BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Topology in the furnace: TDA as a diagnostics tool for process con
 trol systems
DTSTART:20160519T141500
DTEND:20160519T153000
DTSTAMP:20260503T140713Z
UID:c25e84382515a6b0c6a72bc390987b8b412dcd71124dc70377dc1c77
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. Mikael Vejdemo-Johansson\, KTH Royal Institute of Technol
 ogy\, Stockholm\nBio: I am a postdoc specializing in applied algebraic to
 pology. My research is into topological data analysis\, partially of data 
 sets from robotics.\nI have previously been a postdoc in the applied topol
 ogy group at Stanford\, and in the computer algebra group at St Andrews.\n
 Steel smelting is a high-volume\, high-throughput industry\, where the sma
 llest performance gains translate into large dividends. Model construction
  to predict conditions inside the furnace is a centrally important part of
  process control. Machine learning and statistical methods have been shown
  to improve on purely metallurgical models\, but in either case\, the fail
 ure modes of the model are poorly understood\, and tools for analyzing the
 m not well developed.\nWe work in collaboration with Outukumppu Stainless 
 with their electric-arc scrap furnace\, to analyze and improve their tempe
 rature prediction models. Temperature prediction in particular is an impor
 tant model to improve: reference measurements can be done by inserting pro
 bes\, but these are costly and if measuring too early\, more probes will b
 e needed — measuring too late risks overheating the steel and spoiling t
 he entire batch.\nBased on work and ideas from Anthony Bak and Ayasdi\, an
 d in collaboration with Ayasdi\, we are studying the use of the Mapper alg
 orithm to construct intrinsic models of the fibres (preimages) of failed p
 redictions. These models help classify different modes of failure for the 
 models\, and direct attention for improvement or for learning compensation
  transforms to improve precision of temperature detection.\nIn this talk\,
  I will describe the approach we take for modeling and classifying failure
  modes\, and give some examples from our ongoing study of the steel smelti
 ng data. The talk will assume no previous knowledge of topological data an
 alysis\, and will explain Mapper completely and accessibly.  
LOCATION:CO 011
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR