Recent advances in diffusion-based robot policies have demonstrated significant potential in imitating multi-modal behaviors. However, these approaches typically require large quantities of demonstration data paired with corresponding robot action labels, creating a substantial data collection burden. In this work, we propose a plan-then-control framework aimed at improving the action-data efficiency of inverse dynamics controllers by leveraging observational demonstration data. Specifically, we adopt a Deep Koopman Operator framework to model the dynamical system and utilize observation-only trajectories to learn a latent action representation. This latent representation can then be effectively mapped to real high-dimensional continuous actions using a linear action decoder, requiring minimal action-labeled data. Through experiments on simulated robot manipulation tasks and a real robot experiment with multi-modal expert demonstrations, we demonstrate that our approach significantly enhances action-data efficiency and achieves high task success rates with limited action data.

Safe and successful deployment of robots requires not only the ability to generate complex plans but also the capacity to frequently replan and correct execution errors. This paper addresses the challenge of long-horizon trajectory planning under temporally extended objectives in a receding horizon manner. To this end, we propose Doppler, a data-driven hierarchical framework that generates and updates plans based on instruction specified by linear temporal logic (LTL). Our method decomposes temporal tasks into chain of options with hierarchical reinforcement learning from offline non-expert datasets. It leverages diffusion models to generate options with low-level actions. We devise a determinantal-guided posterior sampling technique during batch generation, which improves the speed and diversity of diffusion generated options, leading to more efficient querying. Experiments on robot navigation and manipulation tasks demonstrate that Doppler can generate sequences of trajectories that progressively satisfy the specified formulae for obstacle avoidance and sequential visitation.

Building upon the foundations laid by our previous work, this paper introduces Arena 4.0, a significant advancement of Arena 3.0 [1], Arena-Bench [2], Arena 1.0 [3], and Arena 2.0 [4]. Arena 4.0 provides three main novel contributions: 1) a generative-model-based world and scenario generation approach using large language models (LLMs) and diffusion models, to dynamically generate complex, humancentric environments from text prompts or 2D floorplans that can be used for development and benchmarking of social navigation strategies. 2) A comprehensive 3D model database which can be extended with 3D assets and semantically linked and annotated using a variety of metrics for dynamic spawning and arrangements inside 3D worlds. 3) The complete migration towards ROS 2, which ensures operation with state-of-the-art hardware and functionalities for improved navigation, usability, and simplified transfer towards real robots. We evaluated the platforms performance through a comprehensive user study and its world generation capabilities for benchmarking demonstrating significant improvements in usability and efficiency compared to previous versions. Arena 4.0 is openly available at https://github.com/Arena-Rosnav.

We present OpenRoboCare, a multimodal dataset for robot caregiving, capturing expert occupational therapist demonstrations of Activities of Daily Living (ADLs). Caregiving tasks involve complex physical human-robot interactions, requiring precise perception under occlusions, safe physical contact, and long-horizon planning. While recent advances in robot learning from demonstrations have shown promise, there is a lack of a large-scale, diverse, and expert-driven dataset that captures real-world caregiving routines. To address this gap, we collect data from 21 occupational therapists performing 15 ADL tasks on two manikins. The dataset spans five modalities—RGB-D video, pose tracking, eye-gaze tracking, task and action annotations, and tactile sensing, providing rich multimodal insights into caregiver movement, attention, force application, and task execution strategies. We further analyze expert caregiving principles and strategies, offering insights to improve robot efficiency and task feasibility. Additionally, our evaluations demonstrate that OpenRoboCare presents challenges for state-of-the-art robot perception and human activity recognition methods, both critical for developing safe and adaptive assistive robots, highlighting the value of our contribution. See our website for additional visualizations: https://emprise.cs.cornell.edu/robo-care/.

While most optical tactile sensors rely on measuring surface displacement, insights from continuum mechanics suggest that measuring shear strain provides key information for tactile sensing. In this work, we introduce an optical tactile sensing principle based on shear strain detection. A silicone rubber layer, dyed with color inks, is used to quantify the shear magnitude of the sensing layer. This principle was validated using the NUSense camera-based tactile sensor. The wide-angle camera captures the elongation of the soft pad under mechanical load, a phenomenon attributed to the Poisson effect. We tested the robustness of the sensor by subjecting the outermost layer to multiple load (8 N) cycles using a 5 mm in radius ball head indenter. The physical and optical properties of the inked pad proved essential and remained stable over time, exhibiting only low variance.

In this work, we are interested in automated methods for knowledge graph creation (KGC) from input text.Progress on large language models (LLMs) has prompted a series of recent works applying them to KGC, e.g., via zero/few-shot prompting.Despite successes on small domain-specific datasets, these models face difficulties scaling up to text common in many real-world applications.A principal issue is that, in prior methods, the KG schema has to be included in the LLM prompt to generate valid triplets; larger and more complex schemas easily exceed the LLMs' context window length.Furthermore, there are scenarios where a fixed pre-defined schema is not available and we would like the method to construct a high-quality KG with a succinct selfgenerated schema.To address these problems, we propose a three-phase framework named Extract-Define-Canonicalize (EDC): open information extraction followed by schema definition and post-hoc canonicalization.EDC is flexible in that it can be applied to settings where a pre-defined target schema is available and when it is not; in the latter case, it constructs a schema automatically and applies selfcanonicalization.To further improve performance, we introduce a trained component that retrieves schema elements relevant to the input text; this improves the LLMs' extraction performance in a retrieval-augmented generationlike manner.We demonstrate on three KGC benchmarks that EDC is able to extract highquality triplets without any parameter tuning and with significantly larger schemas compared to prior works.

