Paper introduces CRLMaze, a new benchmark for continual reinforcement learning in 3D non-stationary environments.
problem Challenges of training reinforcement learning agents in high-dimensional, always-changing environments.
method End-to-end model-free continual reinforcement learning strategy.
result Competitive results in a complex 3D non-stationary task, outperforming four baselines.
Agent learns to read maps and navigate mazes using deep reinforcement learning.
problem Teaching a machine to understand and navigate 3D environments from 2D maps.
method Combines A3C with a recurrent localization cell, learns localization from 3D images.
result Agent successfully navigates and localizes in random mazes, generalizing to larger mazes.
Google Research Football: A new 3D physics-based game for reinforcement learning.
problem Training reinforcement learning algorithms in complex, realistic environments.
method Developed a new 3D physics-based football simulator environment.
result Reported baseline results for various reinforcement algorithms.
Solves online 3D bin packing with deep reinforcement learning under constraints.
problem Challenges of packing items immediately without information and constraints.
method Constrained deep reinforcement learning (DRL) with feasibility predictor.
result Significantly outperforms state-of-the-art methods in online 3D bin packing.
Improved RL for 3D games with concurrent actions.
problem Training complex behaviors in 3D games.
method Deep reinforcement learning architecture with concurrent actions.
result 4x improvement in training time and 2.5x in performance.
Paper uses RL for high-level character control in 3D environments.
problem Creating intelligent characters with generalizable behavior.
method Combines traditional animations, heuristics, and reinforcement learning.
result Demonstrates learning of complex behaviors in a 3D environment.
New RL method designs 3D molecules with improved symmetry.
problem Lack of 3D information in molecular design.
method Symmetry-aware actor-critic architecture using spherical harmonics.
result Improves generalization and molecule quality.
Improved reinforcement learning for 3D games using SLAM and object detection.
problem Challenges in 3D game environments, especially partial observability and combinatorial spaces.
method Augmented Deep Q-Learning Network with SLAM and object detection for better policy learning.
result Our approach consistently learns better policies in 3D games like Doom.
High-throughput 3D control training system achieves 100,000 FPS.
problem Lack of efficient, single-machine reinforcement learning systems.
method Sample Factory combines asynchronous sampling and off-policy correction.
result Achieves 100,000 FPS on 3D control problems without sacrificing sample efficiency.
Modular RL modules solve complex 3D Sokoban tasks.
problem Solving complex, integrated tasks combining visual, physical, and abstract reasoning.
method Compose RL modules in a sense-plan-act hierarchy, using only model-free methods.
result Modular RL outperforms state-of-the-art monolithic RL on Mujoban.
Deep RL learns grasping from 2.5D images.
problem Grasping objects from 2.5D images.
method Deep Reinforcement Learning (DRL) in a simulated environment.
result Successfully learned grasping from 2.5D images.
ACTRCE uses natural language to improve reinforcement learning performance.
problem Sparse reward in reinforcement learning.
method Extends HER framework with natural language goal representation.
result ACTRCE can solve challenging 3D navigation tasks and generalize to unseen instructions.
Improves AI agents' 3D navigation by learning from failures and 3D spatial relationships.
problem Challenges in data efficiency, obstacle avoidance, and generalization in 3D visual navigation.
method Incorporates attention on 3D spatial relationships and a target skill extension module into DRL framework.
result Significantly improves navigation performance and generalization across targets and scenes.
This paper uses deep reinforcement learning to optimize UAV-assisted vehicular networks.
problem Optimizing UAV-assisted vehicular networks for efficient communication in smart cities.
method Formulated a Markov decision process (MDP) problem and solved it using deep deterministic policy gradient (DDPG) method.
result Proposed solutions maximize total throughput per unit energy and encourage UAV mobility.
Adversarial reinforcement learning optimizes microswimmers' path-planning in turbulent flows.
problem Optimizing microswimmers' paths in turbulent flows for efficient target reach.
method Adversarial-reinforcement learning scheme applied to 2D and 3D turbulent flows.
result Microswimmers can reach targets faster than a naive approach in turbulent flows.
A new RL method helps designers solve complex tasks.
problem Design process gap between problem and solution.
method Deep Reinforcement Learning (RL) for task-oriented design.
result Method achieves satisfactory design even with multiple goals.
