.png "back to home page")
"Fairness is a process", some very strong and important #AI developments with a long way to go. And a bit of explaining why text, video when audio intelligence is hard. Facebook #F82019 #f8. pic.twitter.com/G8BLht4Dk9
— Wayne Kurtzman (@WayneNH) May 1, 2019
In previous F8s, @facebook 's policy problems discussed on first day, in Zuck keynote. 2nd day tech keynote was bouncy optimistic vision of cool future tech. This year it's reversed--tech keynote is about how AI can fix Facebook.
— Steven Levy (@StevenLevy) May 1, 2019
Facebook has talked a lot about the work it does in AI, to avoid bias, find hate speech on its platform etc. today. That's all very laudable. It is not using its keynotes to really talk about how developers can bring this to their own products, though. #f8 #f82019
— Frederic Lardinois (@fredericl) May 1, 2019
Facebook AI technology is getting steadily better at identifying what's going on in a photo, says Manohar Paluri. The same recognition technology with video, though, is "in its infancy." #F82019 pic.twitter.com/UxeiroVkKf
— Stephen Shankland (@stshank) May 1, 2019
Facebook’s content moderation: its AI falls behind when it comes to picking up hate speech and harassment in particular #f8 pic.twitter.com/nu1PkGBPDd
— Hannah Murphy (@MsHannahMurphy) May 1, 2019
Here the areas where Facebook's AI first started being effective at catching problems to, later down timeline, where it's catching most of the problems. pic.twitter.com/L4fOMuTse8
— Stephen Shankland (@stshank) May 1, 2019
Here's Facebook CTO @schrep at #F82019 conference for developers talking about the company's responsibility to keep toxicity at bay. We're going to be hearing a lot about AI helping this effort, no doubt. pic.twitter.com/4P7wYopc66
— Stephen Shankland (@stshank) May 1, 2019
Inclusive #AI is a framework that measures representation. User studies, algorithm development, system validation are used on a products across age, voice, gender and skin color to deliver a consistent experience for everyone. #F8 #F82019 #Facebook pic.twitter.com/FWdDZUdYeC
— Wayne Kurtzman (@WayneNH) May 1, 2019
Weird social media time capsule given I'm currently at a conference hearing about how AI can now tell the difference between pictures of weed and tempura broccoli on Facebook ?♂️
— Tom Hillman (@wholenewwworld) May 1, 2019
Facebook’s AI isn’t perfect at predicting misinformation, so it applies a threshold (or slider) to subgroups deemed higher risk. #f8 #f82019 fairness is a process... pic.twitter.com/CVV9BMDAMs
— Nick Fuller (@NicholasLFuller) May 1, 2019
It's probably harder for AI (and humans) to get the right answer with potato quality images that are pretty common — cheaper phones, dim lighting, low resolution, motion blur, etc. Interesting that Facebook looks at broader context for photos now for more signals.
— Stephen Shankland (@stshank) May 1, 2019
Exciting news! https://t.co/5Z4SYqt9o1
— Sebastian Thrun (@SebastianThrun) May 1, 2019
Facebook launches machine learning experimentation tool Ax#AI #machinelearning #BigData #IoT #deeplearning #ArtificialIntelligence #deeplearning #developers #softwaredevelopment #100daysofcode #ML #ArtificialIntelligence #DEVcommunity https://t.co/bnwNxHue95 via @VentureBeat
— Reinard Mortlock (@MortlockReinard) May 1, 2019
This is the culmination of a lot of hard work over the last several months. Not only with the core team but all of our partners like @Udacity, @jeremyphoward, Andrew Trask, Nick Felt and Google, and lots of others. Congrats to Team…https://t.co/KtctWyTIgW https://t.co/NollqBj7Aq
— Joseph Spisak (@joespeez) May 1, 2019
Update to Pytorch, including production use https://t.co/4pU2CY3wLi
— Ward Plunet (@StartupYou) May 1, 2019
At F8 Day 2, Facebook open sourced its AI experimentation platform Ax. While it could help competitors, open sourcing is a powerful engineer recruiting strategy https://t.co/FgJTKMMm0k
— Josh Constine (@JoshConstine) May 1, 2019
An overview of my talk today at #f8. Design is not neutral, our responsibility is broad, and the solutions to these complex problems are not easy. It all comes down to values. https://t.co/VC9JSMfIls
— Margaret Stewart (@mags) May 1, 2019
Facebook’s VP of Product Design, Margaret Stewart, shares how FB's values are reflected in what they build. #F8 https://t.co/Y7dHEkBseG
— Havas PR US (@HavasPRUS) May 1, 2019
There are many design challenges that come with building products for diverse, global audiences. Here's how we're working to overcome them: https://t.co/8mIqTRypEW #F8 pic.twitter.com/DrYNZbKdKk
— TechAtFacebook (@techatfacebook) May 1, 2019
When she didn't see the tools she needed to ensure that the AI powering Smart Camera on @PortalFacebook was inclusive, Lade set out to help build them // Read more on @techatfacebook: https://t.co/KHCdPnooh2
— Oculus (@oculus) May 1, 2019
Today at #F8, we talked about a process for inclusive #AI and how it provides a set of working guidelines for teams designing data sets, measuring product performance, and testing systems through the lens of inclusivity: https://t.co/EsEmAdyDCc pic.twitter.com/YufjfCuQje
— TechAtFacebook (@techatfacebook) May 1, 2019
After many years in the making, we're open-sourcing Facebook's Adaptive Experimentation platform, Ax, and BoTorch, a library for Bayesian optimization research, built on PyTorch. https://t.co/1wpFnlqmPg pic.twitter.com/vX8HeLF0Sg
— eytan bakshy (@eytan) May 1, 2019
I'm extremely excited to see BoTorch released: scalable, modular, and flexible Bayesian optimization, integrated with GPyTorch! It's been wonderful working with @eytan and his team.
