Profile

Cover photo
Verified name
Jeff Dean
Works at Google
Attended University of Washington
Lives in Palo Alto
51,258 followers|8,435,004 views
AboutPostsPhotosVideos+1's

Stream

Jeff Dean

Shared publicly  - 
 
The Google Brain team — Looking Back on 2016

I wrote up a blog post about the work the Google Brain team has been doing over 2016. I'm really excited to work with such great colleagues! In writing this up, it's pretty remarkable to me that nearly every other sentence has one or a few links to many more details on significant and impactful work.


90
11
Tony Yu's profile photoAl Roth's profile photoNapatsanun Kulpatwattana's profile photo
3 comments
 
Thank you very much for your kindness and all Google Teacher everyone for all of your kindness support to my improvement Technology skills and gave me a big changing of my life.
Add a comment...

Jeff Dean

Shared publicly  - 
 
An inside look at the Google Brain team

For the past few months, we've had NY Times Magazine reporter +Gideon Lewis-Kraus visit the Google Brain team several times and hang out with us for a few days at a time, with an eye towards writing an article about how our research team operates and what we're working on. We gave him pretty open access to our building, the people in the team, many of our meetings, etc., and over the course of several visits, he decided to focus his story on the origins of the Brain team, and on our in-progress collaboration with the Google Translate team to replace the old phrase-based translation system with a neural machine translation system (essentially part of the article is a behind-the-scenes look at how the scientific work in https://arxiv.org/abs/1609.08144 and https://arxiv.org/abs/1611.04558 came about).

This long article is the result of his visits and synthesis of what he learned. Gideon, I think the article turned out really well!



How Google used artificial intelligence to transform Google Translate, one of its more popular services — and how machine learning is poised to reinvent computing itself.
149
95
Erika French's profile photoFlora Wong's profile photoAndy Man's profile photoAndre Amorim's profile photo
6 comments
 
I would like to have my body in cryonic suspension too ... Just like Hal Finney and Marvin Minsky ... skip some years .. hope for better in the next life (or next awake) https://imgs.xkcd.com/comics/cryogenics.png 
Add a comment...

Jeff Dean

Shared publicly  - 
 
Zero-shot translations: translate Korean->Japanese without ever seeing Korean->Japanese training data.

One of the very interesting aspects of our recent work on neural machine translation is that we can train a single, multi-lingual model that is able to translate between many language pairs (say, English->Japanese, English->Korean, Japanese->English, and Korean->English). This often is able to improve quality for many of the language pairs, and also simplifies many production aspects of the system. However, a really interesting property is that a system trained on these language pairs is actually able to do a reasonable job of translating between language pairs where it has never seen training data before (say, Korean->Japanese).

This blog post gives the high level details:

https://research.googleblog.com/2016/11/zero-shot-translation-with-googles.html

Our Arxiv paper has details about this approach:

https://arxiv.org/abs/1611.04558

119
20
Andre Amorim's profile photoSigfredo Zamorano's profile photoScott MacLeod (Helianth)'s profile photo
9 comments
 
How best to add the remaining 6,994 known living languages, Jeff - https://scott-macleod.blogspot.com/2017/01/zebu-cows-breakthrough-googles-zero.html - and as what I'm calling as a new method "ethno-wiki-virtual-world-graphy" ? Thanks!
Add a comment...

Jeff Dean

Shared publicly  - 
 
I'm very excited to be headed to my first Grace Hopper Conference today through Friday, celebrating women in computer science. Diversity in our field is something I care deeply about, and this will be a great chance to listen and learn. If you're there and see me, please say hi. I'd love to chat about machine learning and our group (g.co/brain), TensorFlow (tensorflow.org), computer systems, diversity in CS, the Google Brain Residency program (g.co/brainresidency), or anything else. I'll be at the Google Career Fair booth (#1730) from 10 AM to 12 PM (Houston time) on Thursday, and around various sessions and events the rest of the time.
108
1
Jessica Liu Strohmann's profile photoRICARDO MARTINEZ's profile photoAmuza Wanyanga's profile photoBritta Petersen's profile photo
4 comments
 
I have got a mail where I have won a lot of money from Google..... Is that for real or fake? Your name is mentioned and so is Sundar Pichai. I don't believe it is real, that is why I need to ask you. Answer at Brittabaypetersen@ihsdata.dk. Thanks.
Add a comment...

Jeff Dean

Shared publicly  - 
 
A blog post from our group (Google Brain team, g.co/brain), in collaboration with [X] and DeepMind, describing three different approaches for robots to pool their experience to acquire basic motor skills. The blog post is an excellent overview of the work, and there are links to four detailed research papers that describe the work in more detail at the end of the blog post.

