arXiv research

Algorithm improves imitation learning from visual data in partially observable environments.

New algorithm improves reinforcement learning from partial observations.

Observational learning is a type of learning that occurs as a function of observing, retaining and possibly replicating or imitating the behaviour of another agent. It is a core mechanism appearing in various instances of social learning and has been found to be employed in several intelligent species, including humans…

Deep learning identifies space objects from uncorrelated observations.

A new algorithm CAP learns optimal policies from observational data with confounding bias and missing observations.

Overfitting occurs when RL agents correlate rewards with spurious observation features.

The paper teaches robots to navigate by learning costs from expert demonstrations.

Solves high-dimensional observation learning for control models.

Paper establishes baselines for offline RL from visual observations.

Controlled interventions provide the most direct source of information for learning causal effects. In particular, a dose-response curve can be learned by varying the treatment level and observing the corresponding outcomes. However, interventions can be expensive and time-consuming. Observational data, where the treat…

Method learns dynamics from noisy partial observations.

Intelligent agents can cope with sensory-rich environments by learning task-agnostic state abstractions. In this paper, we propose an algorithm to approximate causal states, which are the coarsest partition of the joint history of actions and observations in partially-observable Markov decision processes (POMDP). Our m…

In many cases an intelligent agent may want to learn how to mimic a single observed demonstrated trajectory. In this work we consider how to perform such procedural learning from observation, which could help to enable agents to better use the enormous set of video data on observation sequences. Our approach exploits t…

New method identifies causal variables from partially observed data.

New framework for learning from various weak supervision types.

Method estimates observation functions in state-space models without supervision.

New algorithm proves RL from partial obs is feasible.

New algorithm learns optimal decisions from imperfectly observed contexts.

New method combines observational and interventional data for causal model learning.

Guarantees for third-person imitation learning from offline data.

RL struggles with generalization due to implicit partial observability.

RichID learns optimal control policies from nonlinear observations.

New method learns spatiotemporal dynamics from random point process observations.

LUQ learns QoI from dynamical systems for consistent observation inversion.

Imitation by observation is an approach for learning from expert demonstrations that lack action information, such as videos. Recent approaches to this problem can be placed into two broad categories: training dynamics models that aim to predict the actions taken between states, and learning rewards or features for com…

DPFRL uses particle filters for decision making with complex visual observations.

Learning a policy using only observational data is challenging because the distribution of states it induces at execution time may differ from the distribution observed during training. We propose to train a policy by unrolling a learned model of the environment dynamics over multiple time steps while explicitly penali…

CausalRM models rewards from user feedback, overcoming noise and bias.

Paper proposes learning system dynamics from irregularly-sampled partial observations.

A new RL algorithm tackles PO tasks by modeling the environment and improving the policy.

The paper develops a method to learn robust decision policies from observational data, reducing high-cost outcomes.

Imitation from observation (IfO) is the problem of learning directly from state-only demonstrations without having access to the demonstrator's actions. The lack of action information both distinguishes IfO from most of the literature in imitation learning, and also sets it apart as a method that may enable agents to l…

New estimator learns symmetric dynamics from few observations.

Enhanced feedback model improves sample-efficiency in POMDPs.

New algorithms learn POMDPs efficiently with hindsight observability.

Observer learns optimal policy from learner's actions without rewards.

The paper develops a method to learn navigation costs from expert demonstrations in partially observable environments.

FMI uses matching to mimic interventions for causal feature learning.

Agents learning to act autonomously in real-world domains must acquire a model of the dynamics of the domain in which they operate. Learning domain dynamics can be challenging, especially where an agent only has partial access to the world state, and/or noisy external sensors. Even in standard STRIPS domains, existing …

Study learns linear system dynamics from noisy bilinear data.

Climate projections continue to be marred by large uncertainties, which originate in processes that need to be parameterized, such as clouds, convection, and ecosystems. But rapid progress is now within reach. New computational tools and methods from data assimilation and machine learning make it possible to integrate …

The paper extends multiple instance learning to multiclass and regression problems.

We study Imitation Learning (IL) from Observations alone (ILFO) in large-scale MDPs. While most IL algorithms rely on an expert to directly provide actions to the learner, in this setting the expert only supplies sequences of observations. We design a new model-free algorithm for ILFO, Forward Adversarial Imitation Lea…

Method estimates causal effects from incremental data, overcoming missing data challenges.

Imitation from observation is the framework of learning tasks by observing demonstrated state-only trajectories. Recently, adversarial approaches have achieved significant performance improvements over other methods for imitating complex behaviors. However, these adversarial imitation algorithms often require many demo…

EnCF improves data assimilation for implicit, non-smooth observations.

Survey of active learning for data streams.

Modeling dynamical systems is important in many disciplines, e.g., control, robotics, or neurotechnology. Commonly the state of these systems is not directly observed, but only available through noisy and potentially high-dimensional observations. In these cases, system identification, i.e., finding the measurement map…