Since their introduction a year ago, distributional approaches to reinforcement learning (distributional RL) have produced strong results relative to the standard approach which models expected values (expected RL). However, aside from convergence guarantees, there have been few theoretical results investigating the re…
On-device research index
arXiv research
A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
168,657 papers · 148 categories
Trend · papers per month
48 results for “distributed RL”
New RL algorithm minimizes distributional learning error.
problem Improving distributional reinforcement learning for better error minimization.
method Proposes a new model-based algorithm with theoretical minimax optimality.
result Proves minimax optimality for approximating return distributions.
This work bridges offline RL and DRL to address distributional shift.
problem Distributional shift in offline RL due to difference in state-action visitation distributions.
method Proposes offline RL algorithms using DRL framework, characterizes sample complexity under single policy concentrability.
result Demonstrates superior performance of proposed algorithms through simulations.
This work develops efficient methods for continuous-time distributional reinforcement learning.
problem Continuous-time reinforcement learning with return distributions.
method Parameterizing return distributions using quantile representation and showing topological properties.
result Efficient approximation algorithm for continuous-time distributional reinforcement learning.
Distributional Reinforcement Learning (RL) differs from traditional RL in that, rather than the expectation of total returns, it estimates distributions and has achieved state-of-the-art performance on Atari Games. The key challenge in practical distributional RL algorithms lies in how to parameterize estimated distrib…
LaPSRL achieves optimal regret for isoperimetric RL distributions.
problem Designing RL algorithms with sublinear regret for non-log-concave distributions.
method Posterior Sampling (PSRL) and Langevin sampling (LaPSRL) for isoperimetric distributions.
result LaPSRL achieves order-optimal regret and subquadratic complexity.
Offline RL with pre-trained features amplifies errors even under mild shifts.
problem Sample-efficient offline RL with pre-trained features under mild distribution shift.
method Empirical study of offline RL with pre-trained neural representations.
result Substantial error amplification occurs even with pre-trained features, requiring stronger conditions for successful offline RL.
This paper shows RL with KL penalties is equivalent to Bayesian inference for fine-tuning LMs.
problem Fine-tuning large language models to avoid undesirable features.
method Analyzed KL-regularized RL and showed it's equivalent to variational inference.
result KL-regularized RL avoids distribution collapse and is more insightful as Bayesian inference.
Despite the recent progress in deep reinforcement learning field (RL), and, arguably because of it, a large body of work remains to be done in reproducing and carefully comparing different RL algorithms. We present catalyst.RL, an open source framework for RL research with a focus on reproducibility and flexibility. Ma…
State-constrained offline RL expands RL's learning scope.
problem Batch-constrained offline RL limits policies to seen actions.
method Introduces state-constrained offline RL focusing on state distribution.
result Policy can take high-quality out-of-distribution actions.
Catalyst.RL accelerates RL research with efficient training.
problem Efficient reinforcement learning training in complex environments.
method Open-source PyTorch framework with distributed training and RL algorithms.
result Catalyst.RL achieved 2nd place in a computationally expensive RL challenge.
Unified framework for reliable uncertainty quantification in RL.
problem Uncertainty quantification in high-stakes reinforcement learning.
method Unified conformal prediction framework integrating distributional RL and conformal calibration.
result Significantly improved coverage and reliability over standard methods.
Simpler one-step distributional RL framework for control.
problem Lack of a unified theory for DistrRL in control.
method One-step distributional reinforcement learning (OS-DistrRL) framework.
result Unified theory for policy evaluation and control.
Paper introduces a new robust loss function for RL.
problem Heuristic selection of threshold parameters in quantile Huber loss.
method Derived from Wasserstein distance, captures noise in quantile values.
result Enhances robustness against outliers and enables parameter adjustment.
Paper proposes SinkhornDRL for distributional RL using Sinkhorn divergence and regularized Wasserstein loss.
problem Improving distributional reinforcement learning by minimizing Bellman return distribution differences.
method Introduces SinkhornDRL, a distributional RL algorithm using Sinkhorn divergence and regularized Wasserstein loss.
result SinkhornDRL consistently outperforms or matches existing algorithms on Atari games, especially in multi-dimensional reward settings.
Fiber simplifies RL and population-based methods for distributed training.
problem Challenges in RL and population-based methods, including frequent interaction with simulations and dynamic scaling.
method Introducing Fiber, a scalable distributed computing framework.
result Significantly expands accessibility of large-scale parallel computation.