— Andrew Gordon Wilson (@andrewgwils) May 1, 2019
announcement: https://t.co/u65EvH02Ez
website: https://t.co/ll8voLKoQp pic.twitter.com/vMuMo4vJKt
.@facebookAI Open-sourcing #Ax and #BoTorch: New AI tools for adaptive experimentation https://t.co/hSDBf3kBT2
— Michela Paganini (@WonderMicky) May 1, 2019
Congrats to FB’s engineers, researchers, and data scientists on the open source release of the rad experimentation & Bayesian optimization tools, Ax+BoTorch (https://t.co/1OnuJlM1tS).
— Mikhail Popov (@bearloga) May 1, 2019
If you want to learn the basics of BayesOpt, I have a tutorial for you: https://t.co/PmUzrR7MTI
React나 PyTorch를 봐도 알 수 있듯, 아무리 페이스북에 대한 여론이 안 좋더라도, 그 안 기술자의 양과 질을 고려할 때 소프트웨어 ‘벤더’로의 존재감은 구글과 쌍벽. #F8 https://t.co/lCYW8PFiWG
— Goodhyun 김국현 (@goodhyun) May 2, 2019
Facebook wants AI researchers to figure out privacy https://t.co/s1v6cAFW4g pic.twitter.com/Xyou21ebrx
— Rich Tehrani (@rtehrani) May 2, 2019
Facebook’s reputation around privacy may be at rock bottom but this is certainly welcome news. I only hope they employ some of the practices taught in this course. https://t.co/amWAg3ru4t
— Chris Briggs ????? (@confusedmatrix) May 1, 2019
#Facebook launches #machinelearning experimentation tool Ax | VentureBeat https://t.co/dRxj0xGiOr
— Carles Dijous (AAlb) (@carlesdijous) May 1, 2019
Facebook launches machine learning experimentation tool Ax #machinelearning https://t.co/yhV9h86YGx
— mark allen (@MarkAllen_Tech) May 1, 2019
#Facebook steps up work in #AI to systematically tackle its problems https://t.co/88kQ1MawUe pic.twitter.com/cwXuq5nAjG
— #AI (@AI__TECH) May 2, 2019
Facebook launches PyTorch 1.1 #f82019 https://t.co/Wi7ezhX9i7
— Emil Protalinski (@EPro) May 1, 2019
Facebook launches PyTorch 1.1 with TensorBoard support https://t.co/VSnigAzyDs pic.twitter.com/jt2ipOIcrI
— Rich Tehrani (@rtehrani) May 1, 2019
#Facebook open-sources #Ax and #BoTorch to simplify AI model optimization https://t.co/YcRSE2IPkZ#AI #machinelearning #nlp #alexa #deeplearning #skill #scientist #bigdata #artificialintelligence #amazon #apple #facebook #it #cloud #cloudcomputing pic.twitter.com/rdYwMQT9SC
— Rich Tehrani (@rtehrani) May 2, 2019
Facebook launches two new open source AI tools: BoTorch, based on PyTorch, for Bayesian library optimization and Ax, a platform for managing AI experiments (Frederic Lardinois/TechCrunch) https://t.co/qMsrF99r3X
— Chris Heilmann (@codepo8) May 1, 2019
Facebook open-sources Ax and BoTorch to simplify AI model optimization – TechCrunch https://t.co/YcRSE2IPkZ pic.twitter.com/5gZYdAdVn5
— Rich Tehrani (@rtehrani) May 1, 2019
I hope this "inclusive AI" thingy doesn't lead my robot to kill me in future for being a kuffar. Because in modern tech world, a particular religious group dominates such "inclusive" activities.?https://t.co/LLT3r4oMAK
— Talion (@Niderfan) May 1, 2019
It is encouraging to see Facebook's significant investment in Bayesian optimization development and deployment, and its impact on their #machinelearning leadership. We look forward to continuing to give our customers similar capabilities to these leaders. https://t.co/t8NVUweom3
— SigOpt (@SigOpt) May 1, 2019
Facebook: Open-sourcing Ax and BoTorch: New AI tools for adaptive experimentationhttps://t.co/vp2rM2ZKGl
— Frank Denis (@jedisct1) May 2, 2019
Believe it or not, Facebook is using AI to reduce abuse and bias - VentureBeat #AI #DigitalTransformation CC:@MikeQuindazzi https://t.co/fJC2TOZhYK
— ????????????４??? (@digital_trans4m) May 1, 2019
Facebook has a three-part plan
— Spiros Margaris (@SpirosMargaris) May 2, 2019
for making #AI more 'inclusive' https://t.co/glJSVKwaoe #fintech #ArtificialIntelligence #MachineLearning #DeepLearning @abcdedgar @engadget pic.twitter.com/K1ONrAFvru