Subramanian Ramamoorthy

Salvatore Esposito, Matías Mattamala, Daniel Rebain et al.

Continuum robots are advancing bronchoscopy procedures by accessing complex lung airways and enabling targeted interventions. However, their development is limited by the lack of realistic training and test environments: Real data is difficult to collect due to ethical constraints and patient safety concerns, and developing autonomy algorithms requires realistic imaging and physical feedback. We present ROOM (Realistic Optical Observation in Medicine), a comprehensive simulation framework designed for generating photorealistic bronchoscopy training data. By leveraging patient CT scans, our pipeline renders multi-modal sensor data including RGB images with realistic noise and light specularities, metric depth maps, surface normals, optical flow and point clouds at medically relevant scales. We validate the data generated by ROOM in two canonical tasks for medical robotics: multi-view pose estimation and monocular depth estimation, demonstrating diverse challenges that state-of-the-art methods must overcome to transfer to these medical settings. Furthermore, we show that the data produced by ROOM can be used to fine-tune existing depth estimation models to overcome these challenges, also enabling other downstream applications such as navigation. We expect that ROOM will enable large-scale data generation across diverse patient anatomies and procedural scenarios that are challenging to capture in clinical settings. Code and data: https://github.com/iamsalvatore/room.

Dhaminda B. Abeywickrama, Amel Bennaceur, Greg Chance et al.

As autonomous systems (AS) increasingly become part of our daily lives, ensuring their trustworthiness is crucial. In order to demonstrate the trustworthiness of an AS, we first need to specify what is required for an AS to be considered trustworthy. This roadmap paper identifies key challenges for specifying for trustworthiness in AS, as identified during the "Specifying for Trustworthiness" workshop held as part of the UK Research and Innovation (UKRI) Trustworthy Autonomous Systems (TAS) programme. We look across a range of AS domains with consideration of the resilience, trust, functionality, verifiability, security, and governance and regulation of AS and identify some of the key specification challenges in these domains. We then highlight the intellectual challenges that are involved with specifying for trustworthiness in AS that cut across domains and are exacerbated by the inherent uncertainty involved with the environments in which AS need to operate.

Anthony L. Corso, Sydney M. Katz, Craig Innes et al.

Modern autonomous systems rely on perception modules to process complex sensor measurements into state estimates. These estimates are then passed to a controller, which uses them to make safety-critical decisions. It is therefore important that we design perception systems to minimize errors that reduce the overall safety of the system. We develop a risk-driven approach to designing perception systems that accounts for the effect of perceptual errors on the performance of the fully-integrated, closed-loop system. We formulate a risk function to quantify the effect of a given perceptual error on overall safety, and show how we can use it to design safer perception systems by including a risk-dependent term in the loss function and generating training data in risk-sensitive regions. We evaluate our techniques on a realistic vision-based aircraft detect and avoid application and show that risk-driven design reduces collision risk by 37% over a baseline system.

Aravinda Ramakrishnan Srinivasan, Yi-Shin Lin, Morris Antonello et al.

Autonomous vehicles use a variety of sensors and machine-learned models to predict the behavior of surrounding road users. Most of the machine-learned models in the literature focus on quantitative error metrics like the root mean square error (RMSE) to learn and report their models' capabilities. This focus on quantitative error metrics tends to ignore the more important behavioral aspect of the models, raising the question of whether these models really predict human-like behavior. Thus, we propose to analyze the output of machine-learned models much like we would analyze human data in conventional behavioral research. We introduce quantitative metrics to demonstrate presence of three different behavioral phenomena in a naturalistic highway driving dataset: 1) The kinematics-dependence of who passes a merging point first 2) Lane change by an on-highway vehicle to accommodate an on-ramp vehicle 3) Lane changes by vehicles on the highway to avoid lead vehicle conflicts. Then, we analyze the behavior of three machine-learned models using the same metrics. Even though the models' RMSE value differed, all the models captured the kinematic-dependent merging behavior but struggled at varying degrees to capture the more nuanced courtesy lane change and highway lane change behavior. Additionally, the collision aversion analysis during lane changes showed that the models struggled to capture the physical aspect of human driving: leaving adequate gap between the vehicles. Thus, our analysis highlighted the inadequacy of simple quantitative metrics and the need to take a broader behavioral perspective when analyzing machine-learned models of human driving predictions.

Craig Innes, Subramanian Ramamoorthy

Testing black-box perceptual-control systems in simulation faces two difficulties. Firstly, perceptual inputs in simulation lack the fidelity of real-world sensor inputs. Secondly, for a reasonably accurate perception system, encountering a rare failure trajectory may require running infeasibly many simulations. This paper combines perception error models -- surrogates for a sensor-based detection system -- with state-dependent adaptive importance sampling. This allows us to efficiently assess the rare failure probabilities for real-world perceptual control systems within simulation. Our experiments with an autonomous braking system equipped with an RGB obstacle-detector show that our method can calculate accurate failure probabilities with an inexpensive number of simulations. Further, we show how choice of safety metric can influence the process of learning proposal distributions capable of reliably sampling high-probability failures.

Rimvydas Rubavicius, Manisha Dubey, N. Siddharth et al.

Human-centred systems require an understanding of human actions in the physical world. Temporally extended sequences of actions are intentional and structured, yet existing methods for recognising what actions are performed often do not attempt to capture their structure, particularly how the actions are executed. This, however, is crucial for assessing the quality of the action's execution and its differences from other actions. To capture the internal mechanics of actions, we introduce a domain-specific language EXACT that represents human motions as underspecified motion programs, interpreted as reward-generating functions for zero-shot policy inference using forward-backwards representations. By leveraging the compositional nature of EXACT motion programs, we combine individual policies into an executable neuro-symbolic model that uses program structure for compositional modelling. We evaluate the utility of the proposed pipeline for creating executable action models by analysing motion-capture data to understand human actions, for the tasks of human action segmentation and action anomaly detection. Our results show that the use of executable action models improves data efficiency and captures intuitive relationships between actions compared with monolithic, task-specific approaches.

Rimvydas Rubavicius, Peter David Fagan, Alex Lascarides et al.

