Simon Maskell

Darren Cook, Miri Zilka, Heidi DeSandre et al. · cambridge

Social media's growing popularity raises concerns around children's online safety. Interactions between minors and adults with predatory intentions is a particularly grave concern. Research into online sexual grooming has often relied on domain experts to manually annotate conversations, limiting both scale and scope. In this work, we test how well-automated methods can detect conversational behaviors and replace an expert human annotator. Informed by psychological theories of online grooming, we label $6772$ chat messages sent by child-sex offenders with one of eleven predatory behaviors. We train bag-of-words and natural language inference models to classify each behavior, and show that the best performing models classify behaviors in a manner that is consistent, but not on-par, with human annotation.

Efthyvoulos Drousiotis, Paul G. Spirakis, Simon Maskell

Markov Chain Monte Carlo (MCMC) is a well-established family of algorithms primarily used in Bayesian statistics to sample from a target distribution when direct sampling is challenging. Existing work on Bayesian decision trees uses MCMC. Unfortunately, this can be slow, especially when considering large volumes of data. It is hard to parallelise the accept-reject component of the MCMC. None-the-less, we propose two methods for exploiting parallelism in the MCMC: in the first, we replace the MCMC with another numerical Bayesian approach, the Sequential Monte Carlo (SMC) sampler, which has the appealing property that it is an inherently parallel algorithm; in the second, we consider data partitioning. Both methods use multi-core processing with a HighPerformance Computing (HPC) resource. We test the two methods in various study settings to determine which method is the most beneficial for each test case. Experiments show that data partitioning has limited utility in the settings we consider and that the use of the SMC sampler can improve run-time (compared to the sequential implementation) by up to a factor of 343.

Elias J Griffith, Chinmaya Mishra, Jason F. Ralph et al.

The development, benchmarking and validation of aerial Persistent Surveillance (PS) algorithms requires access to specialist Wide Area Aerial Surveillance (WAAS) datasets. Such datasets are difficult to obtain and are often extremely large both in spatial resolution and temporal duration. This paper outlines an approach to the simulation of complex urban environments and demonstrates the viability of using this approach for the generation of simulated sensor data, corresponding to the use of wide area imaging systems for surveillance and reconnaissance applications. This provides a cost-effective method to generate datasets for vehicle tracking algorithms and anomaly detection methods. The system fuses the Simulation of Urban Mobility (SUMO) traffic simulator with a MATLAB controller and an image generator to create scenes containing uninterrupted door-to-door journeys across large areas of the urban environment. This `pattern-of-life' approach provides three-dimensional visual information with natural movement and traffic flows. This can then be used to provide simulated sensor measurements (e.g. visual band and infrared video imagery) and automatic access to ground-truth data for the evaluation of multi-target tracking systems.

Chinmaya Mishra, Simon Maskell, Siu-Kui Au et al.

This paper presents an efficient method for estimating the probability of conflict between air traffic within a block of airspace. Autonomous Sense-and-Avoid is an essential safety feature to enable Unmanned Air Systems to operate alongside other (manned or unmanned) air traffic. The ability to estimate probability of conflict between traffic is an essential part of Sense-and-Avoid. Such probabilities are typically very low. Evaluating low probabilities using naive Direct Monte Carlo generates a significant computational load. This paper applies a technique called Subset Simulation. The small failure probabilities are computed as a product of larger conditional failure probabilities, reducing the computational load whilst improving the accuracy of the probability estimates. The reduction in the number of samples required can be one or more orders of magnitude. The utility of the approach is demonstrated by modeling a series of conflicting and potentially conflicting scenarios based on the standard Rules of the Air.

Conor Rosato, Joshua Murphy, Alessandro Varsi et al.

Sequential Monte Carlo Squared (SMC$^2$) is a Bayesian method which can infer the states and parameters of non-linear, non-Gaussian state-space models. The standard random-walk proposal in SMC$^2$ faces challenges, particularly with high-dimensional parameter spaces. This study outlines a novel approach by harnessing first-order gradients derived from a Common Random Numbers - Particle Filter (CRN-PF) using PyTorch. The resulting gradients can be leveraged within a Langevin proposal without accept/reject. Including Langevin dynamics within the proposal can result in a higher effective sample size and more accurate parameter estimates when compared with the random-walk. The resulting algorithm is parallelized on distributed memory using Message Passing Interface (MPI) and runs in $\mathcal{O}(\log_2N)$ time complexity. Utilizing 64 computational cores we obtain a 51x speed-up when compared to a single core. A GitHub link is given which provides access to the code.

Andrew Millard, Joshua Murphy, Peter Green et al.

Bayesian inference allows us to define a posterior distribution over the weights of a generic neural network (NN). Exact posteriors are usually intractable, in which case approximations can be employed. One such approximation - variational inference - is computationally efficient when using mini-batch stochastic gradient descent as subsets of the data are used for likelihood and gradient evaluations, though the approach relies on the selection of a variational distribution which sufficiently matches the form of the posterior. Particle-based methods such as Markov chain Monte Carlo and Sequential Monte Carlo (SMC) do not assume a parametric family for the posterior by typically require higher computational cost. These sampling methods typically use the full-batch of data for likelihood and gradient evaluations, which contributes to this computational expense. We explore several methods of gradually introducing more mini-batches of data (data annealing) into likelihood and gradient evaluations of an SMC sampler. We find that we can achieve up to $6\times$ faster training with minimal loss in accuracy on benchmark image classification problems using NNs.