Semi-pessimistic RL tackles distributional shift and data scarcity in offline RL.
problem Distributional shift and scarcity of labeled data in offline RL.
method Proposes a semi-pessimistic RL method that simplifies learning by seeking a lower bound of the reward function.
result Demonstrates clear competitiveness and improved policy learning with vast unlabeled data.
This paper studies the potential of the return distribution for exploration in deterministic reinforcement learning (RL) environments. We study network losses and propagation mechanisms for Gaussian, Categorical and Gaussian mixture distributions. Combined with exploration policies that leverage this return distributio…
MOPO optimizes offline RL by penalizing dynamics uncertainty.
problem Learning policies from offline data with distributional shift.
method Modify model-based RL to avoid distributional shift.
result MOPO outperforms model-free and standard model-based RL.
Method learns statistics of return distributions via neural networks and maximum mean discrepancy.
problem Learning probability distributions in reinforcement learning.
method Maximum mean discrepancy (MMD) for learning unrestricted statistics of return distributions.
result Method outperforms standard distributional RL baselines on Atari games.
Deep reinforcement learning (RL) has achieved breakthrough results on many tasks, but agents often fail to generalize beyond the environment they were trained in. As a result, deep RL algorithms that promote generalization are receiving increasing attention. However, works in this area use a wide variety of tasks and e…
New method improves meta-reinforcement learning efficiency.
problem Sample inefficiency in meta-reinforcement learning.
method Hindsight Foresight Relabeling (HFR) method.
result HFR improves performance on various meta-reinforcement learning tasks.
CQL learns conservative Q-functions to improve offline RL performance.
problem Leveraging large, static datasets in reinforcement learning without further interaction.
method Conservative Q-learning (CQL) which learns a conservative Q-function to lower-bound policy values.
result CQL substantially outperforms existing offline RL methods, often achieving 2-5 times higher final returns.
The standard reinforcement learning (RL) formulation considers the expectation of the (discounted) cumulative reward. This is limiting in applications where we are concerned with not only the expected performance, but also the distribution of the performance. In this paper, we introduce micro-objective reinforcement le…
This paper explains why distributional reinforcement learning is better than vanilla RL using small-loss bounds.
problem Understanding when and why distributional reinforcement learning (DistRL) is superior to vanilla reinforcement learning (RL).
method The paper uses small-loss bounds to explain the benefits of DistRL, proposing algorithms and proving bounds for different RL settings.
result Distributional reinforcement learning (DistRL) outperforms vanilla RL when optimal costs are small, as shown by small-loss bounds.
Recently, reinforcement learning (RL) algorithms have demonstrated remarkable success in learning complicated behaviors from minimally processed input. However, most of this success is limited to simulation. While there are promising successes in applying RL algorithms directly on real systems, their performance on mor…
CQL (ReDS) learns from varied driving behaviors, improving offline RL performance.
problem Learning from datasets with non-uniform variability in behavior policies.
method Reweighting the data distribution to allow per-state flexibility in following the behavior policy.
result CQL (ReDS) improves performance in various offline RL tasks.
Paper establishes DRL for high-dimensional rewards.
problem Intractable reinforcement learning with high-dimensional rewards.
method Theoretical foundations and a novel DRL algorithm.
result Bellman operator contraction in high-dimensional spaces.
A simple approach to offline RL without additional complexity.
problem Learning from a fixed dataset of actions with value estimation errors.
method Adding a behavior cloning term to the policy update of an online RL algorithm and normalizing the data.
result Matches the performance of state-of-the-art offline RL algorithms with minimal changes.
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.
Entrocraft addresses RL performance saturation in LLMs by customizing entropy curves.
problem Performance saturation in RL algorithms for LLMs.
method Entrocraft uses rejection sampling to bias advantage distributions for customized entropy schedules.
result Entrocraft significantly improves generalization, output diversity, and long-term training in 4B models.
GAN Q-learningstat.ML
Distributional reinforcement learning (distributional RL) has seen empirical success in complex Markov Decision Processes (MDPs) in the setting of nonlinear function approximation. However, there are many different ways in which one can leverage the distributional approach to reinforcement learning. In this paper, we p…
On Practical Reinforcement Learning: Provable Robustness, Scalability, and Statistical Efficiencycs.LG
This thesis improves practical reinforcement learning methods with robustness, scalability, and efficiency.
problem Improving reinforcement learning methods for practical applications.
method Analyzes and develops robust, scalable, and efficient reinforcement learning algorithms.
result Proves the efficiency and robustness of new RL methods.
