
method Investigates agnostic policy learning with different forms of environment access.
result Agnostic policy learning remains statistically intractable with certain forms of environment access.
New algorithm finds optimal policies without knowing reward functions.
problem Reward-agnostic exploration in reinforcement learning.
method Designs an algorithm that explores without reward information, achieving minimax optimality.
result Achieves provable minimax optimality in finding optimal policies for multiple reward functions.
Improved agnostic boosting with better sample efficiency.
problem Agnostic boosting's poor sample efficiency compared to Empirical Risk Minimization.
method Leverages sample reuse across rounds, guarantees better generalization.
result Substantially more sample-efficient agnostic boosting algorithm.
Paper proposes a hybrid RL algorithm that combines offline and online data without needing reward info.
problem How to efficiently use online data to improve RL policies using only offline data.
method A three-stage hybrid RL algorithm that uses reward-agnostic exploration and model-based offline RL.
result The hybrid RL algorithm outperforms both pure offline and pure online RL in sample complexity.
Task-agnostic RL tackles exploration in MDPs with multiple tasks.
problem Challenges in reinforcement learning with multiple tasks or conflicting objectives.
method Task-agnostic RL framework, UCBZero algorithm.
result UCBZero finds near-optimal policies for multiple tasks efficiently.
A novel framework uses goal-conditioned reinforcement learning to generate diverse samples.
problem Generating high-quality, diverse samples from generative models.
method Two agents: GC-agent learns to reconstruct the training set, S-agent learns to imitate GC-agent without knowing the goals.
result Empirically, the method generates diverse and high-quality samples in image synthesis.
The composition of elementary behaviors to solve challenging transfer learning problems is one of the key elements in building intelligent machines. To date, there has been plenty of work on learning task-specific policies or skills but almost no focus on composing necessary, task-agnostic skills to find a solution to …
Framework for multi-agent RL with human feedback in a Snake game.
problem Improving multi-agent reinforcement learning with human feedback.
method Developed a simulated game environment for offline model training and online competitions. Introduced HILL methods and reward manipulation heuristics.
result Agents with HILL methods outperform those without in online competitions.
A deep RL approach generates counterfactual instances efficiently.
problem Efficient generation of counterfactual instances for large datasets and diverse models.
method Deep reinforcement learning to optimize counterfactual instances in a single forward pass.
result Model-agnostic and scalable counterfactual generation.
While model-based deep reinforcement learning (RL) holds great promise for sample efficiency and generalization, learning an accurate dynamics model is often challenging and requires substantial interaction with the environment. A wide variety of domains have dynamics that share common foundations like the laws of clas…
New RL algorithms learn policies competitive with best in class without assuming optimal policy.
problem Agnostic RL where best policy not necessarily optimal.
method First-order optimization in non-Euclidean space, reducing to policy learning.
result Sample complexity upper bounds for three algorithms under VGD condition.
Inverse reinforcement learning (IRL) enables an agent to learn complex behavior by observing demonstrations from a (near-)optimal policy. The typical assumption is that the learner's goal is to match the teacher's demonstrated behavior. In this paper, we consider the setting where the learner has its own preferences th…
This paper tackles online reinforcement learning for unseen tasks with unknown boundaries.
problem Real-world tasks violate assumptions of task distributions, independence, and clear task delineations.
method A mixture of Gaussian Processes models different dynamics, and a transition prior handles temporal dependencies.
result The approach reliably handles task distribution shifts and outperforms alternatives in non-stationary tasks.
Boosting with unlabeled data achieves optimal sample complexity in agnostic settings.
problem Boosting's sample inefficiency in agnostic learning.
method Designing an agnostic boosting algorithm with unlabeled data to match ERM's sample complexity.
result The total sample complexity is optimal, with a vanishing fraction needing to be labeled.
Learning to infer Bayesian posterior from a few-shot dataset is an important step towards robust meta-learning due to the model uncertainty inherent in the problem. In this paper, we propose a novel Bayesian model-agnostic meta-learning method. The proposed method combines scalable gradient-based meta-learning with non…
Physics-guided reinforcement learning optimizes swimming in turbulent flows.
problem Optimizing swimming efforts to maintain proximity in turbulent environments.
method Physics-informed actor-physicist reinforcement learning algorithm.
result Physics-informed reinforcement learning outperforms standard methods in turbulent flow control.
New method prevents forgetting in learning new tasks.
problem Learning new tasks without forgetting old knowledge.
method Hierarchical information-theoretic optimality principle and Mixture-of-Variational-Experts layer.
result Our method can operate task-agnostically and improve performance in various learning problems.
CUDC collects diverse data for offline RL by predicting future states.
problem Challenges in collecting task-agnostic data for offline RL.
method Adaptive temporal distances for curiosity-driven data collection.
result CUDC outperforms existing unsupervised methods in offline RL tasks.
CARML uses meta-learning to avoid obstacles in 2D vehicle navigation.
problem Collision avoidance in 2D vehicle navigation.
method Model-Agnostic Meta-Learning for multi-objective reinforcement learning.
result CARML outperforms a baseline TD3 solution in obstacle avoidance.
This study optimizes offline reinforcement learning methods for various tasks without rewards.
problem Optimizing offline reinforcement learning for multiple tasks without rewards.
method Designing a new model-based approach with singleton absorbing MDPs to achieve optimal convergence rates.
result Achieved optimal convergence rates for offline reinforcement learning in various settings.
In this paper we target the problem of transferring policies across multiple environments with different dynamics parameters and motor noise variations, by introducing a framework that decouples the processes of policy learning and system identification. Efficiently transferring learned policies to an unknown environme…
Meta-reinforcement learning improves fault-adaptive control efficiency.
problem Adaptive control under abrupt system faults with strict time constraints.
method Model-agnostic meta learning (MAML) with a fault library of prior policies.
result Improved sample efficiency and quick adaptation to new faults.