Daniel James Sumler, Lee Devlin, Simon Maskell et al.

Understanding the confidence with which a machine learning model classifies an input datum is an important, and perhaps under-investigated, concept. In this paper, we propose a new calibration metric, the Entropic Calibration Difference (ECD). Based on existing research in the field of state estimation, specifically target tracking (TT), we show how ECD may be applied to binary classification machine learning models. We describe the relative importance of under- and over-confidence and how they are not conflated in the TT literature. Indeed, our metric distinguishes under- from over-confidence. We consider this important given that algorithms that are under-confident are likely to be 'safer' than algorithms that are over-confident, albeit at the expense of also being over-cautious and so statistically inefficient. We demonstrate how this new metric performs on real and simulated data and compare with other metrics for machine learning model probability calibration, including the Expected Calibration Error (ECE) and its signed counterpart, the Expected Signed Calibration Error (ESCE).

Andrew Millard, Joshua Murphy, Daniel Frisch et al.

Markov chain Monte Carlo (MCMC) methods are a powerful but computationally expensive way of performing non-parametric Bayesian inference. MCMC proposals which utilise gradients, such as Hamiltonian Monte Carlo (HMC), can better explore the parameter space of interest if the additional hyper-parameters are chosen well. The No-U-Turn Sampler (NUTS) is a variant of HMC which is extremely effective at selecting these hyper-parameters but is slow to run and is not suited to GPU architectures. An alternative to NUTS, Change in the Estimator of the Expected Square HMC (ChEES-HMC) was shown not only to run faster than NUTS on GPU but also sample from posteriors more efficiently. Sequential Monte Carlo (SMC) samplers are another sampling method which instead output weighted samples from the posterior. They are very amenable to parallelisation and therefore being run on GPUs while having additional flexibility in their choice of proposal over MCMC. We incorporate (ChEEs-HMC) as a proposal into SMC samplers and demonstrate competitive but faster performance than NUTS on a number of tasks.

Joshua Murphy, Conor Rosato, Andrew Millard et al.

When performing Bayesian inference using Sequential Monte Carlo (SMC) methods, two considerations arise: the accuracy of the posterior approximation and computational efficiency. To address computational demands, Sequential Monte Carlo Squared (SMC$^2$) is well-suited for high-performance computing (HPC) environments. The design of the proposal distribution within SMC$^2$ can improve accuracy and exploration of the posterior as poor proposals may lead to high variance in importance weights and particle degeneracy. The Metropolis-Adjusted Langevin Algorithm (MALA) uses gradient information so that particles preferentially explore regions of higher probability. In this paper, we extend this idea by incorporating second-order information, specifically the Hessian of the log-target. While second-order proposals have been explored previously in particle Markov Chain Monte Carlo (p-MCMC) methods, we are the first to introduce them within the SMC$^2$ framework. Second-order proposals not only use the gradient (first-order derivative), but also the curvature (second-order derivative) of the target distribution. Experimental results on synthetic models highlight the benefits of our approach in terms of step-size selection and posterior approximation accuracy when compared to other proposals.

Conor Rosato, Harvinder Lehal, Simon Maskell et al.

Bayesian inference with Markov Chain Monte Carlo (MCMC) is challenging when the likelihood function is irregular and expensive to compute. We explore several sampling algorithms that make use of subset evaluations to reduce computational overhead. We adapt the subset samplers for this setting where gradient information is not available or is unreliable. To achieve this, we introduce data-driven proxies in place of Taylor expansions and define a novel computation-cost aware adaptive controller. We undertake an extensive evaluation for a challenging disease modelling task and a configurable task with similar irregularity in the likelihood surface. We find our improved version of Hierarchical Importance with Nested Training Samples (HINTS), with adaptive proposals and a data-driven proxy, obtains the best sampling error in a fixed computational budget. We conclude that subset evaluations can provide cheap and naturally-tempered exploration, while a data-driven proxy can pre-screen proposals successfully in explored regions of the state space. These two elements combine through hierarchical delayed acceptance to achieve efficient, exact sampling.

Andrew Millard, Joshua Murphy, Simon Maskell et al.

Partial Bayesian neural networks (pBNNs) have been shown to perform competitively with fully Bayesian neural networks while only having a subset of the parameters be stochastic. Using sequential Monte Carlo (SMC) samplers as the inference method for pBNNs gives a non-parametric probabilistic estimation of the stochastic parameters, and has shown improved performance over parametric methods. In this paper we introduce a new SMC-based training method for pBNNs by utilising a guided proposal and incorporating gradient-based Markov kernels, which gives us better scalability on high dimensional problems. We show that our new method outperforms the state-of-the-art in terms of predictive performance and optimal loss. We also show that pBNNs scale well with larger batch sizes, resulting in significantly reduced training times and often better performance.

