Thinking about grad school next year? Interested in visualization, human/AI interaction, computational art?
Consider Harvard!
@wattenberg
and I are building up our lab and will be looking for students.
Our DMs are open.
Visualize transformer attention! ✨
AttentionViz, created by Catherine Yeh and expanded by Yida Chen, helps you explore transformer self-attention by visualizing query and key vectors in a joint embedding.
Paper:
Website:
I’m thrilled to announce that, as of January 2022, I’ll be joining Harvard as a professor of computer science, with an affiliation with the business school.
@wattenberg
joins in fall 2021! We’ll also continue to work, part-time, at Google and on PAIR.
Our Nature paper on AI Predicting Aftershocks is out! Joint work between Google and Harvard (with Phoebe DeVries,
@brendanjmeade
,
@wattenberg
). An example of how interpretability can help us learn from ML systems to do science we couldn't do before:
…
Our team () is looking for talented front-end engineers as well as data visualization creators. Interest in machine learning is a must.
For either position, please apply here:
The Periodic Table of Elements turns 150 years old this month and the history of its creation is astonishing. In celebration, here's a visualization our team made a few years ago:
Super proud of my team’s work on the first ever AI-powered Google Doodle in celebration of Bach’s birthday. This was a collaboration with the Magenta, Doodle and TensorFlow.js teams. Congrats
@bengiswex
,
@nsthorat
and
@dsmilkov
! Compositions happening around the globe today. 😃
Looking for a postdoc position in data visualization, human-centered machine learning, or both? Consider joining our lab at Harvard!
Here’s the official announcement:
Applications for the 2018 Google AI Residency Program are now open. Come join us for a year! (in case you're wondering: this was formerly known as the Google Brain Residency Program)
Thrilled to announce the People + AI Guidebook! A set of guidelines for UXers and PMs shipping AI products, based on data and insights from 100+ individuals across Google teams, industry experts, and academics:
#pairguidebook
We'd love your feedback.
Excited to see
@QuantaMagazine
profiling
@_beenkim
on how she's developing a “translator for humans” so that we can understand when AI systems break down:
You can now search for datasets on Google:
And, because one must test data tools with topics of importance, I looked for "chocolate+UK." Who knew that the top UK chocolate bars for all ages are Mars and Galaxy and Snickers?
Congrats, Natasha Noy & team!
New from
@adamrpearce
and the
#googlePAIR
team: AI Explorables that let you interactively play with ML mechanics to see their effect on key AI issues such as fairness, explainability, transparency, bias, privacy.
Give it a try:
Looking back at PAIR's first year: how we built technology, undertook research and published articles that aimed to broaden the conversation about the partnership between humans and AI:
Honored to become part of the SIGCHI Academy! In awe of the other awardees and thrilled to see such a strong data visualization cohort:
@jeffrey_heer
and
@magrawala
How does a neural net see depth in paintings? Does it "resolve" the depth depicted in these images in the same way we do?
Read about depth predictions:
PAIR project by
@EllenJiang2
,
@emilyrreif
,
@_beenkim
AI is a powerful tool for improving society. I'm excited about our AI for Social Good program, focusing on both research & empowering the ecosystem.
If you have ideas, submit to the AI Impact Challenge! We're funding the best ideas w/$25M. Learn more at
Analyzing and visualizing syntax trees in the high-dimensional spaces of neural nets.
Check out the new PAIR paper on BERT geometry
And the blog post on “Language, trees, and geometry in neural networks”
Toasters have blinking lights, cars have speedometers. Should chatbots have dashboards too?
A speculative essay👉 The System Model and the User Model: Exploring AI Dashboard Design
Excited to announce the 2020
#PAIRSymposium
, on Nov. 18 from London, Boston and Seattle!
This year we focus on
#BoundaryObjects
as a way to broaden the set of stakeholders in
#ParticipatoryML
.
Set a reminder now at !
Thrilled to see Know Your Data launch! The tool brings a new level of transparency to ML data, code free, allowing a broad set of stakeholders to ask critical questions
Creation of
@nsthorat
and
@dsmilkov
in
#googlePAIR
; collaboration with
@GoogleSRE
team
I recently spoke with Vogue Brasil about AI and explained my view that the technology is neither utopian nor dystopian but something in between, that needs to be developed responsibly.
Happy to see the magazine covering AI:
A tantalizing source of urban history in Amsterdam: a visual catalog of objects found in a riverbed along the centuries. Explore the interactive timeline going back to BCE:
1/3
"We went to Brazil because we wanted to escape" vs
"We went to Japan because we wanted to study"
Check out
@adamrpearce
's language model explorable, showing what models like BERT have learned about the world.
#googlePAIR
Type your own queries:
Excited to see the Language Interpretability Tool (LIT) launched! Data visualization for NLP models; what could be better? :)
This is a collaboration between
#googlePAIR
and NLP researchers at
@googIeresearch
Introducing an early-release version of the Language Interpretability Tool (LIT), a visual, interactive, and extensible open-source tool for analyzing all sorts of NLP models 🔥
Code:
Paper:
#NLProc
(1/4)
Are you working on an AI project to improve the world around you? Google will select projects to support through the AI Impact Challenge (a $25M funding pool). The deadline for proposals is next Tuesday, Jan 22:
Language models have some beautiful spiral plots reflecting positional patterns. And a vision model has heads that arrange images according to brightness and hues. But there’s a lot more to find! What else can you see?
A good color scale makes a ton of difference.
Up until yesterday, this
@nytimes
COVID map showed almost all European countries in red. Today, with an updated color scale, we can see there are major differences.
#DataVisualization
Excited to be speaking at the Bank of England tomorrow! Chief Economist Andy Haldane and I will be talking about data visualization for machine learning.
The webinar is virtual and open to the public.
Interested in data and AI? Join us on 18 September for an interactive webinar with senior researcher at Google and data visualization expert, Fernanda Viegas. Find out more and register for your place here:
I’ll be moderating a roster of brilliant speakers:
- Charlotte Lee
@ducktopian
, University of Washington,
- Annelien Bredenoord
@ALBredenoord
, UMC Utrecht & Uni Utrecht; Dutch Senator
- Catherine D'Ignazio
@kanarinka
, MIT
- Nithya Sambasivan,
@autopoietic
,
@GoogleAI
Are you an academic working in the areas of Participatory ML, Accessibility, Algorithmic Fairness or Higher Education in CS? If so, apply to Google’s Award for Inclusion Research (AIR)!
AIR is an international program 🤩
Application deadline: July 14
With a click of a button you can compare a datapoint to its closest counterfactual--the most similar point where your model predicts a different result.
PAIR just launched a Medium channel to explore AI issues we care about with the broad community, including participatory design, fairness and explicability.
Today: Q&A with Sabelo Mhlambi about Ubuntu ethics & AI
#googlePAIR
How do you visualize techniques for data collection that help preserve privacy?
Thrilled to see this new
#googlePAIR
AI Explorable by
@adamrpearce
and
@EllenJiang2
--launched today!
The availability of giant datasets and faster computers is making it harder to collect and study private information without inadvertently violating people’s privacy.
@EllenJiang2
and I took a look at how random numbers can help.
Hannah Fry
@FryRsquared
does a great job of showing how "humble" graphs such as line charts and scatter plots have changed the world and our understanding of it:
Thanks for the pointer,
@m_pellat
!