?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Stochastic+Optimization+with+Adaptive+Batch+Size%3A%0D%0ADiscrete+Choice+Models+as+a+Case+Study&rft.creator=Lederrey%2C+Gael&rft.creator=Lurkin%2C+Virginie&rft.creator=Hillel%2C+Tim&rft.creator=Bierlaire%2C+Michel&rft.description=The+2.5+quintillion+bytes+of+data+created+each+day+brings+new+opportunities%2C+but+also+new%0D%0Astimulating+challenges+for+the+discrete+choice+community.+Opportunities+because+more+and+more%0D%0Anew+and+larger+data+sets+will+undoubtedly+become+available+in+the+future.+Challenging+because%0D%0Ainsights+can+only+be+discovered+if+models+can+be+estimated%2C+which+is+not+simple+on+these+large%0D%0Adatasets.%0D%0AIn+this+paper%2C+inspired+by+the+good+practices+and+the+intensive+use+of+stochastic+gradient+methods%0D%0Ain+the+ML+field%2C+we+introduce+the+algorithm+called+Window+Moving+Average+-+Adaptive+Batch%0D%0ASize+(WMA-ABS)+which+is+used+to+improve+the+efficiency+of+stochastic+second-order+methods.%0D%0AWe+present+preliminary+results+that+indicate+that+our+algorithms+outperform+the+standard+secondorder+methods%2C+especially+for+large+datasets.+It+constitutes+a+first+step+to+show+that+stochastic%0D%0Aalgorithms+can+finally+find+their+place+in+the+optimization+of+Discrete+Choice+Models.&rft.subject=Optimization%2C+Discrete+Choice+Models%2C+Stochastic+Algorithms%2C+Adaptive+Batch+Size&rft.publisher=Institute+for+Economic+Research%2C+Universit%C3%A0+della+Svizzera+italiana&rft.contributor=Scherer%2C+Patrick&rft.date=2019-05&rft.type=Proceedings+paper&rft.language=eng&rft.source=+++++In%3A+Scherer%2C+Patrick%2C+(ed.)+19th+Swiss+Transport+Research+Conference.++++Institute+for+Economic+Research%2C+Universit%C3%A0+della+Svizzera+italiana%3A+Ascona%2C+Switzerland.+(2019)+++++&rft.format=application%2Fpdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10174125%2F1%2FLederrey_EtAl%2520%25281%2529.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10174125%2F&rft.rights=open