This paper addresses a challenging interactive task learning scenario we call rearrangement under unawareness: an agent must manipulate a rigid-body environment without knowing a key concept necessary for solving the task and must learn about it during deployment. For example, the user may ask to "put the two granny smith apples inside the basket", but the agent cannot correctly identify which objects in the environment are "granny smith" as the agent has not been exposed to such a concept before. We introduce SECURE, an interactive task learning policy designed to tackle such scenarios. The unique feature of SECURE is its ability to enable agents to engage in semantic analysis when processing embodied conversations and making decisions. Through embodied conversation, a SECURE agent adjusts its deficient domain model by engaging in dialogue to identify and learn about previously unforeseen possibilities. The SECURE agent learns from the user's embodied corrective feedback when mistakes are made and strategically engages in dialogue to uncover useful information about novel concepts relevant to the task. These capabilities enable the SECURE agent to generalize to new tasks with the acquired knowledge. We demonstrate in the simulated Blocksworld and the real-world apple manipulation environments that the SECURE agent, which solves such rearrangements under unawareness, is more data-efficient than agents that do not engage in embodied conversation or semantic analysis.

Chen Long-fei, Subramanian Ramamoorthy, Robert B Fisher

Vision-based motion estimation methods show promise in accurately and unobtrusively estimating human body motion for healthcare purposes. However, these methods are not specifically designed for healthcare purposes and face challenges in real-world applications. Human pose estimation methods often lack the accuracy needed for detecting fine-grained, subtle body movements, while optical flow-based methods struggle with poor lighting conditions and unseen real-world data. These issues result in human body motion estimation errors, particularly during critical medical situations where the body is motionless, such as during unconsciousness. To address these challenges and improve the accuracy of human body motion estimation for healthcare purposes, we propose the OPPH operator designed to enhance current vision-based motion estimation methods. This operator, which considers human body movement and noise properties, functions as a multi-stage filter. Results tested on two real-world and one synthetic human motion dataset demonstrate that the operator effectively removes real-world noise, significantly enhances the detection of motionless states, maintains the accuracy of estimating active body movements, and maintains long-term body movement trends. This method could be beneficial for analyzing both critical medical events and chronic medical conditions.

Georgios Kamaras, Craig Innes, Subramanian Ramamoorthy

In robotics, likelihood-free inference (LFI) can provide the domain distribution that adapts a learnt agent in a parametric set of deployment conditions. LFI assumes an arbitrary support for sampling, which remains constant as the initial generic prior is iteratively refined to more descriptive posteriors. However, a potentially misspecified support can lead to suboptimal, yet falsely certain, posteriors. To address this issue, we propose three heuristic LFI variants: EDGE, MODE, and CENTRE. Each interprets the posterior mode shift over inference steps in its own way and, when integrated into an LFI step, adapts the support alongside posterior inference. We first expose the support misspecification issue and evaluate our heuristics using stochastic dynamical benchmarks. We then evaluate the impact of heuristic support adaptation on parameter inference and policy learning for a dynamic deformable linear object (DLO) manipulation task. Inference results in a finer length and stiffness classification for a parametric set of DLOs. When the resulting posteriors are used as domain distributions for sim-based policy learning, they lead to more robust object-centric agent performance.

Peter David Fagan, Subramanian Ramamoorthy

Learning from Demonstration (LfD) is a useful paradigm for training policies that solve tasks involving complex motions, such as those encountered in robotic manipulation. In practice, the successful application of LfD requires overcoming error accumulation during policy execution, i.e. the problem of drift due to errors compounding over time and the consequent out-of-distribution behaviours. Existing works seek to address this problem through scaling data collection, correcting policy errors with a human-in-the-loop, temporally ensembling policy predictions or through learning a dynamical system model with convergence guarantees. In this work, we propose and validate an alternative approach to overcoming this issue. Inspired by reservoir computing, we develop a recurrent neural network layer that includes a fixed nonlinear dynamical system with tunable dynamical properties for modelling temporal dynamics. We validate the efficacy of our neural network layer on the task of reproducing human handwriting motions using the LASA Human Handwriting Dataset. Through empirical experiments we demonstrate that incorporating our layer into existing neural network architectures addresses the issue of compounding errors in LfD. Furthermore, we perform a comparative evaluation against existing approaches including a temporal ensemble of policy predictions and an Echo State Network (ESN) implementation. We find that our approach yields greater policy precision and robustness on the handwriting task while also generalising to multiple dynamics regimes and maintaining competitive latency scores.

Logan Mondal Bhamidipaty, Esmeralda S. Whitammer, David Abel et al.

We propose a novel definition of model exploitation in reinforcement learning. Informally, a world model is exploitable if it implies that one policy should be strictly preferred over another while the environment's true transition model implies the reverse. We analogize our definition with a prior characterization of reward hacking but show that the associated proof of inevitability does not transfer to exploitation. To overcome this obstruction, we develop a general theory of reward hacking and model exploitation that proves that exploitation is essentially unavoidable on large policy sets and yields the corresponding claim for hacking as a special case. Unfortunately, we also find that the conditions that guarantee unhackability in finite policy sets have no counterpart that precludes exploitation. Consequently, we introduce a relaxed notion of exploitation and derive a safe horizon within which it can be avoided. Taken together, our results establish a formal bridge between reward hacking and model exploitation and elucidate the limits of safe planning in world models.

Leonard Hinckeldey, Elliot Fosong, Elle Miller et al.

The development of reinforcement learning (RL) algorithms has been largely driven by ambitious challenge tasks and benchmarks. Games have dominated RL benchmarks because they present relevant challenges, are inexpensive to run and easy to understand. While games such as Go and Atari have led to many breakthroughs, they often do not directly translate to real-world embodied applications. In recognising the need to diversify RL benchmarks and addressing complexities that arise in embodied interaction scenarios, we introduce Assistax: an open-source benchmark designed to address challenges arising in assistive robotics tasks. Assistax uses JAX's hardware acceleration for significant speed-ups for learning in physics-based simulations. In terms of open-loop wall-clock time, Assistax runs up to $370\times$ faster when vectorising training runs compared to CPU-based alternatives. Assistax conceptualises the interaction between an assistive robot and an active human patient using multi-agent RL to train a population of diverse partner agents against which an embodied robotic agent's zero-shot coordination capabilities can be tested. Extensive evaluation and hyperparameter tuning for popular continuous control RL and MARL algorithms provide reliable baselines and establish Assistax as a practical benchmark for advancing RL research for assistive robotics. The code is available at: https://github.com/assistive-autonomy/assistax.

