Showing 1–5 of 5 results for author: Hocking, A

Bridging the Gap: Applying Assurance Arguments to MIL-HDBK-516C Certification of a Neural Network Control System with ASIF Run Time Assurance Architecture

Authors: Jonathan Rowanhill, Ashlie B. Hocking, Aditya Zutshi, Kerianne L. Hobbs

Abstract: Recent advances in artificial intelligence and machine learning may soon yield paradigm-shifting benefits for aerospace systems. However, complexity and possible continued on-line learning makes neural network control systems (NNCS) difficult or impossible to certify under the United States Military Airworthiness Certification Criteria defined in MIL-HDBK-516C. Run time assurance (RTA) is a contro… ▽ More Recent advances in artificial intelligence and machine learning may soon yield paradigm-shifting benefits for aerospace systems. However, complexity and possible continued on-line learning makes neural network control systems (NNCS) difficult or impossible to certify under the United States Military Airworthiness Certification Criteria defined in MIL-HDBK-516C. Run time assurance (RTA) is a control system architecture designed to maintain safety properties regardless of whether a primary control system is fully verifiable. This work examines how to satisfy compliance with MIL-HDBK-516C while using active set invariance filtering (ASIF), an advanced form of RTA not envisaged by the 516c committee. ASIF filters the commands from a primary controller, passing on safe commands while optimally modifying unsafe commands to ensure safety with minimal deviation from the desired control action. This work examines leveraging the core theory behind ASIF as assurance argument explaining novel satisfaction of 516C compliance criteria. The result demonstrates how to support compliance of novel technologies with 516C as well as elaborate how such standards might be updated for emerging technologies. △ Less

Submitted 27 March, 2023; originally announced March 2023.

Report number: AFRL-2022-5579 ACM Class: C.4; D.2.4; D.2.11; B.8.m

Systems Theoretic Process Analysis of a Run Time Assured Neural Network Control System

Authors: Kerianne L. Hobbs, Benjamin K. Heiner, Lillian Busse, Kyle Dunlap, Jonathan Rowanhill, Ashlie B. Hocking, Aditya Zutshi

Abstract: This research considers the problem of identifying safety constraints and develo** Run Time Assurance (RTA) for Deep Reinforcement Learning (RL) Tactical Autopilots that use neural network control systems (NNCS). This research studies a specific use case of an NNCS performing autonomous formation flight while an RTA system provides collision avoidance and geofence assurances. First, Systems Theo… ▽ More This research considers the problem of identifying safety constraints and develo** Run Time Assurance (RTA) for Deep Reinforcement Learning (RL) Tactical Autopilots that use neural network control systems (NNCS). This research studies a specific use case of an NNCS performing autonomous formation flight while an RTA system provides collision avoidance and geofence assurances. First, Systems Theoretic Accident Models and Processes (STAMP) is applied to identify accidents, hazards, and safety constraints as well as define a functional control system block diagram of the ground station, manned flight lead, and surrogate unmanned wingman. Then, Systems Theoretic Process Analysis (STPA) is applied to the interactions of the the ground station, manned flight lead, surrogate unmanned wingman, and internal elements of the wingman aircraft to identify unsafe control actions, scenarios leading to each, and safety requirements to mitigate risks. This research is the first application of STAMP and STPA to an NNCS bounded by RTA. △ Less

Submitted 9 November, 2022; v1 submitted 1 September, 2022; originally announced September 2022.

Galaxy morphological classification in deep-wide surveys via unsupervised machine learning

Authors: Garreth Martin, Sugata Kaviraj, Alex Hocking, Shaun C. Read, James E. Geach

Abstract: Galaxy morphology is a fundamental quantity, that is essential not only for the full spectrum of galaxy-evolution studies, but also for a plethora of science in observational cosmology. While a rich literature exists on morphological-classification techniques, the unprecedented data volumes, coupled, in some cases, with the short cadences of forthcoming 'Big-Data' surveys (e.g. from the LSST), pre… ▽ More Galaxy morphology is a fundamental quantity, that is essential not only for the full spectrum of galaxy-evolution studies, but also for a plethora of science in observational cosmology. While a rich literature exists on morphological-classification techniques, the unprecedented data volumes, coupled, in some cases, with the short cadences of forthcoming 'Big-Data' surveys (e.g. from the LSST), present novel challenges for this field. Large data volumes make such datasets intractable for visual inspection (even via massively-distributed platforms like Galaxy Zoo), while short cadences make it difficult to employ techniques like supervised machine-learning, since it may be impractical to repeatedly produce training sets on short timescales. Unsupervised machine learning, which does not require training sets, is ideally suited to the morphological analysis of new and forthcoming surveys. Here, we employ an algorithm that performs clustering of graph representations, in order to group image patches with similar visual properties and objects constructed from those patches, like galaxies. We implement the algorithm on the Hyper-Suprime-Cam Subaru-Strategic-Program Ultra-Deep survey, to autonomously reduce the galaxy population to a small number (160) of 'morphological clusters', populated by galaxies with similar morphologies, which are then benchmarked using visual inspection. The morphological classifications (which we release publicly) exhibit a high level of purity, and reproduce known trends in key galaxy properties as a function of morphological type at z<1 (e.g. stellar-mass functions, rest-frame colours and the position of galaxies on the star-formation main sequence). Our study demonstrates the power of unsupervised machine learning in performing accurate morphological analysis, which will become indispensable in this new era of deep-wide surveys. △ Less

