Object search is a fundamental skill for household robots, yet the core problem lies in the robot's ability to locate the target object accurately. The dynamic nature of household environments, characterized by the arbitrary placement of daily objects by users, makes it challenging to perform target localization. To efficiently locate the target object, the robot needs to be equipped with knowledge at both the object and room level. However, existing approaches rely solely on one type of knowledge, leading to unsatisfactory object localization performance and, consequently, inefficient object search processes. To address this problem, we propose a commonsense scene graph-based target localization, CSG-TL, to enhance target object search in the household environment. Given the pre-built map with stationary items, the robot models the room-level knowledge with object-level commonsense knowledge generated by a large language model (LLM) to a commonsense scene graph (CSG), supporting both types of knowledge for CSG-TL. To demonstrate the superiority of CSG-TL on target localization, extensive experiments are performed on the real-world ScanNet dataset and the AI2THOR simulator. Moreover, we have extended CSG-TL to an object search framework, CSG-OS, validated in both simulated and real-world environments. Code and videos are available at https://sites.google.com/view/csg-os.

This paper designs novel nonparametric Bellman mappings in reproducing kernel Hilbert spaces (RKHSs) for reinforcement learning (RL). The proposed mappings benefit from the rich approximating properties of RKHSs, adopt no assumptions on the statistics of the data owing to their nonparametric nature, require no knowledge on transition probabilities of Markov decision processes, and may operate without any training data. Moreover, they allow for sampling on-the-fly via the design of trajectory samples, re-use past test data via experience replay, effect dimensionality reduction by random Fourier features, and enable computationally lightweight operations to fit into efficient online or time-adaptive learning. The paper offers also a variational framework to design the free parameters of the proposed Bellman mappings, and shows that appropriate choices of those parameters yield several popular Bellman-mapping designs. As an application, the proposed mappings are employed to offer a novel solution to the problem of countering outliers in adaptive filtering. More specifically, with no prior information on the statistics of the outliers and no training data, a policy-iteration algorithm is introduced to select online, per time instance, the ``optimal'' coefficient p in the least-mean-p-power-error method. Numerical tests on synthetic data showcase, in most of the cases, the superior performance of the proposed solution over several RL and non-RL schemes.

We study evaluating a policy under best- and worst-case perturbations to a Markov decision process (MDP), given transition observations from the original MDP, whether under the same or different policy. This is an important problem when there is the possibility of a shift between historical and future environments, due to e.g. unmeasured confounding, distributional shift, or an adversarial environment. We propose a perturbation model that can modify transition kernel densities up to a given multiplicative factor or its reciprocal, which extends the classic marginal sensitivity model (MSM) for single time step decision making to infinite-horizon RL. We characterize the sharp bounds on policy value under this model, that is, the tightest possible bounds given by the transition observations from the original MDP, and we study the estimation of these bounds from such transition observations. We develop an estimator with several appealing guarantees: it is semiparametrically efficient, and remains so even when certain necessary nuisance functions such as worst-case Q-functions are estimated at slow nonparametric rates. It is also asymptotically normal, enabling easy statistical inference using Wald confidence intervals. In addition, when certain nuisances are estimated inconsistently we still estimate a valid, albeit possibly not sharp bounds on the policy value. We validate these properties in numeric simulations. The combination of accounting for environment shifts from train to test (robustness), being insensitive to nuisance-function estimation (orthogonality), and accounting for having only finite samples to learn from (inference) together leads to credible and reliable policy evaluation.

As humans come to rely on autonomous systems more, ensuring the transparency of such systems is important to their continued adoption. Explainable Artificial Intelligence (XAI) aims to reduce confusion and foster trust in systems by providing explanations of agent behavior. Partially observable Markov decision processes (POMDPs) provide a flexible framework capable of reasoning over transition and state uncertainty, while also being amenable to explanation. This work investigates the use of user-provided counterfactuals to generate contrastive explanations of POMDP policies. Feature expectations are used as a means of contrasting the performance of these policies. We demonstrate our approach in a Search and Rescue (SAR) setting. We analyze and discuss the associated challenges through two case studies.

Large Language Models (LLMs) have shown proficiency in question-answering tasks but often struggle to integrate real-time knowledge updates, leading to potentially outdated or inaccurate responses. This problem becomes even more challenging when dealing with multi-hop questions since they require LLMs to update and integrate multiple knowledge pieces relevant to the questions. To tackle the problem, we propose the Retrieval-Augmented model Editing (RAE) framework tailored for multi-hop question answering. RAE first retrieves edited facts and then refines the language model through in-context learning. Specifically, our retrieval approach, based on mutual information maximization, leverages the reasoning abilities of LLMs to identify chain facts that na\"ive similarity-based searches might miss. Additionally, our framework incorporates a pruning strategy to eliminate redundant information from the retrieved facts, which enhances the editing accuracy and mitigates the hallucination problem. Our framework is supported by theoretical justification for its fact retrieval efficacy. Finally, comprehensive evaluation across various LLMs validates RAE's ability in providing accurate answers with updated knowledge.

