A novel approach learns goal-conditioned policies for locomotion using batch RL.
problem Training goal-conditioned policies for rotation invariant locomotion.
method Data augmentation and Siamese framework for invariance.
result Our approach outperforms existing RL algorithms on 3D locomotion agents.
Generative neural nets learn deep policies conditioned on goals.
problem Learning optimal policies for specific goals in reinforcement learning.
method Goal-conditioned neural nets that generate deep neural policies.
result Single learned policy generator can achieve any desired return.
A new EM framework for goal-conditioned RL improves performance on sparse reward tasks.
problem Handling sparse rewards in goal-conditioned reinforcement learning.
method A graphical model framework with an EM algorithm that includes a learning-in-hindsight E-step and a supervised M-step.
result hEM significantly outperforms model-free baselines on goal-conditioned benchmarks with sparse rewards.
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.
We propose a novel framework for multi-task reinforcement learning (MTRL). Using a variational inference formulation, we learn policies that generalize across both changing dynamics and goals. The resulting policies are parametrized by shared parameters that allow for transfer between different dynamics and goal condit…
Method solves long-horizon robotic tasks via imitation and reinforcement learning.
problem Long-horizon robotic tasks with complex sequences of actions.
method Two-phase approach: imitation learning followed by reinforcement learning.
result Method can scale to challenging long-horizon tasks.
Goal-conditioned policies are used in order to break down complex reinforcement learning (RL) problems by using subgoals, which can be defined either in state space or in a latent feature space. This can increase the efficiency of learning by using a curriculum, and also enables simultaneous learning and generalization…
Sparse reward problems are one of the biggest challenges in Reinforcement Learning. Goal-directed tasks are one such sparse reward problems where a reward signal is received only when the goal is reached. One promising way to train an agent to perform goal-directed tasks is to use Hindsight Learning approaches. In thes…
The paper introduces an adjacency constraint to improve goal-conditioned HRL.
problem Training inefficiency in goal-conditioned HRL due to large action space.
method Restricting the high-level action space to a k-step adjacent region of the current state.
result The adjacency constraint preserves optimal hierarchical policies and improves HRL performance.
Physics-informed GCRL tackles sparse feedback learning with hybrid dynamics.
problem Sparse feedback learning with high-dimensional, hybrid, or contact-dependent dynamics.
method Introduces physics-informed inductive biases into goal-conditioned value learning.
result Contact-rich manipulation tasks degrade existing Pi-GCRL methods.
Agent learns causal relationships from visual data to perform tasks.
problem Performing tasks in novel environments with latent causal structures.
method Learning-based approach to induce causal graphs from visual observations, using attention mechanisms.
result Effective generalization to new tasks with unseen causal structures.
Efficient exploration is necessary to achieve good sample efficiency for reinforcement learning in general. From small, tabular settings such as gridworlds to large, continuous and sparse reward settings such as robotic object manipulation tasks, exploration through adding an uncertainty bonus to the reward function ha…
DAGR improves navigation by refining goal representations conditioned on the current state.
problem Goal-conditioned reinforcement learning lacks state awareness, leading to inefficient policy recovery.
method DAGR refines static goal embeddings into state-conditioned ones using gated cross-attention with a state-goal discrepancy map.
result DAGR improves navigation tasks on OGBench, matching or outperforming base methods.
A method uses RL to learn abstractions for planning, improving robot navigation and manipulation tasks.
problem Planning requires suitable abstractions for states and transitions, which RL struggles with for temporally extended tasks.
method Goal-conditioned policies learned with RL are incorporated into planning, with a latent variable model representing valid states.
result Our method significantly outperforms prior work on image-based robot navigation and manipulation tasks.
In this paper we consider self-supervised representation learning to improve sample efficiency in reinforcement learning (RL). We propose a forward prediction objective for simultaneously learning embeddings of states and action sequences. These embeddings capture the structure of the environment's dynamics, enabling e…
This work improves imitation learning and goal-conditioned RL by estimating value densities.
problem Effective solutions for imitation and goal-conditioned reinforcement learning require reliably reaching specified states or demonstrations.
method The approach uses recent advances in density estimation to learn value functions efficiently and without hindsight bias.
result The method achieves state-of-the-art demonstration sample-efficiency in imitation learning and is both efficient and bias-free in goal-conditioned reinforcement learning.