Nikolaos Tsagkas, Jack Rome, Subramanian Ramamoorthy et al.

Precise manipulation that is generalizable across scenes and objects remains a persistent challenge in robotics. Current approaches for this task heavily depend on having a significant number of training instances to handle objects with pronounced visual and/or geometric part ambiguities. Our work explores the grounding of fine-grained part descriptors for precise manipulation in a zero-shot setting by utilizing web-trained text-to-image diffusion-based generative models. We tackle the problem by framing it as a dense semantic part correspondence task. Our model returns a gripper pose for manipulating a specific part, using as reference a user-defined click from a source image of a visually different instance of the same object. We require no manual grasping demonstrations as we leverage the intrinsic object geometry and features. Practical experiments in a real-world tabletop scenario validate the efficacy of our approach, demonstrating its potential for advancing semantic-aware robotics manipulation. Web page: https://tsagkas.github.io/click2grasp

Rimvydas Rubavicius, Antonio Valerio Miceli-Barone, Alex Lascarides et al.

Cyber-physical systems like autonomous vehicles are tested in simulation before deployment, using domain-specific programs for scenario specification. To aid the testing of autonomous vehicles in simulation, we design a natural language interface, using an instruction-following large language model, to assist a non-coding domain expert in synthesising the desired scenarios and vehicle behaviours. We show that using it to convert utterances to the symbolic program is feasible, despite the very small training dataset. Human experiments show that dialogue is critical to successful simulation generation, leading to a 4.5 times higher success rate than a generation without engaging in extended conversation.

Craig Innes, Subramanian Ramamoorthy

Autonomous Vehicles (AVs) are often tested in simulation to estimate the probability they will violate safety specifications. Two common issues arise when using existing techniques to produce this estimation: If violations occur rarely, simple Monte-Carlo sampling techniques can fail to produce efficient estimates; if simulation horizons are too long, importance sampling techniques (which learn proposal distributions from past simulations) can fail to converge. This paper addresses both issues by interleaving rare-event sampling techniques with online specification monitoring algorithms. We use adaptive multi-level splitting to decompose simulations into partial trajectories, then calculate the distance of those partial trajectories to failure by leveraging robustness metrics from Signal Temporal Logic (STL). By caching those partial robustness metric values, we can efficiently re-use computations across multiple sampling stages. Our experiments on an interstate lane-change scenario show our method is viable for testing simulated AV-pipelines, efficiently estimating failure probabilities for STL specifications based on real traffic rules. We produce better estimates than Monte-Carlo and importance sampling in fewer simulations.

Georgios Kamaras, Subramanian Ramamoorthy

We present an integrated (or end-to-end) framework for the Real2Sim2Real problem of manipulating deformable linear objects (DLOs) based on visual perception. Working with a parameterised set of DLOs, we use likelihood-free inference (LFI) to compute the posterior distributions for the physical parameters using which we can approximately simulate the behaviour of each specific DLO. We use these posteriors for domain randomisation while training, in simulation, object-specific visuomotor policies (i.e. assuming only visual and proprioceptive sensory) for a DLO reaching task, using model-free reinforcement learning. We demonstrate the utility of this approach by deploying sim-trained DLO manipulation policies in the real world in a zero-shot manner, i.e. without any further fine-tuning. In this context, we evaluate the capacity of a prominent LFI method to perform fine classification over the parametric set of DLOs, using only visual and proprioceptive data obtained in a dynamic manipulation trajectory. We then study the implications of the resulting domain distributions in sim-based policy learning and real-world performance.

Jonghyuk Park, Alex Lascarides, Subramanian Ramamoorthy

In this paper, we offer a learning framework in which the agent's knowledge gaps are overcome through corrective feedback from a teacher whenever the agent explains its (incorrect) predictions. We test it in a low-resource visual processing scenario, in which the agent must learn to recognize distinct types of toy truck. The agent starts the learning process with no ontology about what types of trucks exist nor which parts they have, and a deficient model for recognizing those parts from visual input. The teacher's feedback to the agent's explanations addresses its lack of relevant knowledge in the ontology via a generic rule (e.g., "dump trucks have dumpers"), whereas an inaccurate part recognition is corrected by a deictic statement (e.g., "this is not a dumper"). The learner utilizes this feedback not only to improve its estimate of the hypothesis space of possible domain ontologies and probability distributions over them, but also to use those estimates to update its visual interpretation of the scene. Our experiments demonstrate that teacher-learner pairs utilizing explanations and corrections are more data-efficient than those without such a faculty.

Shuai Li, Michael Burke, Subramanian Ramamoorthy et al.

This paper introduces a novel framework to learn data association for multi-object tracking in a self-supervised manner. Fully-supervised learning methods are known to achieve excellent tracking performances, but acquiring identity-level annotations is tedious and time-consuming. Motivated by the fact that in real-world scenarios object motion can be usually represented by a Markov process, we present a novel expectation maximization (EM) algorithm that trains a neural network to associate detections for tracking, without requiring prior knowledge of their temporal correspondences. At the core of our method lies a neural Kalman filter, with an observation model conditioned on associations of detections parameterized by a neural network. Given a batch of frames as input, data associations between detections from adjacent frames are predicted by a neural network followed by a Sinkhorn normalization that determines the assignment probabilities of detections to states. Kalman smoothing is then used to obtain the marginal probability of observations given the inferred states, producing a training objective to maximize this marginal probability using gradient descent. The proposed framework is fully differentiable, allowing the underlying neural model to be trained end-to-end. We evaluate our approach on the challenging MOT17, MOT20, and BDD100K datasets and achieve state-of-the-art results in comparison to self-supervised trackers using public detections.

Chen Long-fei, Muhammad Ahmed Raza, Craig Innes et al.

Aging and chronic conditions affect older adults' daily lives, making early detection of developing health issues crucial. Weakness, common in many conditions, alters physical movements and daily activities subtly. However, detecting such changes can be challenging due to their subtle and gradual nature. To address this, we employ a non-intrusive camera sensor to monitor individuals' daily sitting and relaxing activities for signs of weakness. We simulate weakness in healthy subjects by having them perform physical exercise and observing the behavioral changes in their daily activities before and after workouts. The proposed system captures fine-grained features related to body motion, inactivity, and environmental context in real-time while prioritizing privacy. A Bayesian Network is used to model the relationships between features, activities, and health conditions. We aim to identify specific features and activities that indicate such changes and determine the most suitable time scale for observing the change. Results show 0.97 accuracy in distinguishing simulated weakness at the daily level. Fine-grained behavioral features, including non-dominant upper body motion speed and scale, and inactivity distribution, along with a 300-second window, are found most effective. However, individual-specific models are recommended as no universal set of optimal features and activities was identified across all participants.