New RL approach speeds up training across tasks.
problem Training reinforcement learning models quickly and efficiently.
method Combines planning quasi-metric and task-specific aimers.
result Achieves multiple-fold speed-up on bit-flip and robotic arm tasks.
GeneraLight improves traffic signal control models' generalization ability.
problem Overfitting and lack of generalization ability in RL TSC models.
method GeneraLight uses a meta-RL framework with a traffic flow generator based on GANs.
result GeneraLight significantly boosts generalization performance across different traffic flows.
Algorithm learns to switch control among agents in a team.
problem Learning to switch control among reinforcement learning agents.
method 2-layer Markov decision process, upper confidence bounds, shared confidence bounds.
result Sublinear total regret with shared confidence bounds.
RL learns to ignore factors in factor investing portfolios.
problem Combining factor investing and reinforcement learning for optimal portfolio allocation.
method RL agent learns through sequential allocations based on firms' characteristics using Dirichlet distributions.
result RL-based portfolios are very close to equally-weighted allocations, indicating agnostic factor learning.
Model-agnostic meta-learning (MAML) is a meta-learning technique to train a model on a multitude of learning tasks in a way that primes the model for few-shot learning of new tasks. The MAML algorithm performs well on few-shot learning problems in classification, regression, and fine-tuning of policy gradients in reinf…
A fundamental problem in control is to learn a model of a system from observations that is useful for controller synthesis. To provide good performance guarantees, existing methods must assume that the real system is in the class of models considered during learning. We present an iterative method with strong guarantee…
Proposes PIC and POIC for measuring task difficulty in RL.
problem Lack of metrics to measure task difficulty in RL.
method Introduces policy information capacity (PIC) and policy-optimal information capacity (POIC) as metrics based on mutual information.
result Empirically shows PIC and POIC correlate with task solvability better than alternatives.
Unified policy controls diverse agents through modular neural networks.
problem Learning control policies for various agent morphologies.
method Shared Modular Policies (SMP) with decentralized control and message passing.
result A single modular policy controls multiple agent morphologies.
New approach makes deep reinforcement learning robust without assuming adversary knowledge.
problem Deep reinforcement learning policies are vulnerable to state observation perturbations.
method Proposes an adversary agnostic robust DRL paradigm using policy distillation with two terms: prescription gap maximization and Jacobian regularization.
result Boosts adversarial robustness on five Atari games compared to state-of-the-art methods.
Model-agnostic meta-learners aim to acquire meta-learned parameters from similar tasks to adapt to novel tasks from the same distribution with few gradient updates. With the flexibility in the choice of models, those frameworks demonstrate appealing performance on a variety of domains such as few-shot image classificat…
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…
SEEK algorithm selects minimal state in reinforcement learning for better policy learning.
problem Challenges in obtaining a state representation that is parsimonious and satisfies the Markov property.
method SEEK algorithm estimates the minimal sufficient state in reinforcement learning.
result The SEEK algorithm achieves selection consistency in large samples.
Deep RL learns effective job shop scheduling rules from raw features.
problem Designing effective priority dispatching rules for job shop scheduling is challenging.
method End-to-end deep reinforcement learning using Graph Neural Networks.
result Agent learns high-quality dispatching rules from raw features and generalizes well to unseen instances.
CODA resolves coordination issues in offline multi-agent reinforcement learning.
problem Coordination failure in offline multi-agent reinforcement learning.
method Diffusion-based multi-agent trajectory generator for data augmentation.
result CODA resolves coordination pathologies in continuous polynomial games and complex benchmarks.
Traditional reinforcement learning agents learn from experience, past or present, gained through interaction with their environment. Our approach synthesizes experience, without requiring an agent to interact with their environment, by asking the policy directly "Are there situations X, Y, and Z, such that in these sit…
First-order ANIL learns shared representations even with overparametrization.
problem Lack of theoretical evidence for model-agnostic meta-learning (ANIL) learning shared representations.
method First-order ANIL with a linear two-layer network architecture, showing asymptotically low-rank solutions with overparametrization.
result First-order ANIL learns linear shared representations, even with overparametrization, and performs well in adaptation.
CoinDICE estimates confidence intervals for unknown behavior policies in reinforcement learning.
problem Estimating value of a target policy using only behavior policy data.
method Function space embedding, generalized empirical likelihood method, Lagrangian optimization.
result Valid confidence intervals with tighter and more accurate estimates than existing methods.
We review basic concepts of convex duality, focusing on the very general and supremely useful Fenchel-Rockafellar duality. We summarize how this duality may be applied to a variety of reinforcement learning (RL) settings, including policy evaluation or optimization, online or offline learning, and discounted or undisco…
The paper analyzes MAML's representation using RSA, revealing that feature reuse is not the primary reason for its success.
problem Understanding why model-agnostic meta-learning (MAML) works well in few-shot learning tasks.
method Representation similarity analysis (RSA) applied to MAML's few-shot learning instantiation.
result Feature reuse is not the primary reason for MAML's success; instead, it is the learning task itself that increases representation similarity.
Paper introduces TrufLL for language model training without labeled data.
problem Training language models from scratch with reinforcement learning.
method Dynamic vocabulary truncation in RL for language generation.
result TrufLL successfully learns language generation policies from scratch.
In order to make good decision under uncertainty an agent must learn from observations. To do so, two of the most common frameworks are Contextual Bandits and Markov Decision Processes (MDPs). In this paper, we study whether there exist algorithms for the more general framework (MDP) which automatically provide the bes…