We're quite excited by our growing research efforts around machine learning for robotics.


88
20
Jamshed Khan's profile photoAndre Amorim's profile photoDyna Fitria's profile photo
3 comments
 
Ğu jbjhjjj
Add a comment...

Jeff Dean

Shared publicly  - 
 
Neural Machine Translation: Much better translation quality

Full technical report (23 exciting pages of bedtime reading):
http://arxiv.org/pdf/1609.08144v1.pdf

Research blog post: https://research.googleblog.com/2016/09/a-neural-network-for-machine.html

I'm very excited to announce that our new neural machine translation system closes the quality gap between the existing Google Translate production system and human quality translations by 58% to 87% for a variety of different language pairs (see table below, from the technical report we published today). This work has been a close collaboration between the Google Brain team (g.co/brain) and the Google Translate team.

Thanks to lots of hard engineering work and the computational efficiency of our Tensor Processing Units (see report), we are also rolling these benefits out to users of Google Translate, starting today with Mandarin to English as the first language pair live in production that uses this new system. We'll be rolling out many more language pairs over the coming weeks.

This highlights the success of neural models at more accurately capturing the complexities of real human language, and is a powerful demonstration of the research our group has been doing on language understanding.

(Edit: moved arxiv and blog link above the fold)
144
29
Tau-Mu Yi's profile photoPedro Marcal's profile photoKirti Vashee's profile photoTirumala Krishna's profile photo
7 comments
 
Nice info
Add a comment...

Jeff Dean

Shared publicly  - 
 
In 2016, we welcomed into the Google Brain team our first batch of Google Brain Residents (g.co/brainresidency). We just published a blog post about what the 27 residents have been up roughly halfway into their one year residency program. They're an impressive group and the vitality and energy they've added to our group has been fantastic!

The deadline for applying for this year's program is coming up on January 13th, so if you're interested in coming to the Brain team for a year to do research and be mentored by our research scientists, I would encourage you to apply.


53
2
Kris Oud's profile photoPaolo Pignatelli's profile photo
2 comments
 
Happy New year, Google Research! Thank you for all your work, especially your publicly published research available from your blog posts, and your data-sets. To each and every researcher in the nlu, grazie, merci danke shoen. 
Add a comment...

Jeff Dean

Shared publicly  - 
 
My friend +Nikhil Buduma's company, Remedy, is on https://www.producthunt.com today. They're an affordable, concierge medicine company that use machine learning to automatically generate a series of questions to help triage your symptoms, and then allow you to get advice and/or prescriptions from a real doctor with much less hassle and much lower costs than traditional doctor visits. Check them out!

https://www.remedymedical.com/
https://www.remedymedical.com/concierge


Product Hunt is a curation of the best new products, every day. Discover the latest mobile apps, websites, and technology products that everyone's talking about.
43
4
Add a comment...

Jeff Dean

Shared publicly  - 
 
A machine learning model that is better than the median board-certified ophthalmologist in assessing signs of diabetic retinopathy

The Google Brain team (http://g.co/brain) has been focusing some of our efforts on how machine learning can transform healthcare. We're very excited about the opportunities to provide better and more accessible care, and to save lives and make people healthier. Some of my colleagues have been working on automated systems for assessing retinal images for signs of diabetic retinopathy (DR), a degenerative eye disease that if not caught can cause blindness, and is a great example of where machine learning can really help healthcare providers. They have written a paper that was published today in the Journal of the American Medical Association, titled "Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs".

The automated system achieves an accuracy in assessing retinal images for signs of diabetic retinopathy that is higher than the median accuracy of eight board-certified ophthalmologists. This is an example of the transformative potential of machine learning for healthcare, because in many parts of the world, there simply aren't enough ophthalmologists to screen everyone for DR (the actual cameras to take the retinal images are not that expensive and so the real bottleneck is the time of skilled ophthalmologists to interpret the images).

Research blog post about this work:
https://research.googleblog.com/2016/11/deep-learning-for-detection-of-diabetic.html

The full JAMA article:
http://jamanetwork.com/journals/jama/fullarticle/2588763

A more general overview of our healthcare work: http://g.co/brain/healthcare

Edit: revised language to clarify that the model is better than the median ophthalmologist rather than more accurate than the consensus (which is the ground truth).
157
35
Benjamen Meiers's profile photoTau-Mu Yi's profile photoEric Brewer's profile photoTodd Hoff's profile photo
4 comments
 
It would be cool if the test could be integrated into a pair of glasses/VR. When I take the test my eyes are dilated and there's a person that takes the pics. A lot of people probably don't have access to this level of health care.
Add a comment...

