ML models for adversarial Abuse: A content perspective

Using ML for detection of policy violating content on YouTube

Machine Learning models for policy violating text detection

How Pinterest uses machine learning models to reduce unsafe content

Temporal interaction embeddings for unsafe content

Moderating promotional spam and NSFW content in photos at scale at Yelp

Near-duplicate detection to augment anti-abuse efforts

Metadata-based detection of Child Sexual Abuse Material

How Pinterest uses machine learning models to combat adversarial abuse

 Mining suspicious multi-attribute entity groups with multi-view graphs

Content agnostic adversarial abuse fighting

Peeking into the future: Using hacked account behavioral models to label malicious login requests

Review recommendation at Yelp: ground truth sampling design for reliable model training

Building scaled systems and machine learning models in an adversarial space

Attendees can raise their hand and come off mute to make relevant Trust & Safety announcements such as, "we just released this library" or "check out this book I just published on Trust & Safety best practices" etc