In this work, we focus on object search tasks within unexplored environments. We introduce a framework centered around the Probable Object Location (POLo) score. Utilizing a 3D object probability map, the POLo score allows the agent to make data-driven decisions for efficient object search. We further enhance the framework’s practicality by introducing POLoNet, a neural network trained to approximate the computationally-intensive POLo score. Our approach addresses critical limitations of both end-to-end reinforcement learning methods, which suffer from memory decay over long-horizon tasks, and traditional map-based methods that neglect visibility constraints. Our experiments, involving the first phase of the Open-Vocabulary Mobile Manipulation (OVMM) 2023 challenge, demonstrate that an agent equipped with POLoNet significantly outperforms a range of baseline methods, including end-to-end RL techniques and prior map-based strategies. To provide a comprehensive evaluation, we introduce new performance metrics that offer insights into the efficiency and effectiveness of various agents in object goal navigation.

Out-of-distribution (OOD) detection is a critical task in machine learning that seeks to identify abnormal samples. Traditionally, unsupervised methods utilize a deep generative model for OOD detection. However, such approaches require a new model to be trained for each inlier dataset. This paper explores whether a single model can perform OOD detection across diverse tasks. To that end, we introduce Diffusion Paths (DiffPath), which uses a single diffusion model originally trained to perform unconditional generation for OOD detection. We introduce a novel technique of measuring the rate-of-change and curvature of the diffusion paths connecting samples to the standard normal. Extensive experiments show that with a single model, DiffPath is competitive with prior work using individual models on a variety of OOD tasks involving different distributions. Our code is publicly available at https://github.com/clear-nus/diffpath.

Learning accurate predictive models of real-world dynamic phenomena (e.g., climate, biological) remains a challenging task. One key issue is that the data generated by both natural and artificial processes often comprise time series that are irregularly sampled and/or contain missing observations. In this work, we propose the Neural Continuous-Discrete State Space Model (NCDSSM) for continuous-time modeling of time series through discrete-time observations. NCDSSM employs auxiliary variables to disentangle recognition from dynamics, thus requiring amortized inference only for the auxiliary variables. Leveraging techniques from continuous-discrete filtering theory, we demonstrate how to perform accurate Bayesian inference for the dynamic states. We propose three flexible parameterizations of the latent dynamics and an efficient training objective that marginalizes the dynamic states during inference. Empirical results on multiple benchmark datasets across various domains show improved imputation and forecasting performance of NCDSSM over existing models.

In this work, we focus on the problem of safe policy transfer in reinforcement learning: we seek to leverage existing policies when learning a new task with specified constraints. This problem is important for safety-critical applications where interactions are costly and unconstrained exploration can lead to undesirable or dangerous outcomes, e.g., with physical robots that interact with humans. We propose a Constrained Markov Decision Process (CMDP) formulation that simultaneously enables the transfer of policies and adherence to safety constraints. Our formulation cleanly separates task goals from safety considerations and permits the specification of a wide variety of constraints. Our approach relies on a novel extension of generalized policy improvement to constrained settings via a Lagrangian formulation. We devise a dual optimization algorithm that estimates the optimal dual variable of a target task, thus enabling safe transfer of policies derived from successor features learned on source tasks. Our experiments in simulated domains show that our approach is effective; it visits unsafe states less frequently and outperforms alternative state-of-the-art methods when taking safety constraints into account.

Perspective-taking is the ability to perceive or understand a situation or concept from another individual's point of view, and is crucial in daily human interactions. Enabling robots to perform perspective-taking remains an unsolved problem; existing approaches that use deterministic or handcrafted methods are unable to accurately account for uncertainty in partially-observable settings. This work proposes to address this limitation via a deep world model that enables a robot to perform both perception and conceptual perspective taking, i.e., the robot is able to infer what a human sees and believes. The key innovation is a decomposed multi-modal latent state space model able to generate and augment fictitious observations/emissions. Optimizing the ELBO that arises from this probabilistic graphical model enables the learning of uncertainty in latent space, which facilitates uncertainty estimation from high-dimensional observations. We tasked our model to predict human observations and beliefs on three partially-observable HRI tasks. Experiments show that our method significantly outperforms existing baselines and is able to infer visual observations available to other agent and their internal beliefs.

In this work, we present GraspFlow, a refinement approach for generating context-specific grasps. We formulate the problem of grasp synthesis as a sampling problem: we seek to sample from a context-conditioned probability distribution of successful grasps. However, this target distribution is unknown. As a solution, we devise a discriminator gradient-flow method to evolve grasps obtained from a simpler distribution in a manner that mimics sampling from the desired target distribution. Unlike existing approaches, GraspFlow is modular, allowing grasps that satisfy multiple criteria to be obtained simply by incorporating the relevant discriminators. It is also simple to implement, requiring minimal code given existing auto-differentiation libraries and suitable discriminators. Experiments show that GraspFlow generates stable and executable grasps on a real-world Panda robot for a diverse range of objects. In particular, in 60 trials on 20 different household objects, the first attempted grasp was successful 94% of the time, and 100% grasp success was achieved by the second grasp. Moreover, incorporating a functional discriminator for robot-human handover improved the functional aspect of the grasp by up to 33%.

