The IWSLT 2017 evaluation campaign has organised three tasks. The Multilingual task, which
is about training machine translation systems handling many-to-many language directions…

… The goal of the conference is to organize evaluations and sessions … on the evaluation
campaign organized by IWSLT 2020, … In their respective description papers, the two teams scored …

… This overview provides detailed descriptions of the data and evaluation conditions of
each … In the following, we describe each task in detail, provide a summary of the received …

… To complete the overview on the evolution of the evaluation campaign, we analyse the … if
to apply in the future (starting from 2017) human evaluations based on ranking or Likert scales. …

… This paper reports on the shared tasks organized by the 20th IWSLT Conference. The …
contributors to the overview paper, almost evenly distributed across industry and academia. …

… evaluation campaign featuring shared tasks addressing scientific challenges in SLT. This
paper reports on the 2024 IWSLT Evaluation Campaign, … Each section includes a description of …

… The other observation is about the introduction of human evaluation in our task. While
largely confirming the rankings obtained with automatic metrics, it provides the most reliable …

… evaluated before during IWSLT campaigns. Three well-known metrics were used for scoring
the results: Bilingual Evaluation … permissible for the IWSLT 2015 campaign included: data …

… In this evaluation campaign, we built only sequence-tosequence speech recognition (ASR) …
recognition architectures we employed in the campaign. Afterwards, we give a detailed de…

… IWSLT2020 Speech Translation evaluation campaign. We first … the “constrained” conditions
for the evaluation campaign). It is … 2017) to have better sentence boundaries, such method …