This work tackles force control for contact-rich manipulation tasks with rigid robots using RL.
problem Challenges in working with real robotic hardware, especially position-controlled robots.
method Combines RL with traditional force control techniques, implementing parallel position/force control and admittance control.
result Validated methods on both simulation and real robot (UR3 e-series) for force control.
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.
A new framework enables real-time task trade-off control.
problem Conflict between multiple related tasks in a fixed model capacity.
method Formulates MTL as a preference-conditioned multiobjective optimization problem; uses a hypernetwork-based neural network.
result A single model can handle different trade-off preferences among multiple tasks.
Develops methods to control risk in ordinal classification tasks.
problem Controlling risk in ordinal classification tasks.
method Formulated ordinal classification in conformal risk control framework, proposed loss functions and algorithms.
result Demonstrated effectiveness and analyzed differences in risk control methods.
Reliable and effective multi-task learning is a prerequisite for the development of robotic agents that can quickly learn to accomplish related, everyday tasks. However, in the reinforcement learning domain, multi-task learning has not exhibited the same level of success as in other domains, such as computer vision. In…
Weakly-supervised RL identifies meaningful tasks, improving performance in complex environments.
problem Learning to efficiently explore and distinguish between meaningful and irrelevant tasks.
method Weak supervision to automatically disentangle meaningful tasks from a large space of nonsensical tasks.
result The learned subspace of meaningful tasks leads to substantial performance gains, especially in complex environments.
A language for specifying complex reinforcement learning tasks.
problem Challenges in specifying and shaping reward functions for complex reinforcement learning tasks.
method Proposes a new language and algorithm for automatically generating and shaping reward functions.
result SPECTRL tool outperforms state-of-the-art baselines.
Learning robot tasks or controllers using deep reinforcement learning has been proven effective in simulations. Learning in simulation has several advantages. For example, one can fully control the simulated environment, including halting motions while performing computations. Another advantage when robots are involved…
Using reinforcement learning to learn control policies is a challenge when the task is complex with potentially long horizons. Ensuring adequate but safe exploration is also crucial for controlling physical systems. In this paper, we use temporal logic to facilitate specification and learning of complex tasks. We combi…
This paper presents a reinforcement learning approach to synthesizing task-driven control policies for robotic systems equipped with rich sensory modalities (e.g., vision or depth). Standard reinforcement learning algorithms typically produce policies that tightly couple control actions to the entirety of the system's …
Study uses DRL with Lagrangian relaxation to solve temporal control tasks with STL constraints.
problem Optimal control problems with temporal logic constraints.
method Extended CMDP formulation, Lagrangian relaxation, two-phase constrained DRL algorithm.
result Demonstrated learning performance of the proposed algorithm through simulations.
Many continuous control tasks have easily formulated objectives, yet using them directly as a reward in reinforcement learning (RL) leads to suboptimal policies. Therefore, many classical control tasks guide RL training using complex rewards, which require tedious hand-tuning. We automate the reward search with AutoRL,…
A new approach predicts next observations without explicit decoding for better control.
problem High-dimensional observations and unknown dynamics in real-world control tasks.
method Proposes a novel information-theoretic LCE approach using predictive coding to develop a decoder-free model.
result The model reliably learns a controllable latent space leading to superior performance.
Meta-CVs leverage task similarity to reduce variance with limited data.
problem Reducing variance in Monte Carlo estimators with few samples.
method Meta-learning control variates for related tasks.
result Meta-CVs lead to significant variance reduction in settings with limited data.
Predicting outcomes and planning interactions with the physical world are long-standing goals for machine learning. A variety of such tasks involves continuous physical systems, which can be described by partial differential equations (PDEs) with many degrees of freedom. Existing methods that aim to control the dynamic…
In dynamic environments, learned controllers are supposed to take motion into account when selecting the action to be taken. However, in existing reinforcement learning works motion is rarely treated explicitly; it is rather assumed that the controller learns the necessary motion representation from temporal stacks of …
Multi-task learning uses auxiliary data or knowledge from relevant tasks to facilitate the learning in a new task. Multi-task optimization applies multi-task learning to optimization to study how to effectively and efficiently tackle multiple optimization problems simultaneously. Evolutionary multi-tasking, or multi-fa…
New RL algorithms improve control tasks with data reuse.
problem Real-world control requires performance guarantees and data efficiency.
method Generalized Policy Improvement combining on-policy guarantees and sample reuse.
result Extensive experimental analysis shows benefits of new algorithms.
