SubmissionNumber#=%=#81
FinalPaperTitle#=%=#TLDR at SemEval-2024 Task 2: T5-generated clinical-Language summaries for DeBERTa Report Analysis
ShortPaperTitle#=%=#
NumberOfPages#=%=#10
CopyrightSigned#=%=#Shahriar Noroozizadeh
JobTitle#==#
Organization#==#Carnegie Mellon University, 
5000 Forbes Ave, Pittsburgh, PA 15213
Abstract#==#This paper introduces novel methodologies for the Natural Language Inference for Clinical Trials (NLI4CT) task. We present TLDR (T5-generated clinical-Language summaries for DeBERTa Report Analysis) which incorporates T5-model generated premise summaries for improved entailment and contradiction analysis in clinical NLI tasks. This approach overcomes the challenges posed by small context windows and lengthy premises, leading to a substantial improvement in Macro F1 scores: a 0.184 increase over truncated premises. Our comprehensive experimental evaluation, including detailed error analysis and ablations, confirms the superiority of TLDR in achieving consistency and faithfulness in predictions against semantically altered inputs.
Author{1}{Firstname}#=%=#Spandan
Author{1}{Lastname}#=%=#Das
Author{1}{Username}#=%=#spandand
Author{1}{Email}#=%=#spandand@andrew.cmu.edu
Author{1}{Affiliation}#=%=#Carnegie Mellon University
Author{2}{Firstname}#=%=#Vinay
Author{2}{Lastname}#=%=#Samuel
Author{2}{Username}#=%=#vsamuel13
Author{2}{Email}#=%=#vinaysamuel13@gmail.com
Author{2}{Affiliation}#=%=#Carnegie Mellon University
Author{3}{Firstname}#=%=#Shahriar
Author{3}{Lastname}#=%=#Noroozizadeh
Author{3}{Username}#=%=#snoroozi
Author{3}{Email}#=%=#snoroozi@cs.cmu.edu
Author{3}{Affiliation}#=%=#Carnegie Mellon University

==========
èéáğö