VaST maps complex environments to simpler models for faster learning.
problem Sample inefficiency in deep reinforcement learning.
method Variational State Tabulation (VaST) and prioritized sweeping planning.
result VaST enables rapid learning and adaptation in complex tasks.
Randomized neural networks improve deep RL agents' generalization.
problem Deep RL agents struggle to generalize to new, semantically similar environments.
method Introduce a randomized (convolutional) neural network to perturb input observations.
result Significantly outperforms various regularization and data augmentation methods.
Deep RL framework teaches cars to drive autonomously.
problem Difficult to apply reinforcement learning to autonomous driving.
method Deep reinforcement learning with Recurrent Neural Networks and attention models.
result Framework successfully learned autonomous maneuvers in complex scenarios.
WILD-SCAV benchmarks AI in complex 3D FPS environments.
problem Lack of complexity and diversity in RL environments.
method Developed a 3D open-world FPS game environment.
result Demonstrates effectiveness in benchmarking RL algorithms.
Deep network learns Obstacle Tower challenge without human demonstrations.
problem Master procedurally generated levels that get progressively harder.
method Deep Reinforcement Learning with a simple feed-forward network.
result Performed competitively in a reinforcement learning competition.
Graphical physics network learns intuitive physics using deep reinforcement learning with intrinsic motivation.
problem Teaching intuitive physics to AI agents.
method Integrates deep reinforcement learning with intrinsic reward normalization for efficient learning.
result Agent effectively learns object positions and velocities using intrinsic motivation.
Paper simplifies deep RL for mobile robots on consumer hardware.
problem Training complex reasoning agents in 3D environments.
method Used ViZDoom for fast simulation and varied difficulty scenarios.
result Robust baseline agent architecture trained on consumer hardware.
Agent learns to interpret 3D instructions, generalizing from experience.
problem Grounded language learning in artificial agents.
method Simulated 3D environment, reinforcement and unsupervised learning.
result Agent learns to relate language to physical surroundings and actions.
Generative model learns spatial memory for predicting agent's future in complex environments.
problem Training generative and temporal models in partially observed, 3D environments is challenging.
method Action-conditioned generative model with non-parametric spatial memory and state-space model for dynamics.
result Scalable architecture capable of coherent predictions over hundreds of time steps in 2D and 3D environments.
Researchers classify and visualize 5-cube cubical surfaces.
problem Classifying and visualizing surfaces in a 5-dimensional cube.
method Exhaustive search, classification by genus and demigenus, 3D visualization, reinforcement learning for optimization.
result 2690 connected closed cubical surfaces in the 5-cube, visualized and optimized for 3D printing.
Paper proposes Q-learning for efficient aerial BS placement to improve fairness in mobile networks.
problem Optimal placement of aerial base stations to enhance fairness in a dynamic user mobility environment.
method Reinforcement learning approach to solve the NP-hard problem of 3D placement.
result Simulation results show increased fairness among users with a reasonable computing time and solution close to optimal.
ViZDoom competition tests AI in Doom using only visual input.
problem Creating AI bots that play Doom effectively from raw visual data.
method Multi-player deathmatch in Doom; bots make decisions based on screen buffer.
result Reinforcement learning can produce capable bots but not competitive against humans.
NEURO-DRAM improves neuroimaging classification accuracy.
problem Improper use of traditional computer vision models in neuroimaging.
method 3D recurrent visual attention model trained with reinforcement learning.
result NEURO-DRAM achieves state-of-the-art accuracy in Alzheimer's disease prediction.
Flatland is a simple 2D environment for reinforcement learning.
problem Testing and prototyping reinforcement learning agents.
method Simple, lightweight, customizable environment with physical properties.
result Baseline reinforcement learning agents can solve a navigation task in Flatland.
Model learns multiple tasks using visual and textual representations.
problem Training visual navigation agents for multiple tasks.
method Dual-Attention unit for task-invariant alignment of visual and textual representations.
result Model outperforms baselines on semantic goal navigation and embodied question answering.
New meta-RL method avoids exploration-exploitation trade-off.
problem Learning to explore and exploit simultaneously in meta-RL.
method Developed new objectives for exploration and exploitation.
result DREAM outperforms existing methods on complex tasks.