Improved exploration algorithm for unknown MDPs with reduced sample complexity.
problem Exploration of unknown environments without reward function.
method Incremental model-based approach that interleaves state discovery and policy improvement.
result Achieves sample complexity scaling as O~(L5SL+εΓL+εAε−2). What is a good exploration strategy for an agent that interacts with an environment in the absence of external rewards? Ideally, we would like to get a policy driving towards a uniform state-action visitation (highly exploring) in a minimum number of steps (fast mixing), in order to ease efficient learning of any goal-…
ESPD improves learning efficiency in sparse reward reinforcement learning.
problem Sparse reward reinforcement learning challenges.
method Evolutionary Stochastic Policy Distillation (ESPD) based on drifted random walk insight.
result High learning efficiency demonstrated in MuJoCo robotics control suite experiments.
Paper analyzes neural network complexity for planning problems.
problem Understanding neural network complexity for planning policies.
method Circuit complexity analysis for relational neural networks.
result Three classes of planning problems identified based on network complexity.
Generically learns movement control policies from exploration data.
problem Movement optimization in physically based characters.
method Parameterizes actions as target states, learns low-level control policy.
result Improves movement optimization across multiple tasks and algorithms.
Success conditioning optimizes policies by imitating successful trajectories, solving a trust-region optimization problem.
problem Improving policies through random actions that lead to desired outcomes.
method Success conditioning, which involves collecting and updating policies based on successful trajectories.
result Success conditioning solves a trust-region optimization problem, maximizing policy improvement with a χ2 divergence constraint. This work tackles long-term visual planning by goal-conditioned hierarchical predictors.
problem Current learning approaches fail on long-horizon tasks due to lack of goal information and coarse-to-fine planning.
method Formulate goal-conditioned predictors (GCPs) and hierarchical models to predict trajectories between observations.
result GCPs enable effective long-term planning with much longer horizons than before.
New method tackles parcel routing with AI.
problem Routing parcels efficiently through a network of hubs.
method Combines graph neural networks with model-free RL.
result Extracts small feature graphs from the environment state.
Improved cooperation between levels boosts reinforcement learning performance.
problem Training multi-level policies in hierarchical reinforcement learning.
method Modeling policy optimization as a multi-agent process and inducing cooperation between sub-policies.
result Inducing cooperation between sub-policies leads to stronger and more sample-efficient policies.
In Multi-Goal Reinforcement Learning, an agent learns to achieve multiple goals with a goal-conditioned policy. During learning, the agent first collects the trajectories into a replay buffer, and later these trajectories are selected randomly for replay. However, the achieved goals in the replay buffer are often biase…
A central challenge in reinforcement learning is discovering effective policies for tasks where rewards are sparsely distributed. We postulate that in the absence of useful reward signals, an effective exploration strategy should seek out {\it decision states}. These states lie at critical junctions in the state space …
Model-Based Offline Planning (MBOP) learns models from offline data to control systems directly.
problem Training RL policies from offline data without direct system interaction.
method Generates models from offline data and uses planning to control the system.
result Near-optimal policies found for simulated systems with minimal real-time interaction.
Decouples critic chunk length from policy to improve policy reactivity and performance.
problem Bootstrapping bias and difficulty in extracting optimal policies from chunked critics.
method Optimizes policy against a distilled critic for partial action chunks, allowing shorter chunks for policy.
result Reliably outperforms prior methods on long-horizon offline goal-conditioned tasks.
For an autonomous agent to fulfill a wide range of user-specified goals at test time, it must be able to learn broadly applicable and general-purpose skill repertoires. Furthermore, to provide the requisite level of generality, these skills must handle raw sensory input such as images. In this paper, we propose an algo…
This study analyzes convergence and stability of reinforcement learning algorithms.
problem Understanding the conditions under which reinforcement learning algorithms converge and remain stable.
method Theoretical analysis of convergence and stability of Episodic Upside-Down Reinforcement Learning, Goal-Conditioned Supervised Learning, and Online Decision Transformers.
result The algorithms can achieve near-optimal behavior if the transition kernel is close to a deterministic kernel.