Human models play a crucial role in human-robot interaction (HRI), enabling robots to consider the impact of their actions on people and plan their behavior accordingly. However, crafting good human models is challenging; capturing context-dependent human behavior requires significant prior knowledge and/or large amounts of interaction data, both of which are difficult to obtain. In this work, we explore the potential of large language models (LLMs) — which have consumed vast amounts of human-generated text data — to act as zero-shot human models for HRI. Our experiments on three social datasets yield promising results; the LLMs are able to achieve performance comparable to purpose-built models. That said, we also discuss current limitations, such as sensitivity to prompts and spatial/numerical reasoning mishaps. Based on our findings, we demonstrate how LLM-based human models can be integrated into a social robot's planning process and applied in HRI scenarios focused on the important element of trust. Specifically, we present one case study on a simulated trust-based table-clearing task and replicate past results that relied on custom models. Next, we conduct a new robot utensil-passing experiment ($n=65$) where preliminary results show that planning with an LLM-based human model can achieve gains over a basic myopic plan. In summary, our results show that LLMs offer a promising (but incomplete) approach to human modeling for HRI.

The recent proliferation of large-scale text-to-image models has led to growing concerns that such models may be misused to generate harmful, misleading, and inappropriate content. Motivated by this issue, we derive a technique inspired by continual learning to selectively forget concepts in pretrained deep generative models. Our method, dubbed Selective Amnesia, enables controllable forgetting where a user can specify how a concept should be forgotten. Selective Amnesia can be applied to conditional variational likelihood models, which encompass a variety of popular deep generative frameworks, including variational autoencoders and large-scale text-to-image diffusion models. Experiments across different models demonstrate that our approach induces forgetting on a variety of concepts, from entire classes in standard datasets to celebrity and nudity prompts in text-to-image models.

Reaching consensus and convergence to equilibrium are two major challenges of multi-agent systems. Although each has attracted significant attention, relatively few studies address both challenges at the same time. This paper examines the connection between the notions of consensus and equilibrium in a multi-agent system where multiple interacting sub-populations coexist. We argue that consensus can be seen as an intricate component of intra-population stability, whereas equilibrium can be seen as encoding inter-population stability. We show that smooth fictitious play, a well-known learning model in game theory, can achieve both consensus and convergence to equilibrium in diverse multi-agent settings. Moreover, we show that the consensus formation process plays a crucial role in the seminal thorny problem of equilibrium selection in multi-agent learning.

This paper proposes SCALES, a general framework that translates well-established fairness principles into a common representation based on the Constraint Markov Decision Process (CMDP). With the help of causal language, our framework can place constraints on both the procedure of decision making (procedural fairness) as well as the outcomes resulting from decisions (outcome fairness). Specifically, we show that well-known fairness principles can be encoded either as a utility component, a non-causal component, or a causal component in a SCALES-CMDP. We illustrate SCALES using a set of case studies involving a simulated healthcare scenario and the real-world COMPAS dataset. Experiments demonstrate that our framework produces fair policies that embody alternative fairness principles in single-step and sequential decision-making scenarios.

This workshop aims to bring together researchers to explore and identify ways in which human-robot collaboration can reap the benefits of modern machine learning. The intended outcome is a roadmap that identifies key milestones that will lead us towards fluent effective human-robot teaming. In addition to focus groups and creative brainstorming exercises, this workshop will comprise invited talks, contributed paper talks, a poster session, and a debate. The papers, talks, posters, and roadmap will be made publicly available on our website: https://sites.google.com/view/mlhrc-hri-2022/home.

This work focuses on learning useful and robust deep world models using multiple, possibly unreliable, sensors. We find that current methods do not sufficiently encourage a shared representation between modalities; this can cause poor performance on downstream tasks and over-reliance on specific sensors. As a solution, we contribute a new multi-modal deep latent state-space model, trained using a mutual information lower-bound. The key innovation is a specially-designed density ratio estimator that encourages consistency between the latent codes of each modality. We tasked our method to learn policies (in a self-supervised manner) on multi-modal Natural MuJoCo benchmarks and a challenging Table Wiping task. Experiments show our method significantly outperforms state-of-the-art deep reinforcement learning methods, particularly in the presence of missing observations.

Sequences and time-series often arise in robot tasks, e.g., in activity recognition and imitation learning. In recent years, deep neural networks (DNNs) have emerged as an effective data-driven methodology for processing sequences given sufficient training data and compute resources. However, when data is limited, simpler models such as logic/rule-based methods work surprisingly well, especially when relevant prior knowledge is applied in their construction. However, unlike DNNs, these "structured" models can be difficult to extend, and do not work well with raw unstructured data. In this work, we seek to learn flexible DNNs, yet leverage prior temporal knowledge when available. Our approach is to embed symbolic knowledge expressed as linear temporal logic (LTL) and use these embeddings to guide the training of deep models. Specifically, we construct semantic-based embeddings of automata generated from LTL formula via a Graph Neural Network. Experiments show that these learnt embeddings can lead to improvements on downstream robot tasks such as sequential action recognition and imitation learning.

Humans display the remarkable ability to sense the world through tools and other held objects. For example, we are able to pinpoint impact locations on a held rod and tell apart different textures using a rigid probe. In this work, we consider how we can enable robots to have a similar capacity, i.e., to embody tools and extend perception using standard grasped objects. We propose that vibro-tactile sensing using dynamic tactile sensors on the robot fingers, along with machine learning models, enables robots to decipher contact information that is transmitted as vibrations along rigid objects. This paper reports on extensive experiments using the BioTac micro-vibration sensor and a new event dynamic sensor, the NUSkin, capable of multi-taxel sensing at 4 kHz. We demonstrate that fine localization on a held rod is possible using our approach (with errors less than 1 cm on a 20 cm rod). Next, we show that vibro-tactile perception can lead to reasonable grasp stability prediction during object handover, and accurate food identification using a standard fork. We find that multi-taxel vibro-tactile sensing at a sufficiently high sampling rate (above 2 kHz) led to the best performance across the various tasks and objects. Taken together, our results provide both evidence and guidelines for using vibro-tactile perception to extend tactile perception, which we believe will lead to enhanced competency with tools and better physical human-robot interaction.

