RESEARCH NEWS
Digital twins for industry |
to develop an easier and more cost-effective approach that enables multinationals to harness the power of digital twins . |
|
Quaisr , a start-up created by researchers from Imperial and the Alan Turing Institute in London , has raised US $ 3.1 million to further develop its work on digital twins . The money will be used to expand the team , with a focus on software engineering and business development .
“ Our top priority is to determine productmarket fit , and to do so we need people to be focused on product development while we obtain valuable feedback from our customers ,” said Dr . Assen Batchvarov , the company ’ s product manager .
Quaisr was set up in 2020 by Imperial and Turing researchers Professor Richard Craster , Imperial ’ s Dean of Natural Sciences ; Professor Omar Matar , Head of Imperial ’ s Department of Chemical Engineering ; Dr . Indranil Pan and Dr . Lachlan Mason . The team is expected to grow to a dozen full-time members over the next couple of quarters .
The pre-seed funding round was led by Crane Venture Partners , with the participation
|
of Acequia Capital , Hybris Founder Carsten Thoma , Encord Founder Eric Landau and additional strategic angel investors .
A digital twin is a virtual representation of an object or system that can be updated from real-time data . This can then be used to monitor and predict the performance of the object or the evolution of the system under different conditions , to help decision-making or inform further research and development . Apart from applicability in operations , using a digital twin can significantly shorten the discovery cycles for new materials , for example , or help to optimise the performance of physical systems , such as manufacturing lines or chemical production facilities , in realtime .
However , connecting , scaling and democratising the building blocks that represent digital replicas of assets and processes is time-consuming and resourceintensive . Frustrated by limitations of the tools available , the Quaisr founders set out
|
The result is a platform “ that can leverage cloud technologies or on-premises infrastructure ( depending upon client requirements ) so that domain experts can focus on improving and connecting their models and simulations instead of navigating IT infrastructure , security , authentication and writing non-standardised ‘ glue ’ code .”
“ The Quaisr platform will empower heavy industries to build reliable digital twins of their assets and processes , making their operations efficient , sustainable and more reliable ,” said Dr . Batchvarov .
Quaisr is currently taking part in Aerospace Xelerated , an accelerator run by Boeing , and recently announced a materials discovery partnership with security and defence company , QinetiQ . n www . imperial . ac . uk www . quaisr . com
|
Accurately simulating complex systems |
Researchers often use simulations when designing new algorithms , since testing ideas in the real world can be both costly and risky . But since it ’ s impossible to capture every detail of a complex system in a simulation , they typically collect a small amount of real data that they replay while simulating the components they want to study .
Known as trace-driven simulation ( the small pieces of real data are called traces ), this method sometimes results in biased outcomes . This means researchers might unknowingly choose an algorithm that is not the best one they evaluated , and which will perform worse on real data than the simulation predicted that it should .
MIT ( Massachussetts ’ Institute of Technology ) researchers have developed a new method that eliminates this source of bias in trace-driven simulation . By enabling unbiased trace-driven simulations , the new technique could help researchers design better algorithms for a variety of applications , including improving video quality on the internet and increasing the performance of data processing systems .
The researchers ’ machine-learning algorithm draws on the principles of causality to learn how the data traces were affected by the behaviour of the system . In this way , they can replay the correct , unbiased version of the trace during the simulation .
When compared to a previously developed trace-driven simulator , the researchers ’
Above : A neural network with lines connecting square nodes in three rows . The squares have different scenes of bold shapes in grey , pink and green . On the right row , there are two large square nodes with more complex arrangements of shapes . Image credit : Jose-Luis Olivares / MIT .)
simulation method correctly predicted which newly designed algorithm would be best for video streaming — meaning the one that led to less rebuffering and higher visual quality . Existing simulators that do not account for bias would have pointed researchers to a worseperforming algorithm .
“ Data are not the only thing that matter . The story behind how the data are generated and collected is also important . If you want to answer a counterfactual question , you need to know the underlying data generation story so you only intervene on those things that you really want to simulate ,” said Arash Nasr-Esfahany , an electrical engineering and computer science ( EECS ) graduate student and co-lead author of a paper on this new technique .
He is joined on the paper by co-lead authors and fellow EECS graduate students Abdullah Alomar and Pouya Hamadanian ; recent graduate student Anish Agarwal PhD ’ 21 ; and senior authors Mohammad Alizadeh , an associate professor of electrical engineering and computer science , and Devavrat Shah , the Andrew and Erna Viterbi Professor in EECS and a member of the Institute for Data , Systems and Society and of the Laboratory for Information and Decision Systems .
The new tool / algorithm they developed , dubbed CausalSim , can learn the underlying characteristics of a system using only the trace data . CausalSim takes trace data that were collected through a randomised control trial and estimates the underlying functions that produced those data . The model tells the researchers , under the exact same underlying conditions that a user experienced , how a new algorithm would change the outcome .
Using a typical trace-driven simulator , bias might lead a researcher to select a worse-performing algorithm , even though the simulation indicates it should be better . CausalSim helps researchers select the best algorithm that was tested .
During a ten-month experiment , CausalSim consistently improved simulation accuracy “ resulting in algorithms that made about half as many errors as those designed using baseline methods .” n
https
:// web . mit . edu /
ISMR June 2023 | ismr . net | 17