Anthony Knittel, Morris Antonello, John Redford et al.

The ability to anticipate pedestrian motion changes is a critical capability for autonomous vehicles. In urban environments, pedestrians may enter the road area and create a high risk for driving, and it is important to identify these cases. Typical predictors use the trajectory history to predict future motion, however in cases of motion initiation, motion in the trajectory may only be clearly visible after a delay, which can result in the pedestrian has entered the road area before an accurate prediction can be made. Appearance data includes useful information such as changes of gait, which are early indicators of motion changes, and can inform trajectory prediction. This work presents a comparative evaluation of trajectory-only and appearance-based methods for pedestrian prediction, and introduces a new dataset experiment for prediction using appearance. We create two trajectory and image datasets based on the combination of image and trajectory sequences from the popular NuScenes dataset, and examine prediction of trajectories using observed appearance to influence futures. This shows some advantages over trajectory prediction alone, although problems with the dataset prevent advantages of appearance-based models from being shown. We describe methods for improving the dataset and experiment to allow benefits of appearance-based models to be captured.

Jonghyuk Park, Alex Lascarides, Subramanian Ramamoorthy

Interactive Task Learning (ITL) concerns learning about unforeseen domain concepts via natural interactions with human users. The learner faces a number of significant constraints: learning should be online, incremental and few-shot, as it is expected to perform tangible belief updates right after novel words denoting unforeseen concepts are introduced. In this work, we explore a challenging symbol grounding task--discriminating among object classes that look very similar--within the constraints imposed by ITL. We demonstrate empirically that more data-efficient grounding results from exploiting the truth-conditions of the teacher's generic statements (e.g., "Xs have attribute Z.") and their implicatures in context (e.g., as an answer to "How are Xs and Ys different?", one infers Y lacks attribute Z).

Manu Lahariya, Craig Innes, Chris Develder et al.

Soft actuators offer a safe, adaptable approach to tasks like gentle grasping and dexterous manipulation. Creating accurate models to control such systems however is challenging due to the complex physics of deformable materials. Accurate Finite Element Method (FEM) models incur prohibitive computational complexity for closed-loop use. Using a differentiable simulator is an attractive alternative, but their applicability to soft actuators and deformable materials remains underexplored. This paper presents a framework that combines the advantages of both. We learn a differentiable model consisting of a material properties neural network and an analytical dynamics model of the remainder of the manipulation task. This physics-informed model is trained using data generated from FEM, and can be used for closed-loop control and inference. We evaluate our framework on a dielectric elastomer actuator (DEA) coin-pulling task. We simulate the task of using DEA to pull a coin along a surface with frictional contact, using FEM, and evaluate the physics-informed model for simulation, control, and inference. Our model attains < 5% simulation error compared to FEM, and we use it as the basis for an MPC controller that requires fewer iterations to converge than model-free actor-critic, PD, and heuristic policies.

Luca Viano, Yu-Ting Huang, Parameswaran Kamalaruban et al.

Imitation learning (IL) is a popular paradigm for training policies in robotic systems when specifying the reward function is difficult. However, despite the success of IL algorithms, they impose the somewhat unrealistic requirement that the expert demonstrations must come from the same domain in which a new imitator policy is to be learned. We consider a practical setting, where (i) state-only expert demonstrations from the real (deployment) environment are given to the learner, (ii) the imitation learner policy is trained in a simulation (training) environment whose transition dynamics is slightly different from the real environment, and (iii) the learner does not have any access to the real environment during the training phase beyond the batch of demonstrations given. Most of the current IL methods, such as generative adversarial imitation learning and its state-only variants, fail to imitate the optimal expert behavior under the above setting. By leveraging insights from the Robust reinforcement learning (RL) literature and building on recent adversarial imitation approaches, we propose a robust IL algorithm to learn policies that can effectively transfer to the real environment without fine-tuning. Furthermore, we empirically demonstrate on continuous-control benchmarks that our method outperforms the state-of-the-art state-only IL method in terms of the zero-shot transfer performance in the real environment and robust performance under different testing conditions.

Xin Du, Benedicte Legastelois, Bhargavi Ganesh et al.

Using large pre-trained models for image recognition tasks is becoming increasingly common owing to the well acknowledged success of recent models like vision transformers and other CNN-based models like VGG and Resnet. The high accuracy of these models on benchmark tasks has translated into their practical use across many domains including safety-critical applications like autonomous driving and medical diagnostics. Despite their widespread use, image models have been shown to be fragile to changes in the operating environment, bringing their robustness into question. There is an urgent need for methods that systematically characterise and quantify the capabilities of these models to help designers understand and provide guarantees about their safety and robustness. In this paper, we propose Vision Checklist, a framework aimed at interrogating the capabilities of a model in order to produce a report that can be used by a system designer for robustness evaluations. This framework proposes a set of perturbation operations that can be applied on the underlying data to generate test samples of different types. The perturbations reflect potential changes in operating environments, and interrogate various properties ranging from the strictly quantitative to more qualitative. Our framework is evaluated on multiple datasets like Tinyimagenet, CIFAR10, CIFAR100 and Camelyon17 and for models like ViT and Resnet. Our Vision Checklist proposes a specific set of evaluations that can be integrated into the previously proposed concept of a model card. Robustness evaluations like our checklist will be crucial in future safety evaluations of visual perception modules, and be useful for a wide range of stakeholders including designers, deployers, and regulators involved in the certification of these systems. Source code of Vision Checklist would be open for public use.

Allison McCarn Deiana, Nhan Tran, Joshua Agar et al.

In this community review report, we discuss applications and techniques for fast machine learning (ML) in science -- the concept of integrating power ML methods into the real-time experimental data processing loop to accelerate scientific discovery. The material for the report builds on two workshops held by the Fast ML for Science community and covers three main areas: applications for fast ML across a number of scientific domains; techniques for training and implementing performant and resource-efficient ML algorithms; and computing architectures, platforms, and technologies for deploying these algorithms. We also present overlapping challenges across the multiple scientific domains where common solutions can be found. This community report is intended to give plenty of examples and inspiration for scientific discovery through integrated and accelerated ML solutions. This is followed by a high-level overview and organization of technical advances, including an abundance of pointers to source material, which can enable these breakthroughs.