Many complex time series can be effectively subdivided into distinct regimes that exhibit persistent dynamics. Discovering the switching behavior and the statistical patterns in these regimes is important for understanding the underlying dynamical system. We propose the Recurrent Explicit Duration Switching Dynamical System (RED-SDS), a flexible model that is capable of identifying both state- and time-dependent switching dynamics. State-dependent switching is enabled by a recurrent state-to-switch connection and an explicit duration count variable is used to improve the time-dependent switching behavior. We demonstrate how to perform efficient inference using a hybrid algorithm that approximates the posterior of the continuous states via an inference network and performs exact inference for the discrete switches and counts. The model is trained by maximizing a Monte Carlo lower bound of the marginal log-likelihood that can be computed efficiently as a byproduct of the inference routine. Empirical results on multiple datasets demonstrate that RED-SDS achieves considerable improvement in time series segmentation and competitive forecasting performance against the state of the art.

Affective Computing is a rapidly growing field spurred by advancements in artificial intelligence, but often, held back by the inability to translate psychological theories of emotion into tractable computational models. To address this, we propose a probabilistic programming approach to affective computing, which models psychological-grounded theories as generative models of emotion, and implements them as stochastic, executable computer programs. We first review probabilistic approaches that integrate reasoning about emotions with reasoning about other latent mental states (e.g., beliefs, desires) in context. Recently-developed probabilistic programming languages offer several key desidarata over previous approaches, such as: (i) flexibility in representing emotions and emotional processes; (ii) modularity and compositionality; (iii) integration with deep learning libraries that facilitate efficient inference and learning from large, naturalistic data; and (iv) ease of adoption. Furthermore, using a probabilistic programming framework allows a standardized platform for theory-building and experimentation: Competing theories (e.g., of appraisal or other emotional processes) can be easily compared via modular substitution of code followed by model comparison. To jumpstart adoption, we illustrate our points with executable code that researchers can easily modify for their own models. We end with a discussion of applications and future directions of the probabilistic programming approach.

Integrating deep learning with latent state space models has the potential to yield temporal models that are powerful, yet tractable and interpretable. Unfortunately, current models are not designed to handle missing data or multiple data modalities, which are both prevalent in real-world data. In this work, we introduce a factorized inference method for Multimodal Deep Markov Models (MDMMs), allowing us to filter and smooth in the presence of missing data, while also performing uncertainty-aware multimodal fusion. We derive this method by factorizing the posterior p(z|x) for non-linear state space models, and develop a variational backward-forward algorithm for inference. Because our method handles incompleteness over both time and modalities, it is capable of interpolation, extrapolation, conditional generation, label prediction, and weakly supervised learning of multimodal time series. We demonstrate these capabilities on both synthetic and real-world multimodal data under high levels of data deletion. Our method performs well even with more than 50% missing data, and outperforms existing deep approaches to inference in latent time series.

Implicit Generative Models (IGMs) such as GANs have emerged as effective data-driven models for generating samples, particularly images. In this paper, we formulate the problem of learning an IGM as minimizing the expected distance between characteristic functions. Specifically, we minimize the distance between characteristic functions of the real and generated data distributions under a suitably-chosen weighting distribution. This distance metric, which we term as the characteristic function distance (CFD), can be (approximately) computed with linear time-complexity in the number of samples, in contrast with the quadratic-time Maximum Mean Discrepancy (MMD). By replacing the discrepancy measure in the critic of a GAN with the CFD, we obtain a model that is simple to implement and stable to train. The proposed metric enjoys desirable theoretical properties including continuity and differentiability with respect to generator parameters, and continuity in the weak topology. We further propose a variation of the CFD in which the weighting distribution parameters are also optimized during training; this obviates the need for manual tuning, and leads to an improvement in test power relative to CFD. We demonstrate experimentally that our proposed method outperforms WGAN and MMD-GAN variants on a variety of unsupervised image generation benchmarks.

Tactile perception is crucial for a variety of robot tasks including grasping and in-hand manipulation. New advances in flexible, event-driven, electronic skins may soon endow robots with touch perception capabilities similar to humans. These electronic skins respond asynchronously to changes (e.g., in pressure, temperature), and can be laid out irregularly on the robot's body or end-effector. However, these unique features may render current deep learning approaches such as convolutional feature extractors unsuitable for tactile learning. In this paper, we propose a novel spiking graph neural network for event-based tactile object recognition. To make use of local connectivity of taxels, we present several methods for organizing the tactile data in a graph structure. Based on the constructed graphs, we develop a spiking graph convolutional network. The event-driven nature of spiking neural network makes it arguably more suitable for processing the event-based data. Experimental results on two tactile datasets show that the proposed method outperforms other state-of-the-art spiking methods, achieving high accuracies of approximately 90% when classifying a variety of different household objects.

Common experience suggests that agents who know each other well are better able to work together. In this work, we address the problem of calibrating intention and capabilities in human-robot collaboration. In particular, we focus on scenarios where the robot is attempting to assist a human who is unable to directly communicate her intent. Moreover, both agents may have differing capabilities that are unknown to one another. We adopt a decision-theoretic approach and propose the TICC-POMDP for modeling this setting, with an associated online solver. Experiments show our approach leads to better team performance both in simulation and in a real-world study with human subjects.

