eprintid: 10193135
rev_number: 8
eprint_status: archive
userid: 699
dir: disk0/10/19/31/35
datestamp: 2024-06-07 09:10:47
lastmod: 2024-06-07 09:10:47
status_changed: 2024-06-07 09:10:47
type: article
metadata_visibility: show
sword_depositor: 699
creators_name: Wang, Fanjin
creators_name: Harker, Anthony
creators_name: Edirisinghe, Mohan
creators_name: Parhizkar, Maryam
title: Tackling Data Scarcity Challenge through Active Learning
in Materials Processing with Electrospray
ispublished: inpress
divisions: UCL
divisions: B02
divisions: B04
divisions: C08
divisions: C05
divisions: C06
divisions: D10
divisions: F45
divisions: G08
divisions: F60
keywords: Active learning; machine learning; materials development; materials discovery; small data
note: Copyright © 2024 The Authors. Advanced Intelligent Systems published by Wiley-VCH GmbH
This is an open access article under the terms of the Creative Commons Attribution License, https://creativecommons.org/licenses/by/4.0/, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
abstract: Machine learning (ML) has been harnessed as a promising modelling tool for materials research. However, small data, or data scarcity, is a bottleneck when incorporating ML in studies involving experimentation. Current experiment planning methods show several disadvantages: one-factor-at-a-time (OFAT) experimentation became impractical due to limited laboratory resources; conventional design of experiments (DoE) failed to incorporate high-dimensional features in ML; Surrogate-based or Bayesian optimization (BO) shifted the goal to optimize material properties rather than guiding training data accumulation. The present research proposes leveraging active learning (AL) to strategically select critical data for experimentation. Two AL strategies, query-by-Committee (QBC) algorithm and Greedy method, are benchmarked against random query baseline on various materials datasets. AL is shown to efficiently reduce model prediction errors with minimal additional experiment data. Investigation of hyperparameters revealed benefits of applying AL at an early stage of experimental dataset construction. Moreover, AL is implemented and validated for an in-house materials development task - electrospray modelling. AL exploration as a paradigm is highlighted to guide experiment design for efficient data accumulation purposes, and its potential for further ML modelling. In doing so, the power of ML is expected to be fully unleashed to experimental researchers.
date: 2024-05-23
date_type: published
publisher: Wiley
official_url: http://dx.doi.org/10.1002/aisy.202300798
oa_status: green
full_text_type: pub
language: eng
primo: open
primo_central: open_green
verified: verified_manual
elements_id: 2279048
doi: 10.1002/aisy.202300798
lyricists_name: Wang, Fanjin
lyricists_name: Parhizkar, Maryam
lyricists_name: Harker, Anthony
lyricists_name: Edirisinghe, Mohan
lyricists_id: FWANA07
lyricists_id: MPARH34
lyricists_id: AHHAR69
lyricists_id: MJEDE48
actors_name: Basit, Abdul
actors_name: Harris, Jean
actors_id: ABASI56
actors_id: JAHAR68
actors_role: owner
actors_role: impersonator
full_text_status: public
publication: Advanced Intelligent Systems
article_number: 2300798
issn: 2640-4567
citation:        Wang, Fanjin;    Harker, Anthony;    Edirisinghe, Mohan;    Parhizkar, Maryam;      (2024)    Tackling Data Scarcity Challenge through Active Learning in Materials Processing with Electrospray.                   Advanced Intelligent Systems      , Article 2300798.  10.1002/aisy.202300798 <https://doi.org/10.1002/aisy.202300798>.    (In press).    Green open access   
 
document_url: https://discovery.ucl.ac.uk/id/eprint/10193135/1/Advanced%20Intelligent%20Systems%20-%202024%20-%20Wang.pdf