Curriculum learning and imitation learning improve control over financial time-series data.
problem Improving control performance over complex financial time-series data.
method Data augmentation for curriculum learning and policy distillation for imitation learning.
result Curriculum learning shows significant improvement over time-series control tasks.
Anticipatory model generates music with control over events.
problem Controlling symbolic music generation.
method Interleaving event and control sequences to predict future events.
result Anticipatory model matches autoregressive models in performance and can infill control tasks.
We present a method for fast training of vision based control policies on real robots. The key idea behind our method is to perform multi-task Reinforcement Learning with auxiliary tasks that differ not only in the reward to be optimized but also in the state-space in which they operate. In particular, we allow auxilia…
Improved CEM for fast real-time planning in high-dimensional control tasks.
problem Sampling inefficiency of CEM in real-time planning.
method Novel additions to CEM including temporally-correlated actions and memory.
result 2.7-22x less samples and 1.2-10x performance increase.
Paper identifies objective mismatch in MBRL, affecting control task performance.
problem Objective mismatch in MBRL framework affects control task performance.
method Proposes re-weighting dynamics model training to mitigate mismatch.
result Likelihood of one-step ahead predictions is not always correlated with control performance.
A new framework for controllable generation of discrete masked models.
problem Efficient controllable generation of discrete data models.
method Plug-and-play framework based on importance sampling.
result Demonstrates versatility across multiple domains, including protein design.
Develops a method to plan exploration that learns strong policies with fewer samples.
problem Lack of efficient exploration in reinforcement learning for real-world tasks.
method Plans an action sequence that maximizes information gain about the optimal trajectory.
result 2x fewer samples than exploration baselines and 200x fewer than model-free methods.
Unified causal model improves controllable text generation without bias.
problem Controllable text generation tasks, biased by prior models.
method Unified causal framework for attribute-conditional generation and text attribute transfer.
result Significant superiority over previous conditional models for improved control and reduced bias.
In partially observable (PO) environments, deep reinforcement learning (RL) agents often suffer from unsatisfactory performance, since two problems need to be tackled together: how to extract information from the raw observations to solve the task, and how to improve the policy. In this study, we propose an RL algorith…
Catastrophic forgetting continues to severely restrict the learnability of controllers suitable for multiple task environments. Efforts to combat catastrophic forgetting reported in the literature to date have focused on how control systems can be updated more rapidly, hastening their adjustment from good initial setti…
We propose a neural information processing system which is obtained by re-purposing the function of a biological neural circuit model, to govern simulated and real-world control tasks. Inspired by the structure of the nervous system of the soil-worm, C. elegans, we introduce Neuronal Circuit Policies (NCPs), defined as…
Framework calibrates ML models for risk control in various tasks.
problem Achieving statistical guarantees for model predictions.
method Reframing risk control as multiple hypothesis testing, applying statistical techniques.
result New calibration methods for multi-label classification, instance segmentation, outlier detection, and confidence set coverage.
Distribution grids are currently challenged by frequent voltage excursions induced by intermittent solar generation. Smart inverters have been advocated as a fast-responding means to regulate voltage and minimize ohmic losses. Since optimal inverter coordination may be computationally challenging and preset local contr…
A decentralized deep RL controller improves hexapod locomotion learning.
problem Deep RL struggles with real-world legged robot control.
method Decentralized deep RL on a hexapod robot.
result Decentralized approach learns better and faster.