Touch is arguably the most important sensing modality in physical interactions. However, tactile sensing has been largely under-explored in robotics applications owing to the complexity in making perceptual inferences until the recent advancements in machine learning or deep learning in particular. Touch perception is strongly influenced by both its temporal dimension similar to audition and its spatial dimension similar to vision. While spatial cues can be learned episodically, temporal cues compete against the system's re-sponse/reaction time to provide accurate inferences. In this paper, we propose a fast tactile-based texture classification framework which makes use of the spiking neural network to learn from the neural coding of the conventional tactile sensor readings. The framework is implemented and tested on two independent tactile datasets collected in sliding motion on 20 material textures. Our results show that the framework is able to make much more accurate inferences ahead of time as compared to that by the state-of-the-art learning approaches.

The problem of inverse reinforcement learning (IRL) is relevant to a variety of tasks including value alignment and robot learning from demonstration. Despite significant algorithmic contributions in recent years, IRL remains an ill-posed problem at its core; multiple reward functions coincide with the observed behavior and the actual reward function is not identifiable without prior knowledge or supplementary information. This paper presents an IRL framework called Bayesian optimization-IRL (BO-IRL) which identifies multiple solutions that are consistent with the expert demonstrations by efficiently exploring the reward function space. BO-IRL achieves this by utilizing Bayesian Optimization along with our newly proposed kernel that (a) projects the parameters of policy invariant reward functions to a single point in a latent space and (b) ensures nearby points in the latent space correspond to reward functions yielding similar likelihoods. This projection allows the use of standard stationary kernels in the latent space to capture the correlations present across the reward function space. Empirical results on synthetic and real-world environments (model-free and model-based) show that BO-IRL discovers multiple reward functions while minimizing the number of expensive exact policy optimizations.

Trust in autonomy is essential for effective human-robot collaboration and user adoption of autonomous systems such as robot assistants. This article introduces a computational model that integrates trust into robot decision making. Specifically, we learn from data a partially observable Markov decision process (POMDP) with human trust as a latent variable. The trust-POMDP model provides a principled approach for the robot to (i) infer the trust of a human teammate through interaction, (ii) reason about the effect of its own actions on human trust, and (iii) choose actions that maximize team performance over the long term. We validated the model through human subject experiments on a table clearing task in simulation (201 participants) and with a real robot (20 participants). In our studies, the robot builds human trust by manipulating low-risk objects first. Interestingly, the robot sometimes fails intentionally to modulate human trust and achieve the best team performance. These results show that the trust-POMDP calibrates trust to improve human-robot team performance over the long term. Further, they highlight that maximizing trust alone does not always lead to the best performance.

We address the problem of unsupervised disentanglement of latent representations learnt via deep generative models. In contrast to current approaches that operate on the evidence lower bound (ELBO), we argue that statistical independence in the latent space of VAEs can be enforced in a principled hierarchical Bayesian manner. To this effect, we augment the standard VAE with an inverse-Wishart (IW) prior on the covariance matrix of the latent code. By tuning the IW parameters, we are able to encourage (or discourage) independence in the learnt latent dimensions. Extensive experimental results on a range of datasets (2DShapes, 3DChairs, 3DFaces and CelebA) show our approach to outperform the β-VAE and is competitive with the state-of-the-art FactorVAE. Our approach achieves significantly better disentanglement and reconstruction on a new dataset (CorrelatedEllipses) which introduces correlations between the factors of variation.

In this paper, we present results from a human-subject study designed to explore two facets of human mental models of robots - inferred capability and intention - and their relationship to overall trust and eventual decisions. In particular, we examine delegation situations characterized by uncertainty, and explore how inferred capability and intention are applied across different tasks. We develop an online survey where human participants decide whether to delegate control to a simulated UAV agent. Our study shows that human estimations of robot capability and intent correlate strongly with overall self-reported trust. However, overall trust is not independently sufficient to determine whether a human will decide to trust (delegate) a given task to a robot. Instead, our study reveals that estimations of robot intention, capability, and overall trust are integrated when deciding to delegate. From a broader perspective, these results suggest that calibrating overall trust alone is insufficient; to make correct decisions, humans need (and use) multi-faceted mental models when collaborating with robots across multiple contexts.

The sense of touch is arguably the first human sense to develop. Empowering robots with the sense of touch may augment their understanding of interacted objects and the environment beyond standard sensory modalities (e.g., vision). This paper investigates the effect of hybridizing touch and sliding movements for tactile-based texture classification. We develop three machine-learning methods within a framework to discriminate between surface textures; the first two methods use hand-engineered features, whilst the third leverages convolutional and recurrent neural network layers to learn feature representations from raw data. To compare these methods, we constructed a dataset comprising tactile data from 23 textures gathered using the iCub platform under a loosely constrained setup, i.e., with nonlinear motion. In line with findings from neuroscience, our experiments show that a good initial estimate can be obtained via touch data, which can be further refined via sliding; combining both touch and sliding data results in 98% classification accuracy over unseen test data.

This work contributes both experimental findings and novel computational human-robot trust models for multi-task settings. We describe Bayesian non-parametric and neural models, and compare their performance on data collected from real-world human-subjects study. Our study spans two distinct task domains: household tasks performed by a Fetch robot, and a virtual reality driving simulation of an autonomous vehicle performing a variety of maneuvers. We find that human trust changes and transfers across tasks in a structured manner based on perceived task characteristics. Our results suggest that task-dependent functional trust models capture human trust in robot capabilities more accurately, and trust transfer across tasks can be inferred to a good degree. We believe these models are key for enabling trust-based robot decision-making for natural human-robot interaction.