Marco Fontana, Ángel F. García-Fernández, Simon Maskell

This paper presents a Poisson multi-Bernoulli mixture (PMBM) filter for multi-target filtering based on sensor measurements that are sets of trajectories in the last two-time step window. The proposed filter, the trajectory measurement PMBM (TM-PMBM) filter, propagates a PMBM density on the set of target states. In prediction, the filter obtains the PMBM density on the set of trajectories over the last two time steps. This density is then updated with the set of trajectory measurements. After the update step, the PMBM posterior on the set of two-step trajectories is marginalised to obtain a PMBM density on the set of target states. The filter provides a closed-form solution for multi-target filtering based on sets of trajectory measurements, estimating the set of target states at the end of each time window. Additionally, the paper proposes computationally lighter alternatives to the TM-PMBM filter by deriving a Poisson multi-Bernoulli (PMB) density through Kullback-Leibler divergence minimisation in an augmented space with auxiliary variables. The performance of the proposed filters are evaluated in a simulation study.

Efthyvoulos Drousiotis, Alexander M. Phillips, Paul G. Spirakis et al.

Bayesian Decision Trees (DTs) are generally considered a more advanced and accurate model than a regular Decision Tree (DT) because they can handle complex and uncertain data. Existing work on Bayesian DTs uses Markov Chain Monte Carlo (MCMC) with an accept-reject mechanism and sample using naive proposals to proceed to the next iteration, which can be slow because of the burn-in time needed. We can reduce the burn-in period by proposing a more sophisticated way of sampling or by designing a different numerical Bayesian approach. In this paper, we propose a replacement of the MCMC with an inherently parallel algorithm, the Sequential Monte Carlo (SMC), and a more effective sampling strategy inspired by the Evolutionary Algorithms (EA). Experiments show that SMC combined with the EA can produce more accurate results compared to MCMC in 100 times fewer iterations.

Conor Rosato, Vincent Beraud, Paul Horridge et al.

It has been widely documented that the sampling and resampling steps in particle filters cannot be differentiated. The {\itshape reparameterisation trick} was introduced to allow the sampling step to be reformulated into a differentiable function. We extend the {\itshape reparameterisation trick} to include the stochastic input to resampling therefore limiting the discontinuities in the gradient calculation after this step. Knowing the gradients of the prior and likelihood allows us to run particle Markov Chain Monte Carlo (p-MCMC) and use the No-U-Turn Sampler (NUTS) as the proposal when estimating parameters. We compare the Metropolis-adjusted Langevin algorithm (MALA), Hamiltonian Monte Carlo with different number of steps and NUTS. We consider two state-space models and show that NUTS improves the mixing of the Markov chain and can produce more accurate results in less computational time.

Youcheng Sun, Yifan Zhou, Simon Maskell et al.

This paper studies the reliability of a real-world learning-enabled system, which conducts dynamic vehicle tracking based on a high-resolution wide-area motion imagery input. The system consists of multiple neural network components -- to process the imagery inputs -- and multiple symbolic (Kalman filter) components -- to analyse the processed information for vehicle tracking. It is known that neural networks suffer from adversarial examples, which make them lack robustness. However, it is unclear if and how the adversarial examples over learning components can affect the overall system-level reliability. By integrating a coverage-guided neural network testing tool, DeepConcolic, with the vehicle tracking system, we found that (1) the overall system can be resilient to some adversarial examples thanks to the existence of other components, and (2) the overall system presents an extra level of uncertainty which cannot be determined by analysing the deep learning components only. This research suggests the need for novel verification and validation methods for learning-enabled systems.

Yifan Zhou, Simon Maskell

This paper proposes an approach to detect moving objects in Wide Area Motion Imagery (WAMI), in which the objects are both small and well separated. Identifying the objects only using foreground appearance is difficult since a $100-$pixel vehicle is hard to distinguish from objects comprising the background. Our approach is based on background subtraction as an efficient and unsupervised method that is able to output the shape of objects. In order to reliably detect low contrast and small objects, we configure the background subtraction to extract foreground regions that might be objects of interest. While this dramatically increases the number of false alarms, a Convolutional Neural Network (CNN) considering both spatial and temporal information is then trained to reject the false alarms. In areas with heavy traffic, the background subtraction yields merged detections. To reduce the complexity of multi-target tracker needed, we train another CNN to predict the positions of multiple moving objects in an area. Our approach shows competitive detection performance on smaller objects relative to the state-of-the-art. We adopt a GM-PHD filter to associate detections over time and analyse the resulting performance.

Amir Soleimani, Nasser M. Nasrabadi, Elias Griffith et al.

This paper investigates the problem of aerial vehicle recognition using a text-guided deep convolutional neural network classifier. The network receives an aerial image and a desired class, and makes a yes or no output by matching the image and the textual description of the desired class. We train and test our model on a synthetic aerial dataset and our desired classes consist of the combination of the class types and colors of the vehicles. This strategy helps when considering more classes in testing than in training.