Meta-learning has emerged as an important framework for learning new tasks from just a few examples. The success of any meta-learning model depends on (i) its fast adaptation to new tasks, as well as (ii) having a shared representation across similar tasks. Here we extend the model-agnostic meta-learning (MAML) framewo…
Representation learning becomes especially important for complex systems with multimodal data sources such as cameras or sensors. Recent advances in reinforcement learning and optimal control make it possible to design control algorithms on these latent representations, but the field still lacks a large-scale standard …
Complex environments and tasks pose a difficult problem for holistic end-to-end learning approaches. Decomposition of an environment into interacting controllable and non-controllable objects allows supervised learning for non-controllable objects and universal value function approximator learning for controllable obje…
Iterative method learns unknown constraints for MPC control.
problem Learning to satisfy unknown polyhedral state constraints in iterative MPC.
method Collects and improves estimates of unknown constraints using collected data, designs an MPC controller to satisfy the estimated constraints.
result Robust and probabilistic guarantees of constraint satisfaction as a function of task iterations.
Proposes a new reinforcement learning method to improve agent performance in control tasks.
problem Shortcomings of maximum likelihood estimation in model-based reinforcement learning.
method Directly optimizes expected returns using implicit differentiation of a Bellman optimality function.
result Empirical evidence shows improved performance in model misspecification regime.
New text-to-image diffusion models improve scene understanding for AI agents.
problem Fine-grained scene understanding for AI agents from text and images.
method Pre-trained text-to-image diffusion models optimized for generating images from text prompts.
result Policies learned with Stable Control Representations outperform state-of-the-art approaches on various control tasks.
Improved LLM pre-training performance through better weight and variance control.
problem Improper weight and variance control in LLM pre-training affects downstream task performance.
method Introduced Layer Index Rescaling (LIR) and Target Variance Rescaling (TVR) techniques.
result Substantial improvements in downstream task performance (up to 4.6%) and reduced extreme activation values.
Modern automation systems rely on closed loop control, wherein a controller interacts with a controlled process, based on observations. These systems are increasingly complex, yet most controllers are linear Proportional-Integral-Derivative (PID) controllers. PID controllers perform well on linear and near-linear syste…
New method disentangles perceptual uncertainty and behavioral costs in partially observable systems.
problem Tackles inverse optimal control for non-linear partially observable systems.
method Probabilistic approach using maximum causal entropy formulations and local linearization.
result Disentangles perceptual factors and behavioral costs in sequential decision-making.
As the range of tasks performed by a general vision system expands, executing multiple tasks accurately and efficiently in a single network has become an important and still open problem. Recent computer vision approaches address this problem by branching networks, or by a channel-wise modulation of the network feature…
Deep reinforcement learning has been shown to solve challenging tasks where large amounts of training experience is available, usually obtained online while learning the task. Robotics is a significant potential application domain for many of these algorithms, but generating robot experience in the real world is expens…
New method calibrates diffusion models for image regression tasks.
problem Ensuring reliability of diffusion models for critical applications.
method Risk-Controlling Prediction Sets (RCPS) with convex optimization.
result Calibrated entrywise intervals and risk control with minimal mean interval length.
Simulation-to-real transfer is an important strategy for making reinforcement learning practical with real robots. Successful sim-to-real transfer systems have difficulty producing policies which generalize across tasks, despite training for thousands of hours equivalent real robot time. To address this shortcoming, we…
We present an embedding of stochastic optimal control problems, of the so called path integral form, into reproducing kernel Hilbert spaces. Using consistent, sample based estimates of the embedding leads to a model free, non-parametric approach for calculation of an approximate solution to the control problem. This fo…
Deep reinforcement learning for high dimensional, hierarchical control tasks usually requires the use of complex neural networks as functional approximators, which can lead to inefficiency, instability and even divergence in the training process. Here, we introduce stacked deep Q learning (SDQL), a flexible modularized…
Many advanced Learning from Demonstration (LfD) methods consider the decomposition of complex, real-world tasks into simpler sub-tasks. By reusing the corresponding sub-policies within and between tasks, they provide training data for each policy from different high-level tasks and compose them to perform novel ones. E…