In this work, we aim to leverage prior symbolic knowledge to improve the performance of deep models. We propose a graph embedding network that projects propositional formulae (and assignments) onto a manifold via an augmented Graph Convolutional Network (GCN). To generate semantically-faithful embeddings, we develop techniques to recognize node heterogeneity, and semantic regularization that incorporate structural constraints into the embedding. Experiments show that our approach improves the performance of models trained to perform entailment checking and visual relation prediction. Interestingly, we observe a connection between the tractability of the propositional theory representation and the ease of embedding. Future exploration of this connection may elucidate the relationship between knowledge compilation and vector representation learning.

Critiquing is a method for conversational recommendation that adapts recommendations in response to user preference feedback regarding item attributes. Historical critiquing methods were largely based on constraint- and utility-based methods for modifying recommendations w.r.t. these critiqued attributes. In this paper, we revisit the critiquing approach from the lens of deep learning based recommendation methods and language-based interaction. Concretely, we propose an end-to-end deep learning framework with two variants that extend the Neural Collaborative Filtering architecture with explanation and critiquing components. These architectures not only predict personalized keyphrases for a user and item but also embed language-based feedback in the latent space that in turn modulates subsequent critiqued recommendations. We evaluate the proposed framework on two recommendation datasets containing user reviews. Empirical results show that our modified NCF approach not only provides a strong baseline recommender and high-quality personalized item keyphrase suggestions, but that it also properly suppresses items predicted to have a critiqued keyphrase. In summary, this paper provides a first step to unify deep recommendation and language-based feedback in what we hope to be a rich space for future research in deep critiquing for conversational recommendation.

Trust is essential for human-robot collaboration and user adoption of autonomous systems, such as robot assistants. This paper introduces a computational model which integrates trust into robot decision-making. Specifically, we learn from data a partially observable Markov decision process (POMDP) with human trust as a latent variable. The trust-POMDP model provides a principled approach for the robot to (i) infer the trust of a human teammate through interaction, (ii) reason about the effect of its own actions on human behaviors, and (iii) choose actions that maximize team performance over the long term. We validated the model through human subject experiments on a table-clearing task in simulation (201 participants) and with a real robot (20 participants). The results show that the trust-POMDP improves human-robot team performance in this task. They further suggest that maximizing trust in itself may not improve team performance.

This paper proposes an inference framework based on the Z-transform for a specific class of non-homogeneous point processes. This framework gives an alternative method to maximum likelihood estimation which is omnipresent in the field of point processes. The inference strategy is to couple or match the theoretical Z-transform with its empirical counterpart from the observed samples. This procedure fully characterizes the distribution of the point process since there exists a one-to-one mapping with the Z-transform. We illustrate how to use the methodology to estimate a point process whose intensity is driven by a general neural network.

Consider a movie studio aiming to produce a set of new movies for summer release: What types of movies it should produce? Who would the movies appeal to? How many movies should it make? Similar issues are encountered by a variety of organizations, e.g., mobile-phone manufacturers and online magazines, who have to create new (non-existent) items to satisfy groups of users with different preferences. In this paper, we present a joint problem formalization of these interrelated issues, and propose generative methods that address these questions simultaneously. Specifically, we leverage on the latent space obtained by training a deep generative model---the Variational Autoencoder (VAE)---via a loss function that incorporates both rating performance and item reconstruction terms. We use a greedy search algorithm that utilize this learned latent space to jointly obtain K plausible new items, and user groups that would find the items appealing. An evaluation of our methods on a synthetic dataset indicates that our approach is able to generate novel items similar to highly-desirable unobserved items. As case studies on real-world data, we applied our method on the MART abstract art and Movielens Tag Genome datasets, which resulted in promising results: small and diverse sets of novel items.

Collaborative filtering (CF) has made it possible to build personalized recommendation models leveraging the collective data of large user groups, albeit with prescribed models that cannot easily leverage the existence of known behavioral models in particular settings. In this paper, we facilitate the combination of CF with existing behavioral models by introducing Bayesian Behavioral Collaborative Filtering (BBCF). BBCF works by embedding arbitrary (black-box) probabilistic models of human behavior in a latent variable Bayesian framework capable of collectively leveraging behavioral models trained on all users for personalized recommendation. There are three key advantages of BBCF compared to traditional CF and non-CF methods: (1) BBCF can leverage highly specialized behavioral models for specific CF use cases that may outperform existing generic models used in standard CF, (2) the behavioral models used in BBCF may offer enhanced intepretability and explainability compared to generic CF methods, and (3) compared to non-CF methods that would train a behavioral model per specific user and thus may suffer when individual user data is limited, BBCF leverages the data of all users thus enabling strong performance across the data availability spectrum including the near cold-start case. Experimentally, we compare BBCF to individual and global behavioral models as well as CF techniques; our evaluation domains span sequential and non-sequential tasks with a range of behavioral models for individual users, tasks, or goal-oriented behavior. Our results demonstrate that BBCF is competitive if not better than existing methods while still offering the interpretability and explainability benefits intrinsic to many behavioral models.