A new method learns to prioritize and use multiple views of an environment for better decision-making.
problem Learning from multiple views of an environment to improve decision-making.
method Attention-based deep reinforcement learning to dynamically attend to views of the environment.
result The method improves performance in complex 3D environments with obstacles.
Develops game theory framework for UAS integration into NAS.
problem Predicting outcomes of UAS integration into NAS.
method Game theory, reinforcement learning, level-k reasoning.
result Proposes a modeling framework for human pilot behavior.
ADVISOR dynamically balances imitation and reinforcement learning to overcome the imitation gap.
problem The gap between imitation learning and reinforcement learning when teaching agents have privileged information.
method Adaptive Insubordination (ADVISOR) dynamically weights imitation and reward-based reinforcement learning losses.
result On-the-fly switching with ADVISOR outperforms pure imitation, pure reinforcement learning, and their combinations.
New reinforcement learning framework for adapting to new actions.
problem Making reinforcement learning agents adaptable to new actions without retraining.
method Two-stage framework: infer action representations first, then train a flexible policy.
result Agents can make decisions from new action sets without retraining.
Paper proposes a more robust deep reinforcement learning agent.
problem Creating more robust reinforcement learning agents.
method Modified A3C algorithm with dual input streams.
result Significant reduction in training parameters (30%) with improved robustness.
Proposes a method to improve reinforcement learning in multi-view environments.
problem Improving policy learning in environments with partial observability.
method Actor-Critic-Attention Mechanism for generating a single feature representation from multiple views.
result Our method outperforms state-of-the-art baselines on various complex 3D environments.
Interactive machine learning improves deep RL in Minecraft by giving action advice.
problem Training deep RL agents in high-aliasing environments like Minecraft is computationally expensive.
method Conducted experiments with two RL algorithms, Feedback Arbitration, and Newtonian Action Advice, to give action advice to human teachers.
result Action advice from human teachers can improve agent performance in high-aliasing environments.
Generative models help agents form stable beliefs in complex environments.
problem Forming and maintaining stable beliefs in complex, dynamic environments.
method Train expressive generative models to predict multiple steps ahead.
result Expressive generative models improve data-efficiency in RL tasks.
A new RL method boosts exploration efficiency in Atari games and driving simulators.
problem Efficient exploration in reinforcement learning environments.
method A two-component approach: decaying schedule and upper quantile bonus.
result Significantly outperforms QR-DQN in Atari games and near-optimal safety in CARLA.
Improved model predicts future environment changes efficiently.
problem Efficiently predicting future changes in environments for agents.
method Recurrent neural networks for high-dimensional pixel observations, reducing computational load.
result Model can predict hundreds of time-steps into the future, improving exploration and adaptability.
Curiosity-driven agent predicts future state values to learn faster.
problem Sparse reward learning in reinforcement learning.
method The agent predicts future state values and uses consistency with current values as a regularization term.
result The extended agent learns significantly faster in sparse reward settings.
GENESIS generates and samples 3D scenes by capturing object interactions.
problem Lack of models that explicitly capture object interactions in scene generation.
method Object-centric latent variables, spatial GMM, amortized inference, autoregressive prior.
result First object-centric generative model of 3D visual scenes.
Director learns hierarchical behaviors from pixels, outperforming exploration methods.
problem Challenges in learning long-term hierarchical behaviors from pixels.
method Directly learns hierarchical behaviors from pixels using a latent space world model.
result Director outperforms exploration methods on various tasks.
The paper teaches a reinforcement learning agent to generate diverse programs based on symbolic instructions.
problem Learning to generate diverse programs for diverse scenes given a symbolic instruction.
method Instruction-conditioned reinforced adversarial learning.
result The agent's stochastic policy more accurately captures the diversity in the goal distribution.
Paper proposes a multi-task model for CECT macromolecule classification, segmentation, and recovery.
problem Challenges in recognizing and recovering macromolecular structures due to structural diversity and imaging limitations.
method A novel multi-task 3D CNN model that shares learned features across tasks.
result Multi-task model outperforms single-task methods and discovers novel structures.
LEAPS uses semantic models to improve reinforcement learning in diverse environments.
problem Generalizing and adapting to unseen environments in reinforcement learning.
method Hybrid model-based and model-free approach with a multi-target sub-policy and a Bayesian semantic model.
result LEAPS outperforms baselines in visual navigation tasks using diverse indoor scenes.