Submitted 23 October, 2019; v1 submitted 23 September, 2019; originally announced September 2019.

Comments: Accepted for publication in MNRAS. 19 pages, 12 figures

An automatic taxonomy of galaxy morphology using unsupervised machine learning

Authors: Alex Hocking, James E. Geach, Yi Sun, Neil Davey

Abstract: We present an unsupervised machine learning technique that automatically segments and labels galaxies in astronomical imaging surveys using only pixel data. Distinct from previous unsupervised machine learning approaches used in astronomy we use no pre-selection or pre-filtering of target galaxy type to identify galaxies that are similar. We demonstrate the technique on the HST Frontier Fields. By… ▽ More We present an unsupervised machine learning technique that automatically segments and labels galaxies in astronomical imaging surveys using only pixel data. Distinct from previous unsupervised machine learning approaches used in astronomy we use no pre-selection or pre-filtering of target galaxy type to identify galaxies that are similar. We demonstrate the technique on the HST Frontier Fields. By training the algorithm using galaxies from one field (Abell 2744) and applying the result to another (MACS0416.1-2403), we show how the algorithm can cleanly separate early and late type galaxies without any form of pre-directed training for what an 'early' or 'late' type galaxy is. We then apply the technique to the HST CANDELS fields, creating a catalogue of approximately 60,000 classifications. We show how the automatic classification groups galaxies of similar morphological (and photometric) type, and make the classifications public via a catalogue, a visual catalogue and galaxy similarity search. We compare the CANDELS machine-based classifications to human-based classifications from the Galaxy Zoo: CANDELS project. Although there is not a direct map** between Galaxy Zoo and our hierarchical labelling, we demonstrate a good level of concordance between human and machine classifications. Finally, we show how the technique can be used to identify rarer objects and present new lensed galaxy candidates from the CANDELS imaging. △ Less

Submitted 18 September, 2017; originally announced September 2017.

Comments: 21 pages, 20 figures, accepted for publication in MNRAS. Code available at https://github.com/alexhock/galaxymorphology. This article is a substantially revised and expanded version of arXiv:1507.01589

Teaching a machine to see: unsupervised image segmentation and categorisation using growing neural gas and hierarchical clustering

Authors: Alex Hocking, James E. Geach, Neil Davey, Yi Sun

Abstract: We present a novel unsupervised learning approach to automatically segment and label images in astronomical surveys. Automation of this procedure will be essential as next-generation surveys enter the petabyte scale: data volumes will exceed the capability of even large crowd-sourced analyses. We demonstrate how a growing neural gas (GNG) can be used to encode the feature space of imaging data. Wh… ▽ More We present a novel unsupervised learning approach to automatically segment and label images in astronomical surveys. Automation of this procedure will be essential as next-generation surveys enter the petabyte scale: data volumes will exceed the capability of even large crowd-sourced analyses. We demonstrate how a growing neural gas (GNG) can be used to encode the feature space of imaging data. When coupled with a technique called hierarchical clustering, imaging data can be automatically segmented and labelled by organising nodes in the GNG. The key distinction of unsupervised learning is that these labels need not be known prior to training, rather they are determined by the algorithm itself. Importantly, after training a network can be be presented with images it has never 'seen' before and provide consistent categorisation of features. As a proof-of-concept we demonstrate application on data from the Hubble Space Telescope Frontier Fields: images of clusters of galaxies containing a mixture of galaxy types that would easily be recognised and classified by a human inspector. By training the algorithm using one field (Abell 2744) and applying the result to another (MACS0416.1-2403), we show how the algorithm can cleanly separate image features that a human would associate with early and late type galaxies. We suggest that the algorithm has potential as a tool in the automatic analysis and data mining of next-generation imaging and spectral surveys, and could also find application beyond astronomy. △ Less

Submitted 6 July, 2015; originally announced July 2015.

Comments: 5 pages, submitted to MNRAS