YoVDO

Humpty Dumpty - Controlling Word Meanings via Corpus Poisoning

Offered By: IEEE via YouTube

Tags

IEEE Symposium on Security and Privacy Courses Information Retrieval Courses Data Security Courses Word Embeddings Courses Adversarial Machine Learning Courses

Course Description

Overview

Explore a 17-minute IEEE conference talk that delves into the manipulation of word embeddings through corpus modifications. Learn how attackers can control the semantic meaning of words by altering their positions in embedding spaces. Discover the explicit expression used as a proxy for word distances and its causal relationship with embedding distances. Examine two adversarial objectives: making a word a top-ranked neighbor of another and moving words between semantic clusters. Understand the far-reaching implications of such attacks on various downstream tasks in transfer learning scenarios. Investigate how these manipulations affect query expansion in information retrieval systems, influence named entity recognition models, and impact machine translation. Finally, gain insights into how attackers can generate linguistically plausible corpus modifications to bypass language model-based defenses.

Syllabus

Humpty Dumpty: Controlling Word Meanings via Corpus Poisoning


Taught by

IEEE Symposium on Security and Privacy

Tags

Related Courses

Semantic Web Technologies
openHPI
أساسيات استرجاع المعلومات
Rwaq (رواق)
《gacco特別企画》Evernoteで広がるgaccoの学びスタイル (ga038)
University of Tokyo via gacco
La Web Semántica: Herramientas para la publicación y extracción efectiva de información en la Web
Pontificia Universidad Católica de Chile via Coursera
快速学习
University of Science and Technology of China via Coursera