Jeff Dean

Shared publicly  - 
 
It's the one year anniversary of our initial open source release of TensorFlow. It's been really exciting see the vibrant community of researchers, students, scientists, and engineers across a wide range of companies, universities and other organizations that have all come together to both use TensorFlow in their own work, as well as to work together to improve TensorFlow for everyone. TensorFlow had achieved a lot for a one-year-old, but it's still early in its life!
84
15
Stan Bileschi's profile photoRICARDO MARTINEZ's profile photo
2 comments
 
Thanks a lot +Jeff Dean 
Add a comment...

Jeff Dean

Shared publicly  - 
 
Some very nice work by my Google Research colleagues +George Toderici​, Damien Vincent, Nick Johnston, Sung Jin Hwang, David Minnen, Joel Shor and Michele Novell on using recurrent neural networks to do image compression significantly better than JPEG (at the same image quality). They released a TensorFlow implementation of this system, as well. Nice work!

Arxiv paper: https://arxiv.org/abs/1608.05148
TensorFlow code: https://github.com/tensorflow/models/tree/master/compression
95
19
Tom Hume's profile photoRICARDO MARTINEZ's profile photoPaolo Pignatelli's profile photo
3 comments
 
How closely is your group working with the language group? Since languages and images are both compressed representations internally, (the "mind"), that internal representation could/should be shared. Tensorflow is a fantastic common tool that could generate programming "objects" that would form a common core for "machine knowledge" that in turn could truly "compress" images ("how similar is this 'violin' to the archetypal violin I have stored?")... .
Add a comment...

Jeff Dean

Shared publicly  - 
 
The Show and Tell work from the Google Brain team (http://g.co/brain) a bit more than a year ago, along with concurrent work on the same topic by Berkeley, Toronto, Microsoft Research, and others, was some of the most exciting ML work in recent years. The notion that image models could not only classify an image ("train"), but could actually generate plausible whole sentences about an image purely from the raw pixels ("_A blue and yellow train is traveling down the tracks_") is pretty remarkable. Today, Chris Shallue from our group has released an open source implementation of an improved version of this model in TensorFlow. Happy captioning!

Edit: fix typo
93
7
Stewart Henderson's profile photo周老三's profile photoRICARDO MARTINEZ's profile photoKyylaa JByran's profile photo
4 comments
 
Well done!
Add a comment...
Story
Tagline
Google Senior Fellow
Introduction
I build large-scale computer systems.  I joined Google in 1999 and am currently a Google Fellow working in the Systems Infrastructure Group. While at Google, I have designed and implemented large portions of the company's advertising, crawling, indexing and query serving systems, along with various pieces of the distributed computing infrastructure that sits underneath most of Google's products. At various times, I've also worked on improving search quality, statistical machine translation, and various internal software development tools, and I've had significant involvement in the engineering hiring process.

Prior to joining Google, I was at DEC/Compaq's Western Research Laboratory, where I worked on profiling tools, microprocessor architecture, and information retrieval. Prior to graduate school, I worked at the World Health Organization's Global Programme on AIDS, developing software for statistical modeling and forecasting of the HIV/AIDS pandemic.

I earned a B.S. in computer science and economics (summa cum laude) from the University of Minnesota and received a Ph.D. and a M.S. in computer science from the University of Washington. I was elected to the National Academy of Engineering in 2009, which recognized my work on "the science and engineering of large-scale distributed computer systems."

Education
  • University of Washington
    Computer Science
  • University of Minnesota
    Computer Science and Economics
Basic Information
Gender
Male
Other names
Jeffrey Dean
Work
Occupation
Google Senior Fellow
Employment
  • Google
    Google Senior Fellow, present
Places
Map of the places this user has livedMap of the places this user has livedMap of the places this user has lived
Currently
Palo Alto
Previously
Seattle, Washington - Geneva, Switzerland - Minneapolis, Minnesota - Atlanta, Georgia - Mogadishu, Somalia - Honolulu, Hawaii - Little Rock, Arkansas - Arua, Uganda - Boston, Massachusetts
Links
Contributor to
Jeff Dean's +1's are the things they like, agree with, or want to recommend.
Improving Photo Search: A Step Across the Semantic Gap
googleresearch.blogspot.com

Posted by Chuck Rosenberg, Image Search Team Last month at Google I/O, we showed a major upgrade to the photos experience: you can now easil

The Tree of Life: YHGTBFKM: Ecological Society of America letter regardi...
phylogenomics.blogspot.com

The Tree of Life. Blog of Jonathan A. Eisen, evolutionary biologist, microbiologist and genomics researcher, Open Access and Open Science ad

Google Gives Back 2011
www.google.com

العربية, Български, Català, Čeština, Dansk, Deutsch, UK English, US English, Ελληνικά, Español - España, Español - Latinoamerica, suomi, Fil