Group items tagged

The blind application of machine learning runs the risk of amplifying biases present in data. Such a danger is facing us with word embedding, a popular framework to represent text data as vectors which has been used in many machine learning and natural language processing tasks. We show that even word embeddings trained on Google News articles exhibit female/male gender stereotypes to a disturbing extent. This raises concerns because their widespread use, as we describe, often tends to amplify these biases. Geometrically, gender bias is first shown to be captured by a direction in the word embedding. Second, gender neutral words are shown to be linearly separable from gender definition words in the word embedding. Using these properties, we provide a methodology for modifying an embedding to remove gender stereotypes, such as the association between between the words receptionist and female, while maintaining desired associations such as between the words queen and female. We define metrics to quantify both direct and indirect gender biases in embeddings, and develop algorithms to "debias" the embedding. Using crowd-worker evaluation as well as standard benchmarks, we empirically demonstrate that our algorithms significantly reduce gender bias in embeddings while preserving the its useful properties such as the ability to cluster related concepts and to solve analogy tasks. The resulting embeddings can be used in applications without amplifying gender bias.

These principles were developed in conjunction with the 2017 Asilomar conference

" "I Agree" is an art installation that prints the terms of service for common apps on scrolls of colored paper, creating a bar chart of the fine print that neither you, nor anyone else in the history of the world, has ever read. "

"EPSRC-funded research project, we use the concept of the 'Digital Lifespan' to investigate how digital identities are created and managed across the human lifespan by UK citizens"

A couple of years ago, I was heavily involved in analysing and reporting on the massive VTech hack, the one where millions of records were exposed including kids' names, genders, ages, photos and the relationship to parents' records which included their home address. Part of this data was collected via an IoT device called the InnoTab which is a wifi connected tablet designed for young kids; think Fisher Price designing an iPad... then totally screwing up the security.

"This is one of the minor dangers of owning an Amazon Echo. Most anyone can activate the device by just saying the wake word, including Cartman. And in the season opener of South Park, that's exactly what happened. If someone watched that episode of South Park in the same room as an Echo, their Amazon shopping list was filled with random, gross items. This example shows the potential danger of having a voice-activated shopping assistant. It's easy to imagine a potential rogue advertisement, online or elsewhere, that could, in theory, say the right words to order a particular product - like a South Park box set."

Pay as you play, No monthly fees, Own music again