SubmissionNumber#=%=#236
FinalPaperTitle#=%=#LMU-BioNLP at SemEval-2024 Task 2: Large Diverse Ensembles for Robust Clinical NLI
ShortPaperTitle#=%=#
NumberOfPages#=%=#7
CopyrightSigned#=%=#Leon Weber-Genzel
JobTitle#==#
Organization#==#
Abstract#==#In this paper, we describe our submission for the NLI4CT 2024 shared task on robust Natural Language Inference over clinical trial reports. Our system is an ensemble of nine diverse models which we aggregate via majority voting. The models use a large spectrum of different approaches ranging from a straightforward Convolutional Neural Network over fine-tuned Large Language Models to few-shot-prompted language models using chain-of-thought reasoning.
Surprisingly, we find that some individual ensemble members are not only more accurate than the final ensemble model but also more robust.
Author{1}{Firstname}#=%=#Zihang
Author{1}{Lastname}#=%=#Sun
Author{1}{Email}#=%=#Zihang.Sun@campus.lmu.de
Author{1}{Affiliation}#=%=#LMU Munich
Author{2}{Firstname}#=%=#Danqi
Author{2}{Lastname}#=%=#Yan
Author{2}{Email}#=%=#Danqi.Yan@campus.lmu.de
Author{2}{Affiliation}#=%=#LMU Munich
Author{3}{Firstname}#=%=#Anyi
Author{3}{Lastname}#=%=#Wang
Author{3}{Email}#=%=#Anyi.Wang@campus.lmu.de
Author{3}{Affiliation}#=%=#LMU Munich
Author{4}{Firstname}#=%=#Tanalp
Author{4}{Lastname}#=%=#Agustoslu
Author{4}{Email}#=%=#T.Agustoslu@campus.lmu.de
Author{4}{Affiliation}#=%=#LMU Munich
Author{5}{Firstname}#=%=#Qi
Author{5}{Lastname}#=%=#Feng
Author{5}{Email}#=%=#Q.Feng@campus.lmu.de
Author{5}{Affiliation}#=%=#LMU Munich
Author{6}{Firstname}#=%=#Chengzhi
Author{6}{Lastname}#=%=#Hu
Author{6}{Email}#=%=#Chengzhi.Hu@campus.lmu.de
Author{6}{Affiliation}#=%=#LMU Munich
Author{7}{Firstname}#=%=#Longfei
Author{7}{Lastname}#=%=#Zuo
Author{7}{Email}#=%=#Zuo.Longfei@campus.lmu.de
Author{7}{Affiliation}#=%=#LMU Munich
Author{8}{Firstname}#=%=#Shijia
Author{8}{Lastname}#=%=#Zhou
Author{8}{Email}#=%=#Zhou.Shijia@campus.lmu.de
Author{8}{Affiliation}#=%=#LMU Munich
Author{9}{Firstname}#=%=#Hermine
Author{9}{Lastname}#=%=#Kleiner
Author{9}{Email}#=%=#H.Kleiner@campus.lmu.de
Author{9}{Affiliation}#=%=#LMU Munich
Author{10}{Firstname}#=%=#Pingjun
Author{10}{Lastname}#=%=#Hong
Author{10}{Email}#=%=#Pingjun.Hong@campus.lmu.de
Author{10}{Affiliation}#=%=#LMU Munich
Author{11}{Firstname}#=%=#Suteera
Author{11}{Lastname}#=%=#Seeha
Author{11}{Email}#=%=#S.Seeha@campus.lmu.de
Author{11}{Affiliation}#=%=#LMU Munich
Author{12}{Firstname}#=%=#Sebastian
Author{12}{Lastname}#=%=#Loftus
Author{12}{Email}#=%=#S.Loftus@campus.lmu.de
Author{12}{Affiliation}#=%=#LMU Munich
Author{13}{Firstname}#=%=#Anna
Author{13}{Lastname}#=%=#Barwig
Author{13}{Email}#=%=#A.Barwig@campus.lmu.de
Author{13}{Affiliation}#=%=#LMU Munich
Author{14}{Firstname}#=%=#Oliver
Author{14}{Lastname}#=%=#Kraus
Author{14}{Email}#=%=#O.Kraus2@campus.lmu.de
Author{14}{Affiliation}#=%=#LMU Munich
Author{15}{Firstname}#=%=#Jona
Author{15}{Lastname}#=%=#Volohonsky
Author{15}{Email}#=%=#J.Volohonsky@campus.lmu.de
Author{15}{Affiliation}#=%=#LMU Munich
Author{16}{Firstname}#=%=#Yang
Author{16}{Lastname}#=%=#Sun
Author{16}{Email}#=%=#Yang.Sun@campus.lmu.de
Author{16}{Affiliation}#=%=#LMU Munich
Author{17}{Firstname}#=%=#Leopold
Author{17}{Lastname}#=%=#Martin
Author{17}{Email}#=%=#Leopold.Martin@campus.lmu.de
Author{17}{Affiliation}#=%=#LMU Munich
Author{18}{Firstname}#=%=#Lena
Author{18}{Lastname}#=%=#Altinger
Author{18}{Email}#=%=#L.Altinger@campus.lmu.de
Author{18}{Affiliation}#=%=#LMU Munich
Author{19}{Firstname}#=%=#Jing
Author{19}{Lastname}#=%=#Wang
Author{19}{Email}#=%=#Jing.Wang1@campus.lmu.de
Author{19}{Affiliation}#=%=#LMU Munich
Author{20}{Firstname}#=%=#Leon
Author{20}{Lastname}#=%=#Weber
Author{20}{Username}#=%=#leonweber
Author{20}{Email}#=%=#leonweber@posteo.de
Author{20}{Affiliation}#=%=#LMU Munich

==========
èéáğö