DesignofExperimentsforModelDiscrimination using Gaussian...
1
Design of Experiments for Model Discrimination using Gaussian Process Surrogate Models Simon Olofsson , Marc Peter Deisenroth and Ruth Misener Dept of Computing, Imperial College London, SW7 2AZ, UK. 1. Model Discrimination Assume we are studying a complex physical system (e.g. a human organ interacting with a new pharmaceutical). We may have rival models (i.e. hypotheses) for some underlying system mechanism. We perform model discrimination by discarding models whose predictions do not agree with experimental results. Time t f (t) Example of rival models. Experiment Model 1 Model 2 Common in industry to lack sufficient experimental results to perform model discrimination. This issue e.g. slows down devel- opment of new pharmaceuticals. ⇒ Need additional experiments! Challenge: Design the most informative experiments to help with model discrimination. Classical approach vs. Data-driven approach How can we find the most informative next experiment, taking model uncertainty into account? 2.1 Classical Approach Find optimal next experiment by exploiting the models’ mathematical structure. Time t f (t) / Information Model 1 (with uncertainty) Model 2 (with uncertainty) Predicted information content + Computationally cheap. – Limits our choice of model structure. 2.2 Data-Driven Approach Exploit power of computers to find the optimal next experiment through brute force. Time t f (t) / Information Model 1 sample Model 2 sample Predicted information content + Flexible with regards to model structure. – Computationally expensive. Using machine learning, we hybridise the two approaches 3. Novel Approach We introduce data-driven surrogate models, whose mathematical structure can be exploited. Time t f (t) / Information Model 1 (with uncertainty) Model 2 (with uncertainty) Predicted information content + Flexible with regards to models structure. + Computationally cheap. 4. Conclusion Difficulties with model discrimination often slows down e.g. development of new pharmaceuticals. We hybridise two existing approaches, leveraging advantages of both, by use of machine learning. The novel approach is implemented and available as free open-source software. References S. Olofsson et al., 2019. arXiv e-print: 1810.02561. S. Olofsson et al., 2018. Proc Mach Learn Res 80. S. Olofsson et al., 2018. Comput Aided Chem Eng 44. Acknowledgement. This work has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement no. 675251, and an EPSRC Research Fellowship to R.M. (EP/P016871/1).
Transcript of DesignofExperimentsforModelDiscrimination using Gaussian...
Design of Experiments for Model Discriminationusing Gaussian Process Surrogate ModelsSimon Olofsson, Marc Peter Deisenroth and Ruth MisenerDept of Computing, Imperial College London, SW7 2AZ, UK.1. Model DiscriminationAssume we are studying a complexphysical system (e.g. a human organinteracting with a new pharmaceutical).
We may have rival models (i.e.hypotheses) for some underlyingsystem mechanism. We perform modeldiscrimination by discarding modelswhose predictions do not agree withexperimental results.
Time t
f(t)
Example of rival models.
ExperimentModel 1Model 2
Common in industry to lacksufficient experimental results toperform model discrimination.This issue e.g. slows down devel-opment of new pharmaceuticals.⇒ Need additional experiments!
Challenge: Design the mostinformative experiments to helpwith model discrimination.
Classical approach vs. Data-driven approachHow can we find the most informative next experiment, taking model uncertainty into account?
2.1 Classical Approach
Find optimal next experiment by exploiting themodels’ mathematical structure.
Time t
f(t)
/Inf
orm
atio
n Model 1 (withuncertainty)
Model 2 (withuncertainty)
Predictedinformationcontent
+ Computationally cheap.– Limits our choice of model structure.
2.2 Data-Driven Approach
Exploit power of computers to find the optimal nextexperiment through brute force.
Time t
f(t)
/Inf
orm
atio
n Model 1sample
Model 2sample
Predictedinformationcontent
+ Flexible with regards to model structure.– Computationally expensive.
Using machine learning, we hybridise the two approaches3. Novel Approach
We introduce data-driven surrogate models, whosemathematical structure can be exploited.
Time t
f(t)
/Inf
orm
atio
n Model 1 (withuncertainty)
Model 2 (withuncertainty)
Predictedinformationcontent
+ Flexible with regards to models structure.+ Computationally cheap.
4. ConclusionDifficulties with model discrimination often slowsdown e.g. development of new pharmaceuticals.
We hybridise two existing approaches, leveragingadvantages of both, by use of machine learning.
The novel approach is implemented and available asfree open-source software.
ReferencesS. Olofsson et al., 2019. arXiv e-print: 1810.02561.
S. Olofsson et al., 2018. Proc Mach Learn Res 80.
S. Olofsson et al., 2018. Comput Aided Chem Eng 44.
Acknowledgement. This work has received funding from the European Union’s Horizon 2020 research and innovation programmeunder the Marie Skłodowska-Curie grant agreement no. 675251, and an EPSRC Research Fellowship to R.M. (EP/P016871/1).