Jack Geary, Henry Gouk, Subramanian Ramamoorthy

Safe interaction between vehicles requires the ability to choose actions that reveal the preferences of the other vehicles. Since exploratory actions often do not directly contribute to their objective, an interactive vehicle must also able to identify when it is appropriate to perform them. In this work we demonstrate how Active Learning methods can be used to incentivise an autonomous vehicle (AV) to choose actions that reveal information about the altruistic inclinations of another vehicle. We identify a property, Information Sufficiency, that a reward function should have in order to keep exploration from unnecessarily interfering with the pursuit of an objective. We empirically demonstrate that reward functions that do not have Information Sufficiency are prone to inadequate exploration, which can result in sub-optimal behaviour. We propose a reward definition that has Information Sufficiency, and show that it facilitates an AV choosing exploratory actions to estimate altruistic tendency, whilst also compensating for the possibility of conflicting beliefs between vehicles.

Xin Du, Subramanian Ramamoorthy, Wouter Duivesteijn et al.

Local decision rules are commonly understood to be more explainable, due to the local nature of the patterns involved. With numerical optimization methods such as gradient boosting, ensembles of local decision rules can gain good predictive performance on data involving global structure. Meanwhile, machine learning models are being increasingly used to solve problems in high-stake domains including healthcare and finance. Here, there is an emerging consensus regarding the need for practitioners to understand whether and how those models could perform robustly in the deployment environments, in the presence of distributional shifts. Past research on local decision rules has focused mainly on maximizing discriminant patterns, without due consideration of robustness against distributional shifts. In order to fill this gap, we propose a new method to learn and ensemble local decision rules, that are robust both in the training and deployment environments. Specifically, we propose to leverage causal knowledge by regarding the distributional shifts in subpopulations and deployment environments as the results of interventions on the underlying system. We propose two regularization terms based on causal knowledge to search for optimal and stable rules. Experiments on both synthetic and benchmark datasets show that our method is effective and robust against distributional shifts in multiple environments.

Craig Innes, Subramanian Ramamoorthy

Many robot control scenarios involve assessing system robustness against a task specification. If either the controller or environment are composed of "black-box" components with unknown dynamics, we cannot rely on formal verification to assess our system. Assessing robustness via exhaustive testing is also often infeasible if the space of environments is large compared to experiment cost. Given limited budget, we provide a method to choose experiment inputs which give greatest insight into system performance against a given specification across the domain. By combining smooth robustness metrics for signal temporal logic with techniques from adaptive experiment design, our method chooses the most informative experimental inputs by incrementally constructing a surrogate model of the specification robustness. This model then chooses the next experiment to be in an area where there is either high prediction error or uncertainty. Our experiments show how this adaptive experimental design technique results in sample-efficient descriptions of system robustness. Further, we show how to use the model built via the experiment design process to assess the behaviour of a data-driven control system under domain shift.

Simón C. Smith, Subramanian Ramamoorthy

Understanding a controller's performance in different scenarios is crucial for robots that are going to be deployed in safety-critical tasks. If we do not have a model of the dynamics of the world, which is often the case in complex domains, we may need to approximate a performance function of the robot based on its interaction with the environment. Such a performance function gives us insights into the behaviour of the robot, allowing us to fine-tune the controller with manual interventions. In high-dimensionality systems, where the actionstate space is large, fine-tuning a controller is non-trivial. To overcome this problem, we propose a performance function whose domain is defined by external features and parameters of the controller. Attainment regions are defined over such a domain defined by feature-parameter pairs, and serve the purpose of enabling prediction of successful execution of the task. The use of the feature-parameter space -in contrast to the action-state space- allows us to adapt, explain and finetune the controller over a simpler (i.e., lower dimensional space). When the robot successfully executes the task, we use the attainment regions to gain insights into the limits of the controller, and its robustness. When the robot fails to execute the task, we use the regions to debug the controller and find adaptive and counterfactual changes to the solutions. Another advantage of this approach is that we can generalise through the use of Gaussian processes regression of the performance function in the high-dimensional space. To test our approach, we demonstrate learning an approximation to the performance function in simulation, with a mobile robot traversing different terrain conditions. Then, with a sample-efficient method, we propagate the attainment regions to a physical robot in a similar environment.

Josiah P. Hanna, Arrasy Rahman, Elliot Fosong et al.

Recognising the goals or intentions of observed vehicles is a key step towards predicting the long-term future behaviour of other agents in an autonomous driving scenario. When there are unseen obstacles or occluded vehicles in a scenario, goal recognition may be confounded by the effects of these unseen entities on the behaviour of observed vehicles. Existing prediction algorithms that assume rational behaviour with respect to inferred goals may fail to make accurate long-horizon predictions because they ignore the possibility that the behaviour is influenced by such unseen entities. We introduce the Goal and Occluded Factor Inference (GOFI) algorithm which bases inference on inverse-planning to jointly infer a probabilistic belief over goals and potential occluded factors. We then show how these beliefs can be integrated into Monte Carlo Tree Search (MCTS). We demonstrate that jointly inferring goals and occluded factors leads to more accurate beliefs with respect to the true world state and allows an agent to safely navigate several scenarios where other baselines take unsafe actions leading to collisions.

Todor Davchev, Sarah Bechtle, Subramanian Ramamoorthy et al.

Inverse reinforcement learning is a paradigm motivated by the goal of learning general reward functions from demonstrated behaviours. Yet the notion of generality for learnt costs is often evaluated in terms of robustness to various spatial perturbations only, assuming deployment at fixed speeds of execution. However, this is impractical in the context of robotics and building, time-invariant solutions is of crucial importance. In this work, we propose a formulation that allows us to 1) vary the length of execution by learning time-invariant costs, and 2) relax the temporal alignment requirements for learning from demonstration. We apply our method to two different types of cost formulations and evaluate their performance in the context of learning reward functions for simulated placement and peg in hole tasks executed on a 7DoF Kuka IIWA arm. Our results show that our approach enables learning temporally invariant rewards from misaligned demonstration that can also generalise spatially to out of distribution tasks.