In human-robot collaboration, shared control presents an opportunity to teleoperate robotic manipulation to improve the efficiency of manufacturing and assembly processes. Robots are expected to assist in executing the user's intentions. To this end, robust and prompt intention estimation is needed, relying on behavioral observations. The framework presents an intention estimation technique at hierarchical levels i.e., low-level actions and high-level tasks, by incorporating multi-scale hierarchical information in neural networks. Technically, we employ hierarchical dependency loss to boost overall accuracy. Furthermore, we propose a multi-window method that assigns proper hierarchical prediction windows of input data. An analysis of the predictive power with various inputs demonstrates the predominance of the deep hierarchical model in the sense of prediction accuracy and early intention identification. We implement the algorithm on a virtual reality (VR) setup to teleoperate robotic hands in a simulation with various assembly tasks to show the effectiveness of online estimation.

Kakade's natural policy gradient method has been studied extensively in the last years showing linear convergence with and without regularization. We study another natural gradient method which is based on the Fisher information matrix of the state-action distributions and has received little attention from the theoretical side. Here, the state-action distributions follow the Fisher-Rao gradient flow inside the state-action polytope with respect to a linear potential. Therefore, we study Fisher-Rao gradient flows of linear programs more generally and show linear convergence with a rate that depends on the geometry of the linear program. Equivalently, this yields an estimate on the error induced by entropic regularization of the linear program which improves existing results. We extend these results and show sublinear convergence for perturbed Fisher-Rao gradient flows and natural gradient flows up to an approximation error. In particular, these general results cover the case of state-action natural policy gradients.

We consider a setting in which $N$ agents aim to speedup a common Stochastic Approximation (SA) problem by acting in parallel and communicating with a central server. We assume that the up-link transmissions to the server are subject to asynchronous and potentially unbounded time-varying delays. To mitigate the effect of delays and stragglers while reaping the benefits of distributed computation, we propose \texttt{DASA}, a Delay-Adaptive algorithm for multi-agent Stochastic Approximation. We provide a finite-time analysis of \texttt{DASA} assuming that the agents' stochastic observation processes are independent Markov chains. Significantly advancing existing results, \texttt{DASA} is the first algorithm whose convergence rate depends only on the mixing time $\tau_{mix}$ and on the average delay $\tau_{avg}$ while jointly achieving an $N$-fold convergence speedup under Markovian sampling. Our work is relevant for various SA applications, including multi-agent and distributed temporal difference (TD) learning, Q-learning and stochastic optimization with correlated data.

Motivated by applications in large-scale and multi-agent reinforcement learning, we study the non-asymptotic performance of stochastic approximation (SA) schemes with delayed updates under Markovian sampling. While the effect of delays has been extensively studied for optimization, the manner in which they interact with the underlying Markov process to shape the finite-time performance of SA remains poorly understood. In this context, our first main contribution is to show that under time-varying bounded delays, the delayed SA update rule guarantees exponentially fast convergence of the \emph{last iterate} to a ball around the SA operator's fixed point. Notably, our bound is \emph{tight} in its dependence on both the maximum delay $\tau_{max}$, and the mixing time $\tau_{mix}$. To achieve this tight bound, we develop a novel inductive proof technique that, unlike various existing delayed-optimization analyses, relies on establishing uniform boundedness of the iterates. As such, our proof may be of independent interest. Next, to mitigate the impact of the maximum delay on the convergence rate, we provide the first finite-time analysis of a delay-adaptive SA scheme under Markovian sampling. In particular, we show that the exponent of convergence of this scheme gets scaled down by $\tau_{avg}$, as opposed to $\tau_{max}$ for the vanilla delayed SA rule; here, $\tau_{avg}$ denotes the average delay across all iterations. Moreover, the adaptive scheme requires no prior knowledge of the delay sequence for step-size tuning. Our theoretical findings shed light on the finite-time effects of delays for a broad class of algorithms, including TD learning, Q-learning, and stochastic gradient descent under Markovian sampling.

In general, reliable communication via multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) requires accurate channel estimation at the receiver. The existing literature largely focuses on denoising methods for channel estimation that depend on either (i) channel analysis in the time-domain with prior channel knowledge or (ii) supervised learning techniques which require large pre-labeled datasets for training. To address these limitations, we present a frequency-domain denoising method based on a reinforcement learning framework that does not need a priori channel knowledge and pre-labeled data. Our methodology includes a new successive channel denoising process based on channel curvature computation, for which we obtain a channel curvature magnitude threshold to identify unreliable channel estimates. Based on this process, we formulate the denoising mechanism as a Markov decision process, where we define the actions through a geometry-based channel estimation update, and the reward function based on a policy that reduces mean squared error (MSE). We then resort to Q-learning to update the channel estimates. Numerical results verify that our denoising algorithm can successfully mitigate noise in channel estimates. In particular, our algorithm provides a significant improvement over the practical least squares (LS) estimation method and provides performance that approaches that of the ideal linear minimum mean square error (LMMSE) estimation with perfect knowledge of channel statistics.