Retrieving relevant context to align representations for cross-lingual event detection