Description
Your task is to find the mislabeled entries and submit their IDs to get the flag.
File Provided: emoji_dataset.csv
Solution
The challenge provides a Jupyter interface (the URL I was given, although obviously now it won’t work, was: https://86bdf9b5277d2eaf.chal.ctf.ae/?folder=/home/workspace/project)
We are given emoji_dataset.csv as an example dataset. The target is to create an algorithm that gives as an input the row number of misclassified emojis.
The following code worked to get the flag:
import pandas as pd
events = pd.read_csv("./data/emoji_dataset.csv")
# expected mapping
correct_mapping = {
"😺": "cat",
"🐶": "dog",
"🍎": "apple",
"🚗": "car",
}
mislabelled = events[events.apply(lambda r: correct_mapping.get(r['emoji']) != r['label'], axis=1)]
result = sorted(mislabelled['id'].tolist())
# show result
result