Eikonal-Constrained QRL improves goal-reaching in reinforcement learning.
problem Reward design and out-of-distribution generalization in reinforcement learning.
method Eikonal-Constrained Quasimetric Reinforcement Learning (Eik-QRL) using the Eikonal PDE.
result Eik-QRL achieves state-of-the-art performance in offline goal-conditioned navigation and manipulation tasks.
HiDe learns hierarchical control for complex tasks by separating planning and control.
problem Solving long horizon control tasks with generalization to unseen scenarios.
method Functional decomposition of state-action spaces, RL-based planner, modular transfer of policy layers.
result Generalizes across unseen test environments and scales to longer horizons.
The current dominant paradigm for imitation learning relies on strong supervision of expert actions to learn both 'what' and 'how' to imitate. We pursue an alternative paradigm wherein an agent first explores the world without any expert supervision and then distills its experience into a goal-conditioned skill policy …
Learning to control an environment without hand-crafted rewards or expert data remains challenging and is at the frontier of reinforcement learning research. We present an unsupervised learning algorithm to train agents to achieve perceptually-specified goals using only a stream of observations and actions. Our agent s…
This paper proposes a new method to connect language and physical actions in reinforcement learning.
problem Connecting linguistic representations to the physical world in embodied agents.
method Language-conditioned goal generators to decouple sensorimotor learning from language acquisition.
result Agents can demonstrate a diversity of behaviors for any given instruction.
Pre-trained language model boosts RL efficiency.
problem Low sample efficiency in RL, especially in lifelong learning.
method Use a pre-trained task-independent language model for transfer learning.
result Goal-conditional RL agents become more sample efficient.
UDRL fails to converge in stochastic environments with episodic resets.
problem UDRL's convergence in stochastic environments with resets is questioned.
method UDRL is a supervised learning approach that does not use value functions.
result UDRL diverges in a simple stochastic environment with resets.
Designing rewards for Reinforcement Learning (RL) is challenging because it needs to convey the desired task, be efficient to optimize, and be easy to compute. The latter is particularly problematic when applying RL to robotics, where detecting whether the desired configuration is reached might require considerable sup…
Agent learns goals and rewards through language and curiosity.
problem Autonomous agents lack intrinsic motivations and reward functions.
method LE2 algorithm using NL interactions and intrinsic motivations.
result Agent autonomously discovers and grounds goals in real behavior.
Representation learning is a central challenge across a range of machine learning areas. In reinforcement learning, effective and functional representations have the potential to tremendously accelerate learning progress and solve more challenging problems. Most prior work on representation learning has focused on gene…
In many real-world scenarios, an autonomous agent often encounters various tasks within a single complex environment. We propose to build a graph abstraction over the environment structure to accelerate the learning of these tasks. Here, nodes are important points of interest (pivotal states) and edges represent feasib…
AMIGo uses adversarial intrinsic goals to teach RL agents new skills.
problem Learning in sparse reward environments.
method Adversarial intrinsic goals to generate a curriculum for a student policy.
result AMIGo enables agents to learn new skills without extrinsic rewards.
CoDA augments data with counterfactuals from local causal structures.
problem Improving sample efficiency in RL with complex dynamic processes.
method Local causal models (LCMs) and Counterfactual Data Augmentation (CoDA).
result CoDA significantly improves RL agent performance in locally factored tasks.
KEMP predicts long-term trajectories for autonomous driving using keyframes.
problem Predicting future trajectories of road agents for autonomous driving.
method Keyframe-based hierarchical end-to-end deep learning framework.
result Ranked 1st on Waymo Open Motion Dataset Leaderboard.
Reinforcement learning algorithms use correlations between policies and rewards to improve agent performance. But in dynamic or sparsely rewarding environments these correlations are often too small, or rewarding events are too infrequent to make learning feasible. Human education instead relies on curricula--the break…
Plan2Vec learns image representations without labels, improving control tasks.
problem Learning image representations without labeled data.
method Constructs a weighted graph using near-neighbor distances and extrapolates to global embedding.
result Plan2Vec achieves accurate long-term value estimates in control tasks with reduced computational and memory costs.
Conventionally, model-based reinforcement learning (MBRL) aims to learn a global model for the dynamics of the environment. A good model can potentially enable planning algorithms to generate a large variety of behaviors and solve diverse tasks. However, learning an accurate model for complex dynamical systems is diffi…