Paola Ardón, Èric Pairet, Katrin S. Lohan et al.

Affordances describe the possibilities for an agent to perform actions with an object. While the significance of the affordance concept has been previously studied from varied perspectives, such as psychology and cognitive science, these approaches are not always sufficient to enable direct transfer, in the sense of implementations, to artificial intelligence (AI)-based systems and robotics. However, many efforts have been made to pragmatically employ the concept of affordances, as it represents great potential for AI agents to effectively bridge perception to action. In this survey, we review and find common ground amongst different strategies that use the concept of affordances within robotic tasks, and build on these methods to provide guidance for including affordances as a mechanism to improve autonomy. To this end, we outline common design choices for building representations of affordance relations, and their implications on the generalisation capabilities of an agent when facing previously unseen scenarios. Finally, we identify and discuss a range of interesting research directions involving affordances that have the potential to improve the capabilities of an AI agent.

Michael Burke, Subramanian Ramamoorthy

Data association is a fundamental component of effective multi-object tracking. Current approaches to data-association tend to frame this as an assignment problem relying on gating and distance-based cost matrices, or offset the challenge of data association to a problem of tracking by detection. The latter is typically formulated as a supervised learning problem, and requires labelling information about tracked object identities to train a model for object recognition. This paper introduces an expectation maximisation approach to train neural models for data association, which does not require labelling information. Here, a Sinkhorn network is trained to predict assignment matrices that maximise the marginal likelihood of trajectory observations. Importantly, networks trained using the proposed approach can be re-used in downstream tracking applications.

John Hartley, Hubert P. H. Shum, Edmond S. L. Ho et al.

Existing studies on formation control for unmanned aerial vehicles (UAV) have not considered encircling targets where an optimum coverage of the target is required at all times. Such coverage plays a critical role in many real-world applications such as tracking hostile UAVs. This paper proposes a new path planning approach called the Flux Guided (FG) method, which generates collision-free trajectories for multiple UAVs while maximising the coverage of target(s). Our method enables UAVs to track directly toward a target whilst maintaining maximum coverage. Furthermore, multiple scattered targets can be tracked by scaling the formation during flight. FG is highly scalable since it only requires communication between sub-set of UAVs on the open boundary of the formation's surface. Experimental results further validate that FG generates UAV trajectories $1.5 \times$ shorter than previous work and that trajectory planning for 9 leader/follower UAVs to surround a target in two different scenarios only requires 0.52 seconds and 0.88 seconds, respectively. The resulting trajectories are suitable for robotic controls after time-optimal parameterisation; we demonstrate this using a 3d dynamic particle system that tracks the desired trajectories using a PID controller.

Craig Innes, Subramanian Ramamoorthy

Robotic control tasks are often first run in simulation for the purposes of verification, debugging and data augmentation. Many methods exist to specify what task a robot must complete, but few exist to specify what range of environments a user expects such tasks to be achieved in. ProbRobScene is a probabilistic specification language for describing robotic manipulation environments. Using the language, a user need only specify the relational constraints that must hold between objects in a scene. ProbRobScene will then automatically generate scenes which conform to this specification. By combining aspects of probabilistic programming languages and convex geometry, we provide a method for sampling this space of possible environments efficiently. We demonstrate the usefulness of our language by using it to debug a robotic controller in a tabletop robot manipulation environment.

Henry Pulver, Francisco Eiras, Ludovico Carozza et al.

Achieving a proper balance between planning quality, safety and efficiency is a major challenge for autonomous driving. Optimisation-based motion planners are capable of producing safe, smooth and comfortable plans, but often at the cost of runtime efficiency. On the other hand, naively deploying trajectories produced by efficient-to-run deep imitation learning approaches might risk compromising safety. In this paper, we present PILOT -- a planning framework that comprises an imitation neural network followed by an efficient optimiser that actively rectifies the network's plan, guaranteeing fulfilment of safety and comfort requirements. The objective of the efficient optimiser is the same as the objective of an expensive-to-run optimisation-based planning system that the neural network is trained offline to imitate. This efficient optimiser provides a key layer of online protection from learning failures or deficiency in out-of-distribution situations that might compromise safety or comfort. Using a state-of-the-art, runtime-intensive optimisation-based method as the expert, we demonstrate in simulated autonomous driving experiments in CARLA that PILOT achieves a seven-fold reduction in runtime when compared to the expert it imitates without sacrificing planning quality.

Paola Ardón, Maria E. Cabrera, Èric Pairet et al.

Reasoning about object handover configurations allows an assistive agent to estimate the appropriateness of handover for a receiver with different arm mobility capacities. While there are existing approaches for estimating the effectiveness of handovers, their findings are limited to users without arm mobility impairments and to specific objects. Therefore, current state-of-the-art approaches are unable to hand over novel objects to receivers with different arm mobility capacities. We propose a method that generalises handover behaviours to previously unseen objects, subject to the constraint of a user's arm mobility levels and the task context. We propose a heuristic-guided hierarchically optimised cost whose optimisation adapts object configurations for receivers with low arm mobility. This also ensures that the robot grasps consider the context of the user's upcoming task, i.e., the usage of the object. To understand preferences over handover configurations, we report on the findings of an online study, wherein we presented different handover methods, including ours, to $259$ users with different levels of arm mobility. We find that people's preferences over handover methods are correlated to their arm mobility capacities. We encapsulate these preferences in a statistical relational model (SRL) that is able to reason about the most suitable handover configuration given a receiver's arm mobility and upcoming task. Using our SRL model, we obtained an average handover accuracy of $90.8\%$ when generalising handovers to novel objects.

Simón C. Smith, Subramanian Ramamoorthy

We propose an architecture for training generative models of counterfactual conditionals of the form, 'can we modify event A to cause B instead of C?', motivated by applications in robot control. Using an 'adversarial training' paradigm, an image-based deep neural network model is trained to produce small and realistic modifications to an original image in order to cause user-defined effects. These modifications can be used in the design process of image-based robust control - to determine the ability of the controller to return to a working regime by modifications in the input space, rather than by adaptation. In contrast to conventional control design approaches, where robustness is quantified in terms of the ability to reject noise, we explore the space of counterfactuals that might cause a certain requirement to be violated, thus proposing an alternative model that might be more expressive in certain robotics applications. So, we propose the generation of counterfactuals as an approach to explanation of black-box models and the envisioning of potential movement paths in autonomous robotic control. Firstly, we demonstrate this approach in a set of classification tasks, using the well known MNIST and CelebFaces Attributes datasets. Then, addressing multi-dimensional regression, we demonstrate our approach in a reaching task with a physical robot, and in a navigation task with a robot in a digital twin simulation.

