BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Statistics and Machine Learning in (Bio) Chemical Engineering - An
  Open Workshop - Prof Alexei Lapkin et al
DTSTART:20180607T083000Z
DTEND:20180607T160000Z
UID:TALK106240@talks.cam.ac.uk
CONTACT:Alexei Lapkin
DESCRIPTION:9:30	Introduction to the Challenge-led Talks/Discussion (A. La
 pkin)\n\nTalks and discussion on the topics of surrogate functions\, optim
 isation\, classification\, data mining\, automation and more in-depth topi
 cs\, such as specific methods of tackling various uncertainties in model d
 evelopment.\n\n“Interpretable ML for Chemistry: Designing algorithms wit
 h statistical physics and extracting chemical knowledge from results”\, 
 Dr Alpha Lee\, Cavendish Laboratory.\n\n13:00	Lunch break\n\n14:00	Seminar
  talk: “Closed-loop automatic experimentation for optimisation”\nDave 
 Woods\, Professor of Statistics in the Southampton Statistical Sciences Re
 search Institute\n\nAbstract: Automated experimental systems\, involving m
 inimal human intervention\, are becoming more popular and common\, providi
 ng economical and fast data collection. We discuss some statistical issues
  around the design of experiments and data modelling for such systems. Our
  application is to “closed-loop” optimisation of chemical processes\, 
 where automation of reaction synthesis\, chemical analysis and statistical
  design and modelling increases lab efficiency and allows 24/7 use of equi
 pment.\nOur approach uses nonparametric regression modelling\, specificall
 y Gaussian process regression\, to allow flexible and robust modelling of 
 potentially complex relationships between reaction conditions and measured
  responses. A Bayesian approach is adopted to uncertainty quantification\,
  facilitated through computationally efficient Sequential Monte Carlo algo
 rithms for the approximation of the posterior predictive distribution. We 
 propose a new criterion\, Expected Gain in Utility (EGU)\, for optimisatio
 n of a noisy response via fully-sequential design of experiments\, and we 
 compare the performance of EGU to extensions of the Expected Improvement c
 riterion\, which is popular for optimisation of deterministic functions. W
 e also show how the modelling and design can be adapted to identify\, and 
 then down-weight\, potentially outlying observations to obtain a more robu
 st analysis.\n\n15:00	Introduction to the EPSRC project “Combining Chemi
 cal Robotics and Statistical Methods to Discover Complex Functional Produc
 ts”\, a collaboration between Universities of Cambridge\, Glasgow and So
 uthampton\n15:30	Coffee/Networking\n16:00	Afternoon workshop session\n17:0
 0	Close of workshop\n\nPlease register at\nhttps://www.eventbrite.co.uk/e/
 statistics-and-machine-learning-in-bio-chemical-engineering-tickets-460756
 27442
LOCATION:Maxwell Centre\, Cavendish Laboratory\, JJ Thomson Ave\, Cambridg
 e CB3 0HE. Small Lecture Theatre
END:VEVENT
END:VCALENDAR