We present Structured World Modeling for Policy Optimization (SWMPO), a framework for unsupervised learning of neurosymbolic Finite State Machines (FSM) that capture environmental structure for policy optimization. Traditional unsupervised world modeling methods rely on unstructured representations, such as neural networks, that do not explicitly represent high-level patterns within the system (e.g., patterns in the dynamics of regions such as \emph{water} and \emph{land}). Instead, SWMPO models the environment as a finite state machine (FSM), where each state corresponds to a specific region with distinct dynamics. This structured representation can then be leveraged for tasks like policy optimization. Previous works that synthesize FSMs for this purpose have been limited to discrete spaces, not continuous spaces. Instead, our proposed FSM synthesis algorithm operates in an unsupervised manner, leveraging low-level features from unprocessed, non-visual data, making it adaptable across various domains. The synthesized FSM models are expressive enough to be used in a model-based Reinforcement Learning scheme that leverages offline data to efficiently synthesize environment-specific world models. We demonstrate the advantages of SWMPO by benchmarking its environment modeling capabilities in simulated environments.

We study the problem of sequential decision-making in environments governed by evolving causal mechanisms. Prior work on structural causal bandits--formulations that integrate causal graphs into multi-armed bandit problems to guide intervention selection--has shown that leveraging the causal structure can reduce unnecessary interventions by identifying possibly-optimal minimal intervention sets (POMISs). However, such formulations fall short in dynamic settings where reward distributions may vary over time, as their static, hence myopic, nature focuses on immediate rewards and overlooks the long-term effects of interventions. In this work, we propose a non-stationary structural causal bandit framework that leverages temporal structural causal models to capture evolving dynamics over time. We characterize how interventions propagate over time by developing graphical tools and assumptions, which form the basis for identifying non-myopic intervention strategies. Within this framework, we devise POMIS$^+$, which captures the existence of variables that contribute to maximizing both immediate and long-term rewards. Our framework provides a principled way to reason about temporally-aware interventions by explicitly modeling information propagation across time. Empirical results validate the effectiveness of our approach, demonstrating improved performance over myopic baselines.

Abstract To mitigate the current COVID-19 pandemic, policy makers at the Greater London Authority, the regional governance body of London, UK, are reliant upon prompt, accurate and actionable estimations of lockdown and social distancing policy adherence. Transport for London, the local transportation department, reports they implemented over 700 interventions such as greater signage and expansion of pedestrian zoning at the height of the pandemic’s first wave with our platform providing key data for those decisions. Large well-defined heterogeneous compositions of pedestrian footfall and physical proximity are difficult to acquire, yet necessary to monitor city-wide activity (busyness) and consequently discern actionable policy decisions. To meet this challenge, we leverage our existing large-scale data processing urban air quality machine learning infrastructure to process over 900 camera feeds in near real-time to generate new estimates of social distancing adherence, group detection and camera stability. In this work, we describe our development and deployment of a computer vision and machine learning pipeline. It provides near immediate sampling and contextualization of activity and physical distancing on the streets of London via live traffic camera feeds. We introduce a platform for inspecting, calibrating and improving upon existing methods, describe the active deployment on real-time feeds and provide analysis over an 18 month period.

We study the problem of globally optimizing the causal effect on a target variable of an unknown causal graph in which interventions can be performed. This problem arises in many areas of science including biology, operations research and healthcare. We propose Causal Entropy Optimization (CEO), a framework that generalizes Causal Bayesian Optimization (CBO) to account for all sources of uncertainty, including the one arising from the causal graph structure. CEO incorporates the causal structure uncertainty both in the surrogate models for the causal effects and in the mechanism used to select interventions via an information-theoretic acquisition function. The resulting algorithm automatically trades-off structure learning and causal effect optimization, while naturally accounting for observation noise. For various synthetic and real-world structural causal models, CEO achieves faster convergence to the global optimum compared with CBO while also learning the graph. Furthermore, our joint approach to structure learning and causal optimization improves upon sequential, structure-learning-first approaches.

We study the problem of performing a sequence of optimal interventions in a dynamic causal system where both the target variable of interest, and the inputs, evolve over time. This problem arises in a variety of domains including healthcare, operational research and policy design. Our approach, which we call Dynamic Causal Bayesian Optimisation (DCBO), brings together ideas from decision making, causal inference and Gaussian process (GP) emulation. DCBO is useful in scenarios where the causal effects are changing over time. Indeed, at every time step, DCBO identifies a local optimal intervention by integrating both observational and past interventional data collected from the system. We give theoretical results detailing how one can transfer interventional information across time steps and define a dynamic causal GP model which can be used to find optimal interventions in practice. Finally, we demonstrate how DCBO identifies optimal interventions faster than competing approaches in multiple settings and applications.

There is urgent need for non-intrusive tests that can detect early signs of Parkinson's disease (PD), a debilitating neurodegenerative disorder that affects motor control. Recent promising research has focused on disease markers evident in the fine-motor behaviour of typing. Most work to date has focused solely on the timing of keypresses without reference to the linguistic content. In this paper we argue that the identity of the key combinations being produced should impact how they are handled by people with PD, and provide evidence that natural language processing methods can thus be of help in identifying signs of disease. We test the performance of a bi-directional LSTM with convolutional features in distinguishing people with PD from age-matched controls typing in English and Span-ish, both in clinics and online. 1

With ever more data becoming available, there has been a recent drive to develop modelling tools for heterogeneous data sets such as electronic healthcare records. Therein appears both Bayesian non-parametric latent feature models, as well as methods for automatically determining the statistical data type (e.g. ordinal or categorical) of the attributes present in the data. We present a model which combines both of these attractive features in an end-to-end framework. By jointly learning the model complexity and statistical type of the data, we demonstrate that redundant information can be discarded while higher accuracy statistical type estimates are produced on real experimental data.

In this paper we combine Gaussian process regression and impedance control, to illicit robust, anthropomorphic, adaptive control of a powered ankle prosthesis. We learn the non-linear manifolds which guide how locomotion variables temporally evolve, and regress that surface over a velocity range to create a manifold. The joint set of manifolds, as well as the temporal evolution of the gait-cycle duration is what we term a locomotion envelope. Current powered prostheses have problems adapting across speeds. It is likely that humans rely upon a control strategy which is adaptable, can become more robust and accurate with more data and provides a nonparametric approach which allows the strategy to grow with the number of observations. We demonstrate such a strategy in this study and successfully simulate locomotion well beyond our training data. The method we propose is based on common physical features observed in numerous human subjects walking at different speeds. Based on the derived locomotion envelopes we show that ankle power increases monotonically with speed among all subjects. We demonstrate our methods in simulation and human experiments, on a powered ankle foot prosthesis to demonstrate the effectiveness of the method.

Human locomotion and activity recognition systems form a critical part in a robot's ability to safely and effectively operate in a environment populated with human end users. Previous work in this area relies upon strong assumptions about the labels in the training data; e.g. that are noise-free and that they exist at all. Our approach does not predefine the relevant behaviours or their number, as both are learned directly from observations, similar to real-world human-robot interactions, where labels are neither available. Instead we introduce models that make no assumptions about the state space, by presenting a fully unsupervised nonparametric Bayesian recognition approach, in which we leverage recent advances in state space modelling with automatic inference using probabilistic programming. We demonstrate the utility of full model optimisation using Bayesian optimisation and validate our approach on several challenging problems, using different feature modalities.