Amazon Research Awards Honor Outstanding Academic Projects in Artificial Intelligence
May 17 2018 - 9:00AM
Business Wire
Broad range of applications demonstrates the
diversity of research for AI technology solutions
49 scientific groups from 28 institutions will
be awarded with a total of 3.7 million US-Dollars
The Amazon Research Awards (ARA) are granted in
11 categories to foster innovation and collaboration with major
research institutions
(NASDAQ: AMZN) – Amazon today announced the winners of the
Amazon Research Awards (ARA) in 2017, a program designed to support
independent external research in areas relevant for Amazon
customers. The funded research is in the fields of computer science
and related topics including machine learning, computer vision,
robotics, and natural language processing. In the third year of
ARA, more than 800 research groups, universities and scientific
institutions from North America and Europe took part in the open
call for proposals in fall 2017. Out of all these applicants, 49
projects will be supported with the Amazon Research Awards with up
to $80,000 per project.
“With the Amazon Research Awards, we aim to deepen the existing
ties of Amazon Research teams across the world to research
institutions and to fund research themes that will define the
future of artificial intelligence,” said Ralf Herbrich, Director of
Machine Learning at Amazon. “We are proud of the large number of
highly qualified applications, which demonstrate the breadth and
depth of the AI research field. With Amazon Research Awards, we
want scientists to further their research in foundational areas
that enable innovations for our customers and make these results
available on an open source basis. We are especially proud of the
fact that the applicants represent a very wide range of schools,
from smaller schools like Rice, to large institutions like the
University of Washington and Inria.”
These are some of the projects that are funded this year:
- Computer Vision: Cordelia
Schmid, Research Director at Inria, the French National Institute
for computer science and applied mathematics, and her team conduct
research in computer vision, and more particularly the automatic
interpretation of digital images and videos. The funded project
will construct the first realistic dataset for 3D pose evaluation
of humans in action. To do so, the Inria researchers will capture
3D models of humans in real-life situations while performing
actions and manipulating objects using a multi-camera platform.
They will then render these models in synthetic scenes simulating a
moving camera. The resulting data are more realistic as they
capture real hair and clothing deformations, while showing real
motion blur, truncations by the image boundaries and occlusions by
manipulated objects and elements of the scene.
- Machine Learning: Thorsten
Joachims, Professor in the Department of Computer Science and in
the Department of Information Science at Cornell University in
Ithaca, New York, and his team will develop new machine learning
algorithms that can learn from partial information, such as user
feedback in log data. Logged user interactions are one of the most
ubiquitous forms of data available, as they can be recorded from a
variety of systems like search engines or recommender systems at
little cost. The interaction logs of such systems (e.g., a
personalized newspaper) typically contain a record of the input to
the system (e.g., features describing the user), the action taken
by the system (e.g., presented ranking of news articles) and the
feedback (e.g., clicks on the articles). When a user clicks on a
search result, it often does not mean that the result is good on
some absolute level, just that it is better than the higher ranked
results. The proposal eliminates the need to randomize while
collecting the feedback, making the learning systems more
applicable to cases like search, advertising, or recommendations
that may use deterministic logging.
- Robotics: With the ARA grant,
the team of Sven Koenig, professor in computer science at the
University of Southern California, will study how a high number of
robots find their paths efficiently and effectively in highly
filled spaces for warehouse automation. The robots must be able to
make good decisions in complex situations that involve a
substantial degree of uncertainty, yet find solutions in a timely
manner despite a large number of potential contingencies. Congested
spaces exist in Amazon fulfillment centers, for example, in front
of the picking stations since every robot has to cross this region,
whether it delivers a shelf to a picking station or returns it to
the warehouse. USC’s research intends to provide a strong
foundation for building resilient algorithms for such robot systems
to end collision-free paths for all robots from their respective
start vertices to their respective goal vertices.
The other ARA categories are General AI, Knowledge Management
and Data Quality, Machine Translation, Personalization, Search and
Information Retrieval, Security, Privacy and Abuse Prevention, and
Speech.
“On behalf of my team, I’d like to thank Amazon for granting us
the ARA already for the second time,” said Cordelia Schmid,
Research Director at Inria, the French National Institute for
computer science and applied mathematics. “The grant will fund work
on real-world challenges we are researching on at Inria. I value
these interactions with companies like Amazon because they provide
feedback from the industry and help assess the applicability of our
research and information on industry problems.”
The output of the funded projects will be made publicly
available both in the form of academic publications and open source
code contributions. ARA also facilitates training for students and
temporary research positions for faculty. These include graduate
student and post-doc internships offered by the industry partner,
and visiting researcher arrangements when the scientist temporarily
leaves their home university to work as an industry researcher.
The full list of winners with details on their projects is
available on the ARA website:
https://ara.amazon-ml.com/recipients/#2017.
About Amazon
Amazon is guided by four principles: customer obsession rather
than competitor focus, passion for invention, commitment to
operational excellence, and long-term thinking. Customer reviews,
1-Click shopping, personalized recommendations, Prime, Fulfillment
by Amazon, AWS, Kindle Direct Publishing, Kindle, Fire tablets,
Fire TV, Amazon Echo, and Alexa are some of the products and
services pioneered by Amazon. For more information, visit
www.amazon.com/about and follow @AmazonNews.
View source
version on businesswire.com: https://www.businesswire.com/news/home/20180517005315/en/
Amazon.com, Inc.Media
HotlineAmazon-pr@amazon.comwww.amazon.com/pr
Amazon.com (NASDAQ:AMZN)
Historical Stock Chart
From Mar 2024 to Apr 2024
Amazon.com (NASDAQ:AMZN)
Historical Stock Chart
From Apr 2023 to Apr 2024