Correction for Paper Metadata: {2024.iwslt-1.20} #3778 (#3795)

data/xml/2024.iwslt.xml

@@ -306,7 +306,12 @@
       <title><fixed-case>CMU</fixed-case>’s <fixed-case>IWSLT</fixed-case> 2024 Simultaneous Speech Translation System</title>
       <author><first>Xi</first><last>Xu</last><affiliation>Carnegie Mellon University</affiliation></author>
       <author><first>Siqi</first><last>Ouyang</last><affiliation>Carnegie Mellon University</affiliation></author>
+      <author><first>Brian</first><last>Yan</last><affiliation>Carnegie Mellon University</affiliation></author>
+      <author><first>Patrick</first><last>Fernandes</last><affiliation>Carnegie Mellon University</affiliation></author>
+      <author><first>William</first><last>Chen</last><affiliation>Carnegie Mellon University</affiliation></author>
       <author><first>Lei</first><last>Li</last><affiliation>Carnegie Mellon University</affiliation></author>
+      <author><first>Graham</first><last>Neubig</last><affiliation>Carnegie Mellon University</affiliation></author>
+      <author><first>Shinji</first><last>Watanabe</last><affiliation>Carnegie Mellon University</affiliation></author>
       <pages>154-159</pages>
       <abstract>This paper describes CMU’s submission to the IWSLT 2024 Simultaneous Speech Translation (SST) task for translating English speech to German text in a streaming manner. Our end-to-end speech-to-text (ST) system integrates the WavLM speech encoder, a modality adapter, and the Llama2-7B-Base model as the decoder. We employ a two-stage training approach: initially, we align the representations of speech and text, followed by full fine-tuning. Both stages are trained on MuST-c v2 data with cross-entropy loss. We adapt our offline ST model for SST using a simple fixed hold-n policy. Experiments show that our model obtains an offline BLEU score of 31.1 and a BLEU score of 29.5 under 2 seconds latency on the MuST-C-v2 tst-COMMON.</abstract>
       <url hash="fd2a633a">2024.iwslt-1.20</url>