Adaptive user-interfaces (AUIs) can enhance the usability of complex software by providing real-time contextual adaptation and assistance. Ideally, AUIs should be personalized and versatile, i.e., able to adapt to each user who may perform a variety of complex tasks. But this is difficult to achieve with many interaction elements when data-per-user is sparse. In this paper, we propose an architecture for personalized AUIs that leverages upon developments in (1) deep learning, particularly gated recurrent units, to efficiently learn user interaction patterns, (2) collaborative filtering techniques that enable sharing of data among users, and (3) fast approximate nearest-neighbor methods in Euclidean spaces for quick UI control and/or content recommendations. Specifically, interaction histories are embedded in a learned space along with users and interaction elements; this allows the AUI to query and recommend likely next actions based on similar usage patterns across the user base. In a comparative evaluation on user-interface, web-browsing and e-learning datasets, the deep recurrent neural-network (DRNN) outperforms state-of-the-art tensor-factorization and metric embedding methods.

Decision support systems for network security represent a critical element in the safe operation of computer networks. Unfortunately, due to their complexity, it can be difficult to implement and empirically assess novel techniques for displaying networks. This paper details an open source adaptive user interface that hopes to fill this gap. This system supports agile development and offers a wide latitude for human factors and machine learning design modifications. The intent of this system is to serve as an experimental testbed for determining the efficacy of different human factors and machine learning initiatives on operator performance in network monitoring.

In this paper, we present a framework, probabilistic model, and algorithm for learning shared control policies by observing an assistant. This is a methodology we refer to as Learning Assistance by Demonstration (LAD). As a subset of robot Learning by Demonstration (LbD), LAD focuses on the assistive element by explicitly capturing how and when to help. The latter is especially important in assistive scenarios---such as rehabilitation and training---where there exists multiple and possibly conflicting goals. We formalize these notions in a probabilistic model and develop an efficient online mixture of experts (OME) algorithm, based on sparse Gaussian processes (GPs), for learning the assistive policy. Focusing on smart mobility, we couple the LAD methodology with a novel paired-haptic-controllers setup for helping smart wheelchair users navigate their environment. Experimental results with 15 able-bodied participants demonstrate that our learned shared control policy improved driving performance (as measured in lap seconds) by 43 s (a speedup of 191%). Furthermore, survey results indicate that the participants not only performed better quantitatively, but also qualitatively felt the model assistance helped them complete the task.

Successful biological systems adapt to change. In this paper, we are principally concerned with adaptive systems that operate in environments where data arrives sequentially and is multivariate in nature, for example, sensory streams in robotic systems. We contribute two reservoir inspired methods: 1) the online echostate Gaussian process (OESGP) and 2) its infinite variant, the online infinite echostate Gaussian process (OIESGP) Both algorithms are iterative fixed-budget methods that learn from noisy time series. In particular, the OESGP combines the echo-state network with Bayesian online learning for Gaussian processes. Extending this to infinite reservoirs yields the OIESGP, which uses a novel recursive kernel with automatic relevance determination that enables spatial and temporal feature weighting. When fused with stochastic natural gradient descent, the kernel hyperparameters are iteratively adapted to better model the target system. Furthermore, insights into the underlying system can be gleamed from inspection of the resulting hyperparameters. Experiments on noisy benchmark problems (one-step prediction and system identification) demonstrate that our methods yield high accuracies relative to state-of-the-art methods, and standard kernels with sliding windows, particularly on problems with irrelevant dimensions. In addition, we describe two case studies in robotic learning-by-demonstration involving the Nao humanoid robot and the Assistive Robot Transport for Youngsters (ARTY) smart wheelchair.

Crafting a proper assistance policy is a difficult endeavour but essential for the development of robotic assistants. Indeed, assistance is a complex issue that depends not only on the task-at-hand, but also on the state of the user, environment and competing objectives. As a way forward, this paper proposes learning the task of assistance through observation; an approach we term Learning Assistance by Demonstration (LAD). Our methodology is a subclass of Learning-by-Demonstration (LbD), yet directly addresses difficult issues associated with proper assistance such as when and how to appropriately assist. To learn assistive policies, we develop a probabilistic model that explicitly captures these elements and provide efficient, online, training methods. Experimental results on smart mobility assistance - using both simulation and a real-world smart wheelchair platform - demonstrate the effectiveness of our approach; the LAD model quickly learns when to assist (achieving an AUC score of 0.95 after only one demonstration) and improves with additional examples. Results show that this translates into better task-performance; our LAD-enabled smart wheelchair improved participant driving performance (measured in lap seconds) by 20.6s (a speedup of 137%), after a single teacher demonstration.

Recent studies on underactuated manipulation usually describe the system with a Kinematic Model (KM), which is built by adding external constraints to the standard manipulation analysis method. However, such external constraints are easily violated in a real-world dexterous manipulation task which results in significant control errors. In this work, the Enhanced Kinematic Model (E-KM), an integrated model of the KM and the Sparse Online Gaussian Process (SOGP) is proposed. The E-KM can compensate the shortfalls of the KM by on-the-fly training the SOGP on the residual between the prediction of the KM and the ground truth data. Based on the E-KM, we further contribute an optimal controller for underactuated manipulations. This optimal E-KM controller is implemented and tested on the iCub, a humanoid robot with two anthropomorphic underactuated hands. Two sets of real-world experiments are carried out to verify our method. The results demonstrate that the controller using E-KM statistically can achieve higher control accuracy than using solely using the KM for a wide range of objects.

