Detect clickbait without labels, then supervise

Q: Detect clickbait without labels, then supervise

This question evaluates a data scientist's competency in weak supervision, feature engineering from behavioral signals, model evaluation, and robustness to adversarial manipulation within the Machine Learning domain.

Q: How do I approach Machine Learning interview questions?

Machine Learning questions require understanding of core concepts and practice. PracHub provides solutions with explanations to help you master machine learning interviews.

Q: What difficulty level is this interview question?

This is a hard difficulty Machine Learning question, commonly asked during Onsite rounds.

Q: What role is this question designed for?

This question is commonly asked for Data Scientist candidates during technical interviews.

Question

Detecting Clickbait Ads Without Labeled Data

Context

You are asked to detect clickbait ad creatives when there is no labeled training data. You have impression/click logs, post-click signals (e.g., dwell time, bounce), and metadata (creative text/image, destination/publisher). The goal is to bootstrap weak labels from behavior, then train a supervised model that avoids over-blocking legitimate, high-quality novelty and is robust to adversarial gaming.

Tasks

Feature engineering: Propose features that capture short-term CTR and its decay among users who previously clicked a creative (e.g., 14-day within-viewer CTR delta), as well as post-click signals (dwell time on landing page, bounce).
Weak labeling via clustering: Cluster creatives in the CTR × decay space. Specify how you select the "likely clickbait" cluster while minimizing over-blocking of high-quality novelty.
Supervised model: Convert weak labels into a supervised model. Describe model features (e.g., text embeddings, historical user–ad interaction, publisher quality), training regimen, evaluation metrics (e.g., precision@k, uplift on long-term engagement), and guardrails.
Abuse prevention and monitoring: Outline how to prevent gaming (e.g., rotating creatives) and how to monitor for regression to the mean over time.

Detect clickbait without labels, then supervise

Quick Overview

Detect clickbait without labels, then supervise

Detecting Clickbait Ads Without Labeled Data

Context

Tasks

Write your answer