Hierarchical generative modelling for autonomous robots

Humans can produce complex movements when interacting with their surroundings. This relies on the planning of various movements and subsequent execution. In this paper, we investigated this fundamental aspect of motor control in the setting of autonomous robotic operations. We consider hierarchical generative modelling—for autonomous task completion—that mimics the deep temporal architecture of human motor control. Here, temporal depth refers to the nested time scales at which successive levels of a forward or generative model unfold: for example, the apprehension and delivery of an object requires both a global plan that contextualises the fast coordination of multiple local limb movements. This separation of temporal scales can also be motivated from a robotics and control perspective. Specifically, to ensure versatile sensorimotor control, it is necessary to hierarchically structure high-level planning and low-level motor control of individual limbs. We use numerical experiments to establish the efficacy of this formulation and demonstrate how a humanoid robot can autonomously solve a complex task requiring locomotion, manipulation, and grasping, using a hierarchical generative model. In particular, the humanoid robot can retrieve and deliver a box, open and walk through a door to reach the final destination. Our approach, and experiments, illustrate the effectiveness of using human-inspired motor control algorithms, which provide a scalable hierarchical architecture for autonomous performance of complex goal-directed tasks.

A data-centric approach to generative modelling for 3D-printed steel

The emergence of additive manufacture (AM) for metallic material enables components of near arbitrary complexity to be produced. This has potential to disrupt traditional engineering approaches. However, metallic AM components exhibit greater levels of variation in their geometric and mechanical properties compared to standard components, which is not yet well understood. This uncertainty poses a fundamental barrier to potential users of the material, since extensive post-manufacture testing is currently required to ensure safety standards are met. Taking an interdisciplinary approach that combines probabilistic mechanics and uncertainty quantification, we demonstrate that intrinsic variation in AM steel can be well described by a generative statistical model that enables the quality of a design to be predicted before manufacture. Specifically, the geometric variation in the material can be described by an anisotropic spatial random field with oscillatory covariance structure, and the mechanical behaviour by a stochastic anisotropic elasto-plastic material model. The fitted generative model is validated on a held-out experimental dataset and our results underscore the need to combine both statistical and physics-based modelling in the characterization of new AM steel products.

F-Divergences and Cost Function Locality in Generative Modelling with Quantum Circuits

Generative modelling is an important unsupervised task in machine learning. In this work, we study a hybrid quantum-classical approach to this task, based on the use of a quantum circuit born machine. In particular, we consider training a quantum circuit born machine using f-divergences. We first discuss the adversarial framework for generative modelling, which enables the estimation of any f-divergence in the near term. Based on this capability, we introduce two heuristics which demonstrably improve the training of the born machine. The first is based on f-divergence switching during training. The second introduces locality to the divergence, a strategy which has proved important in similar applications in terms of mitigating barren plateaus. Finally, we discuss the long-term implications of quantum devices for computing f-divergences, including algorithms which provide quadratic speedups to their estimation. In particular, we generalise existing algorithms for estimating the Kullback–Leibler divergence and the total variation distance to obtain a fault-tolerant quantum algorithm for estimating another f-divergence, namely, the Pearson divergence.

Beyond Simple Images: Human Knowledge-Guided GANs for Clinical Data Generation

While Generative Adversarial Networks (GANs) have accelerated the use of generative modelling within the machine learning community, most of the adaptations of GANs are restricted to images. The use of GANs to generate clinical data has been rare due to the inability of GANs to faithfully capture the intrinsic relationships between features given a small amount of observational data. We hypothesize and verify that this challenge can be mitigated by incorporating rich domain knowledge in the form of expert advice in the generative process. Specifically, we propose human-allied GANs that uses correlation advice from humans to create synthetic clinical data. We construct a system that takes a symbolic representation of the expert advice and converts it into constraints on correlation of the features during the generative process. Our empirical evaluation demonstrates (a) the superiority of our approach over other GAN models, (b) the importance of incorporating advice over instance noise and (c) an initial framework for incorporation of privacy in our model while capturing the relationships between features.

EXTENDING 3D QUALITY MODELLING FOR EARTHQUAKE-DAMAGED STONE MASONRY WALL: COMBINED DIGITAL MODELS FOR BUILDING ARCHAEOLOGY

The paper proposes an approach for defining a generative modelling process of complex objects and their sharing. The case study is the Stronghold of Arquata del Tronto, a monument of extraordinary historical, cultural and landscape value, damaged by the earthquake in 2016. The first step has been data acquisition on a geometrical level, through laser scanner and UAV photogrammetry, and on a historical level, through archival research to understand construction phases and transformations. The Stronghold was probably built between the 11 and the 12th century on a hill to control the territory. It underwent several transformations and neglection phases over the centuries. The second phase has been the generative modelling following the scan-to-BIM approach. The three-dimensional model is intended to support the design phases, from preliminary analysis to the construction site.For this reason, the Stronghold has been modelled with different Grade of Generation (GOG). The study of the eastern curtain wall, where the signs suffered by the structure due to the earthquake are most evident, was deepened through a Building Archaeology preliminary analysis. The third phase aimed at orienting the HBIM towards three digital information-sharing solutions such as Common Data Environment (CDE), and Virtual Reality (VR) to enhance the cultural and historical values, supporting the reopening of the Stronghold as a venue for conferences and exhibitions.