Update annotations for Iraa/paper_11.txt

annotations/Iraa/paper_11.txt.json

@@ -14,5 +14,13 @@
     "label": "Unsupported claim",
     "user": "Iraa",
     "text": "multilingual BERT"
+  },
+  {
+    "file": "paper_11.txt",
+    "start": 1961,
+    "end": 2746,
+    "label": "Coherence",
+    "user": "Iraa",
+    "text": "Can a highperformance CLIR model be trained that can operate without having to rely on MT? To answer the question, instead of viewing the MT-based approach as a competing one, we propose to leverage its strength via knowledge distillation (KD) into an end-to-end CLIR model. KD (Hinton et al., 2014) is a powerful supervision technique typically used to distill the knowledge of a large teacher model about some task into a smaller student model (Mukherjee and Awadallah, 2020;Turc et al., 2020). Here we propose to use it in a slightly different context, where the teacher and the student retriever are identical in size, but the former has superior performance simply due to utilizing MT output and consequently operating in a high-resource and lowdifficulty monolingual environment."
   }
 ]