Gradient-based meta-learners such as Model-Agnostic Meta-Learning (MAML) have shown strong few-shot performance in supervised and reinforcement learning settings. However, specifically in the case of meta-reinforcement learning (meta-RL), we can show that gradient-based meta-learners are sensitive to task distributions…
New algorithms for multivariate RL improve decision-making in complex systems.
problem Complex multi-objective decision-making in reinforcement learning.
method Oracle-free and computationally-tractable algorithms for multivariate distributional RL.
result Convergence rates match scalar reward settings and provide insights into reward dimensionality.
New insights show coverage conditions are crucial for efficient online reinforcement learning.
problem The role of coverage conditions in determining sample complexity of offline reinforcement learning.
method Established a connection between coverage conditions and sample efficiency in online reinforcement learning.
result Coverability, a structural property of MDPs, enables sample-efficient exploration in online reinforcement learning.
New algorithm for RL using mean embeddings of return distributions.
problem Improving reinforcement learning algorithms for dynamic programming.
method Mean embeddings of return distributions, novel algorithms for RL.
result Asymptotic convergence and improved performance in deep RL.
A common problem in Machine Learning and statistics consists in detecting whether the current sample in a stream of data belongs to the same distribution as previous ones, is an isolated outlier or inaugurates a new distribution of data. We present a hierarchical Bayesian algorithm that aims at learning a time-specific…
Robust RL improves controller robustness to dynamics variations using adversarial populations.
problem Robustness issues in RL when dynamics are perturbed.
method Adversarial population augmentation to the Robust RL formulation.
result Population-based adversarial approach yields more robust and generalizable policies.
Time-variant value function transfer method for RL.
problem Transfer learning with time-variant task distributions.
method Variational approach leveraging temporal structure.
result The proposed method outperforms time-invariant approach in experiments.
A new method improves actor-critic RL by integrating HMC, enhancing policy distribution and exploration.
problem Actor-critic RL yields suboptimal policies due to amortization gap and insufficient exploration.
method Integrating Hamiltonian Monte Carlo (HMC) into the actor-critic RL framework.
result Improves policy distribution and exploration, leading to better policy estimates and higher returns.
DRL agents perform poorly at high decision frequencies, but a new algorithm improves performance.
problem DRL agents struggle at high decision frequencies, leading to poor performance.
method Proved that DRL agents' action-conditioned return distributions collapse to their policy's return distribution as decision frequency increases. Defined superiority as a probabilistic generalization of advantage for high-frequency value-based RL.
result Proper modeling of superiority distribution improves performance of controllers at high decision frequencies.
Understanding Reinforcement Learning-Based Fine-Tuning of Diffusion Models: A Tutorial and Reviewcs.LG
This tutorial reviews RL-based methods for optimizing diffusion models to maximize specific metrics.
problem Optimizing diffusion models to generate samples that maximize specific metrics in practical applications.
method Various RL algorithms including PPO, differentiable optimization, reward-weighted MLE, value-weighted sampling, and path consistency learning.
result Exploration of strengths and limitations of RL-based fine-tuning algorithms and their benefits compared to non-RL-based approaches.
Simplifies BCQ to match and outperform state-of-the-art in offline RL benchmarks.
problem Sample efficiency in offline reinforcement learning.
method Introduces EMaQ, a novel backup operator for offline RL.
result EMaQ matches and outperforms prior state-of-the-art in offline RL benchmarks.
This paper sets communication complexity bounds for distributed RL.
problem Establishing minimum communication requirements for distributed RL.
method Information-theoretic lower bounds and algorithm development.
result Developed algorithms achieving optimal risk up to logarithmic factors.
This paper deals with distributed policy optimization in reinforcement learning, which involves a central controller and a group of learners. In particular, two typical settings encountered in several applications are considered: multi-agent reinforcement learning (RL) and parallel RL, where frequent information exchan…
NetHack Learning Environment (NLE) tests RL algorithms, offering scalable, complex, and challenging gameplay.
problem Challenging environments for testing RL algorithms.
method Procedurally generated, stochastic, rich, and complex NetHack environment.
result Demonstrates empirical success for early stages of NetHack using RL.
Most kernel-based methods, such as kernel or Gaussian process regression, kernel PCA, ICA, or -means clustering, do not scale to large datasets, because constructing and storing the kernel matrix requires at least time and space for samples. Recent works show that sampling point…