Todor Davchev, Kevin Sebastian Luck, Michael Burke et al.

Manipulation skills involving contact and friction are inherent to many robotics tasks. Using the class of motor primitives for peg-in-hole like insertions, we study how robots can learn such skills. Dynamic Movement Primitives (DMP) are a popular way of extracting such policies through behaviour cloning (BC) but can struggle in the context of insertion. Policy adaptation strategies such as residual learning can help improve the overall performance of policies in the context of contact-rich manipulation. However, it is not clear how to best do this with DMPs. As a result, we consider several possible ways for adapting a DMP formulation and propose ``residual Learning from Demonstration`` (rLfD), a framework that combines DMPs with Reinforcement Learning (RL) to learn a residual correction policy. Our evaluations suggest that applying residual learning directly in task space and operating on the full pose of the robot can significantly improve the overall performance of DMPs. We show that rLfD offers a gentle to the joints solution that improves the task success and generalisation of DMPs \rb{and enables transfer to different geometries and frictions through few-shot task adaptation}. The proposed framework is evaluated on a set of tasks. A simulated robot and a physical robot have to successfully insert pegs, gears and plugs into their respective sockets. Other material and videos accompanying this paper are provided at https://sites.google.com/view/rlfd/.

Michael Burke, Kartic Subr, Subramanian Ramamoorthy

Humans can easily reason about the sequence of high level actions needed to complete tasks, but it is particularly difficult to instil this ability in robots trained from relatively few examples. This work considers the task of neural action sequencing conditioned on a single reference visual state. This task is extremely challenging as it is not only subject to the significant combinatorial complexity that arises from large action sets, but also requires a model that can perform some form of symbol grounding, mapping high dimensional input data to actions, while reasoning about action relationships. This paper takes a permutation perspective and argues that action sequencing benefits from the ability to reason about both permutations and ordering concepts. Empirical analysis shows that neural models trained with latent permutations outperform standard neural architectures in constrained action sequencing tasks. Results also show that action sequencing using visual permutations is an effective mechanism to initialise and speed up traditional planning techniques and successfully scales to far greater action set sizes than models considered previously.

Simón C. Smith, Subramanian Ramamoorthy

Semi-supervised learning improves the performance of supervised machine learning by leveraging methods from unsupervised learning to extract information not explicitly available in the labels. Through the design of a system that enables a robot to learn inspection strategies from a human operator, we present a hybrid semi-supervised system capable of learning interpretable and verifiable models from demonstrations. The system induces a controller program by learning from immersive demonstrations using sequential importance sampling. These visual servo controllers are parametrised by proportional gains and are visually verifiable through observation of the position of the robot in the environment. Clustering and effective particle size filtering allows the system to discover goals in the state space. These goals are used to label the original demonstration for end-to-end learning of behavioural models. The behavioural models are used for autonomous model predictive control and scrutinised for explanations. We implement causal sensitivity analysis to identify salient objects and generate counterfactual conditional explanations. These features enable decision making interpretation and post hoc discovery of the causes of a failure. The proposed system expands on previous approaches to program synthesis by incorporating repellers in the attribution prior of the sampling process. We successfully learn the hybrid system from an inspection scenario where an unmanned ground vehicle has to inspect, in a specific order, different areas of the environment. The system induces an interpretable computer program of the demonstration that can be synthesised to produce novel inspection behaviours. Importantly, the robot successfully runs the synthesised program on an unseen configuration of the environment while presenting explanations of its autonomous behaviour.

Paola Ardón, Èric Pairet, Ronald P. A. Petrick et al.

Reasoning about object grasp affordances allows an autonomous agent to estimate the most suitable grasp to execute a task. While current approaches for estimating grasp affordances are effective, their prediction is driven by hypotheses on visual features rather than an indicator of a proposal's suitability for an affordance task. Consequently, these works cannot guarantee any level of performance when executing a task and, in fact, not even ensure successful task completion. In this work, we present a pipeline for SAGAT based on prior experiences. We visually detect a grasp affordance region to extract multiple grasp affordance configuration candidates. Using these candidates, we forward simulate the outcome of executing the affordance task to analyse the relation between task outcome and grasp candidates. The relations are ranked by performance success with a heuristic confidence function and used to build a library of affordance task experiences. The library is later queried to perform one-shot transfer estimation of the best grasp configuration on new objects. Experimental evaluation shows that our method exhibits a significant performance improvement up to 11.7% against current state-of-the-art methods on grasp affordance detection. Experiments on a PR2 robotic platform demonstrate our method's highly reliable deployability to deal with real-world task affordance problems.

Yordan Hristov, Subramanian Ramamoorthy

Robotic manipulation tasks, such as wiping with a soft sponge, require control from multiple rich sensory modalities. Human-robot interaction, aimed at teaching robots, is difficult in this setting as there is potential for mismatch between human and machine comprehension of the rich data streams. We treat the task of interpretable learning from demonstration as an optimisation problem over a probabilistic generative model. To account for the high-dimensionality of the data, a high-capacity neural network is chosen to represent the model. The latent variables in this model are explicitly aligned with high-level notions and concepts that are manifested in a set of demonstrations. We show that such alignment is best achieved through the use of labels from the end user, in an appropriately restricted vocabulary, in contrast to the conventional approach of the designer picking a prior over the latent variables. Our approach is evaluated in the context of two table-top robot manipulation tasks performed by a PR2 robot -- that of dabbing liquids with a sponge (forcefully pressing a sponge and moving it along a surface) and pouring between different containers. The robot provides visual information, arm joint positions and arm joint efforts. We have made videos of the tasks and data available - see supplementary materials at: https://sites.google.com/view/weak-label-lfd.