In this work, we contribute the online echo state gaussian process (OESGP), a novel Bayesian-based online method that is capable of iteratively learning complex temporal dynamics and producing predictive distributions (instead of point predictions). Our method can be seen as a combination of the echo state network with a sparse approximation of Gaussian processes (GPs). Extensive experiments on the one-step prediction task on well-known benchmark problems show that OESGP produced statistically superior results to current online ESNs and state-of-the-art regression methods. In addition, we characterise the benefits (and drawbacks) associated with the considered online methods, specifically with regards to the trade-off between computational cost and accuracy. For a high-dimensional action recognition task, we demonstrate that OESGP produces high accuracies comparable to a recently published graphical model, while being fast enough for real-time interactive scenarios.

In this work, we are primarily concerned with robotic systems that learn online and continuously from multi-variate data-streams. Our first contribution is a new recursive kernel, which we have integrated into a sparse Gaussian Process to yield the Spatio-Temporal Online Recursive Kernel Gaussian Process (STORK-GP). This algorithm iteratively learns from time-series, providing both predictions and uncertainty estimates. Experiments on benchmarks demonstrate that our method achieves high accuracies relative to state-of-the-art methods. Second, we contribute an online tactile classifier which uses an array of STORK-GP experts. In contrast to existing work, our classifier is capable of learning new objects as they are presented, improving itself over time. We show that our approach yields results comparable to highly-optimised offline classification methods. Moreover, we conducted experiments with human subjects in a similar online setting with true-label feedback and present the insights gained.

Plans and decisions in many real-world scenarios are made under uncertainty and to satisfy multiple, possibly conflicting, objectives. In this work, we contribute the multi-reward partially-observable Markov decision process (MR-POMDP) as a general modelling framework. To solve MR-POMDPs, we present two hybrid (memetic) multi-objective evolutionary algorithms that generate non-dominated sets of policies (in the form of stochastic finite state controllers). Performance comparisons between the methods on multi-objective problems in robotics (with 2, 3 and 5 objectives), web-advertising (with 3, 4 and 5 objectives) and infectious disease control (with 3 objectives), revealed that memetic variants outperformed their original counterparts. We anticipate that the MR-POMDP along with multi-objective evolutionary solvers will prove useful in a variety of theoretical and real-world applications.

The discovery of low-energy stable and meta-stable molecular structures remains an important and unsolved problem in search and optimization. In this paper, we contribute two stochastic algorithms, the archiving molecular memetic algorithm (AMMA) and the archiving basin hopping algorithm (ABHA) for sampling low-energy isomers on the landscapes of pure water clusters (H2O)n. We applied our methods to two sophisticated empirical water cluster models, TTM2.1-F and OSS2, and generated archives of low-energy water isomers (H2O)n n=3-15. Our algorithms not only reproduced previously-found best minima, but also discovered new global minima candidates for sizes 9-15 on OSS2. Further numerical results show that AMMA and ABHA outperformed a baseline stochastic multistart local search algorithm in terms of convergence and isomer archival. Noting a performance differential between TTM2.1-F and OSS2, we analyzed both model landscapes to reveal that the global and local correlation properties of the empirical models differ significantly. In particular, the OSS2 landscape was less correlated and hence, more difficult to explore and optimize. Guided by our landscape analyses, we proposed and demonstrated the effectiveness of a hybrid local search algorithm, which significantly improved the sampling performance of AMMA on the larger OSS2 landscapes. Although applied to pure water clusters in this paper, AMMA and ABHA can be easily modified for subsequent studies in computational chemistry and biology. Moreover, the landscape analyses conducted in this paper can be replicated for other molecular systems to uncover landscape properties and provide insights to both physical chemists and evolutionary algorithmists.

The design and engineering of complex materials and products often requires intricate interactions between domain experts in science, material and engineering as well as the utilization of diverse software systems for discovery and optimization. If left as it is, design engineers would most likely be at a loss on how to engage the entire entourage of the multi- disciplinary processes as well as the compute-intensive and data-intensive nature of the activities involved. This paper describes a possible solution through the development of a Grid-based Problem Solving Environment for Engineering of Materials (GPEM). The GPEM aims to provide a one-stop platform where engineers will perform material discovery, design optimization and material characterization, with grid computing as the enabling technology. Upon describing the details of the process workflow and the adopted architecture design, the paper will present the current implementation of GPEM, in the design optimization of fractal structures.

Dengue represents a serious health threat in the Tropics, owing to the year-round presence of Aedes mosquito vectors, and the lack of any anti-viral drugs or vaccines. Climatic factors are important in influencing the incidence of dengue. It is important to determine the relationships between climatic factors and disease incidence trends, which would be helpful for relevant environment and health agencies in planning appropriate pre-emptive control measures. Climatic factors and dengue case records vary over time. It is therefore difficult to justify the time-lag when a climatic factor affects the mosquito-to-human and human-to-mosquito loops. In this paper, we propose to use support vector machine (SVM) classifiers for analyzing the time- series dengue data and genetic algorithm (GA), to determine the time-lags and subset of climatic factors as effective factors influencing the spread of dengue. It is shown that the proposed model is able to detect important climatic factors and their time-lags which affect the disease, and the GA-based SVM classifiers could improve the classification accuracy significantly.

This paper describes a multi-objective parameter-less genetic algorithm (moPGA), which combines several recent developments including efficient non-dominated sorting, linkage learning, isin-Dominance, building-block mutation and convergence detection. Additionally, a novel method of clustering in the objective space using an isin-Pareto Set is introduced. Comparisons with well-known multi-objective GAs on scalable benchmark problems indicate that the algorithm scales well with problem size in terms of number of function evaluations and quality of solutions found. moPGA was built for easy usage and hence, in addition to the problem function and encoding, there are only two required user defined parameters; (1) the maximum running time or generations and (2) the precision of the desired solutions (isin).