Daniel Angelov, Yordan Hristov, Subramanian Ramamoorthy

Learning models of user behaviour is an important problem that is broadly applicable across many application domains requiring human-robot interaction. In this work, we show that it is possible to learn generative models for distinct user behavioural types, extracted from human demonstrations, by enforcing clustering of preferred task solutions within the latent space. We use these models to differentiate between user types and to find cases with overlapping solutions. Moreover, we can alter an initially guessed solution to satisfy the preferences that constitute a particular user type by backpropagating through the learned differentiable models. An advantage of structuring generative models in this way is that we can extract causal relationships between symbols that might form part of the user's specification of the task, as manifested in the demonstrations. We further parameterize these specifications through constraint optimization in order to find a safety envelope under which motion planning can be performed. We show that the proposed method is capable of correctly distinguishing between three user types, who differ in degrees of cautiousness in their motion, while performing the task of moving objects with a kinesthetically driven robot in a tabletop environment. Our method successfully identifies the correct type, within the specified time, in 99% [97.8 - 99.8] of the cases, which outperforms an IRL baseline. We also show that our proposed method correctly changes a default trajectory to one satisfying a particular user specification even with unseen objects. The resulting trajectory is shown to be directly implementable on a PR2 humanoid robot completing the same task.

Paola Ardón, Èric Pairet, Katrin S. Lohan et al.

Affordances are key attributes of what must be perceived by an autonomous robotic agent in order to effectively interact with novel objects. Historically, the concept derives from the literature in psychology and cognitive science, where affordances are discussed in a way that makes it hard for the definition to be directly transferred to computational specifications useful for robots. This review article is focused specifically on robotics, so we discuss the related literature from this perspective. In this survey, we classify the literature and try to find common ground amongst different approaches with a view to application in robotics. We propose a categorisation based on the level of prior knowledge that is assumed to build the relationship among different affordance components that matter for a particular robotic task. We also identify areas for future improvement and discuss possible directions that are likely to be fruitful in terms of impact on robotics practice.

Stefano V. Albrecht, Cillian Brewitt, John Wilhelm et al.

We propose an integrated prediction and planning system for autonomous driving which uses rational inverse planning to recognise the goals of other vehicles. Goal recognition informs a Monte Carlo Tree Search (MCTS) algorithm to plan optimal maneuvers for the ego vehicle. Inverse planning and MCTS utilise a shared set of defined maneuvers and macro actions to construct plans which are explainable by means of rationality principles. Evaluation in simulations of urban driving scenarios demonstrate the system's ability to robustly recognise the goals of other vehicles, enabling our vehicle to exploit non-trivial opportunities to significantly reduce driving times. In each scenario, we extract intuitive explanations for the predictions which justify the system's decisions.

Francisco Eiras, Majd Hawasly, Stefano V. Albrecht et al.

Recent road trials have shown that guaranteeing the safety of driving decisions is essential for the wider adoption of autonomous vehicle technology. One promising direction is to pose safety requirements as planning constraints in nonlinear, non-convex optimization problems of motion synthesis. However, many implementations of this approach are limited by uncertain convergence and local optimality of the solutions achieved, affecting overall robustness. To improve upon these issues, we propose a novel two-stage optimization framework: in the first stage, we find a solution to a Mixed-Integer Linear Programming (MILP) formulation of the motion synthesis problem, the output of which initializes a second Nonlinear Programming (NLP) stage. The MILP stage enforces hard constraints of safety and road rule compliance generating a solution in the right subspace, while the NLP stage refines the solution within the safety bounds for feasibility and smoothness. We demonstrate the effectiveness of our framework via simulated experiments of complex urban driving scenarios, outperforming a state-of-the-art baseline in metrics of convergence, comfort and progress.

Michael Burke, Katie Lu, Daniel Angelov et al.

Informative path-planning is a well established approach to visual-servoing and active viewpoint selection in robotics, but typically assumes that a suitable cost function or goal state is known. This work considers the inverse problem, where the goal of the task is unknown, and a reward function needs to be inferred from exploratory example demonstrations provided by a demonstrator, for use in a downstream informative path-planning policy. Unfortunately, many existing reward inference strategies are unsuited to this class of problems, due to the exploratory nature of the demonstrations. In this paper, we propose an alternative approach to cope with the class of problems where these sub-optimal, exploratory demonstrations occur. We hypothesise that, in tasks which require discovery, successive states of any demonstration are progressively more likely to be associated with a higher reward, and use this hypothesis to generate time-based binary comparison outcomes and infer reward functions that support these ranks, under a probabilistic generative model. We formalise this \emph{probabilistic temporal ranking} approach and show that it improves upon existing approaches to perform reward inference for autonomous ultrasound scanning, a novel application of learning from demonstration in medical imaging while also being of value across a broad range of goal-oriented learning from demonstration tasks. \keywords{Visual servoing \and reward inference \and probabilistic temporal ranking

Craig Innes, Subramanian Ramamoorthy

Most current methods for learning from demonstrations assume that those demonstrations alone are sufficient to learn the underlying task. This is often untrue, especially if extra safety specifications exist which were not present in the original demonstrations. In this paper, we allow an expert to elaborate on their original demonstration with additional specification information using linear temporal logic (LTL). Our system converts LTL specifications into a differentiable loss. This loss is then used to learn a dynamic movement primitive that satisfies the underlying specification, while remaining close to the original demonstration. Further, by leveraging adversarial training, our system learns to robustly satisfy the given LTL specification on unseen inputs, not just those seen in training. We show that our method is expressive enough to work across a variety of common movement specification patterns such as obstacle avoidance, patrolling, keeping steady, and speed limitation. In addition, we show that our system can modify a base demonstration with complex specifications by incrementally composing multiple simpler specifications. We also implement our system on a PR-2 robot to show how a demonstrator can start with an initial (sub-optimal) demonstration, then interactively improve task success by including additional specifications enforced with our differentiable LTL loss.

Emmanuel Kahembwe, Subramanian Ramamoorthy

This work presents an analysis of the discriminators used in Generative Adversarial Networks (GANs) for Video. We show that unconstrained video discriminator architectures induce a loss surface with high curvature which make optimisation difficult. We also show that this curvature becomes more extreme as the maximal kernel dimension of video discriminators increases. With these observations in hand, we propose a family of efficient Lower-Dimensional Video Discriminators for GANs (LDVD GANs). The proposed family of discriminators improve the performance of video GAN models they are applied to and demonstrate good performance on complex and diverse datasets such as UCF-101. In particular, we show that they can double the performance of Temporal-GANs and provide for state-of-the-art performance on a single GPU.