As the AI landscape continues to evolve, a new version of the popular Caffe open source deep learning framework has been released. Caffe2 is backed by Facebook and features a wide array of partnerships to make it as flexible and scalable as possible. But is that enough to make Caffe2 a winner?
Health information technology and innovative uses of data are priorities for the Trump administration, which is committed to the free flow of electronic health-care data, HHS Secretary Tom Price said April 27.
The administration aims to give physicians incentives to use health IT and make it easier for electronic health systems to securely exchange information, also known as interoperability, Price said at Health Datapalooza, an annual health IT and data conference in Washington.
Recognizing and confirming the identities of people in the hot zone is just half of the problem; the ability to track them, monitor their sentiments, and predict the next flare-up of the epidemic is another. But we’re working on that, too.
Epidemico, a health analytics subsidiary Booz Allen acquired in 2015, has a product that provides real-time analysis and insights into population health. The company’s MedWatcher scrapes news feeds and administers a data processing algorithm to tag, filter, analyze, and visualize any health or threat alert, including those in traditional and social media, anywhere on the globe. The data is aggregated, analyzed, and then visualized on maps to project where an outbreak may move. That means we not only know who’s sick, but who’s most likely to get sick, based on data.
The Penn Wharton Budget Model (PWBM) is pleased to announce a $6.6 million commitment from the Laura and John Arnold Foundation (LJAF) to develop new, high-quality tools to evaluate the economic and budgetary impact of federal policy proposals. PWBM’s new tools will be powered by more sophisticated algorithms, granular data on corporations, and modeling of global transactions. The funding will also allow PWBM to develop forecasts regarding the impact of policy — on topics including taxes, international trade, Social Security, and immigration — over a longer time horizon than the standard ten-year budget window.
Anthony Levandowski, the head of Uber’s self-driving group, is stepping aside in face of trade-theft accusations from his former employer, Waymo.
In an email obtained by Business Insider, Levandowski said he will no longer be working on work related to Lidar, the specialized radar sensors that autonomous vehicles rely on to map their surroundings and to navigate on their own.
Levandowski will remain at Uber and will retain his other responsibilities overseeing things like operations and safety.
In the 100 days since President Trump took office, concerned Americans have downloaded over 2 million government datasets. Their goal? To back up information they believe is in danger of going dark: climate science research, discriminatory housing reports, gun violence statistics. But public data preservation isn’t just a job for citizens working in university libraries and on digital archiving Githubs. Now, Washington is getting in on the action.
On Thursday, Senators Gary Peters and Cory Gardner introduced a bipartisan bill that would make it much, much harder for any administration to disappear public data. If passed, the Preserving Government Data Act of 2017 would affect the availability of everything from census numbers to sea level rise.
Deep Sentinel, a home security startup, today announced that it had closed a $7.4 million Series A led by Shasta Ventures with participation from Bezos Expeditions, Lux Capital and UP2398. Founded by serial entrepreneur David Selinger, Deep Sentinel is betting that its emphasis on user experience can provide needed differentiation in the crowded space.
Over tea in the historic Palace Hotel in San Francisco, Selinger laid out the philosophy behind his latest company. Selinger believes that the nature of property crimes has changed. Without ever entering a home, burglars can cost homeowners thousands by simply nabbing an Amazon delivery resting on a front porch. This shift in behavior necessitates a shift in the services provided by home security companies.
“[Videogame] AI is still in the dark ages,” Epic CEO Tim Sweeney told a crowd gathered for GamesBeat’s 2017 industry summit.
As it is, videogame A.I. still relies on heuristics, with little in the way of “deep learning”–a term coined to describe recent advances in A.I. Deep learning is responsible for some of the gaudier headlines of late concerning A.I., including the first instance in which a computer mastered the notoriously complex game of Go
By now just about everyone has seen the video of Makoto Koike’s deep learning-powered cucumber sorter. Hobbyists around the world are hacking solutions to their problems using machine learning. The latest, Swedish beekeeper and inventor Björn Lagerman, is building BeeScanning with the help of a team of engineers and researchers. BeeScanning is an app that applies some clever computer vision to ordinary smartphone photos to alert beekeepers to the potentially dangerous presence of Varroa mites in their bee colonies.
[Sunny] Tang is playing alongside the rest of the Kronos Quartet, the iconic San Francisco string ensemble known for its unorthodox experimentation, and the AI is obeying orders from Trevor Paglen, the American artist who poses big questions about technology and surveillance through nearly any medium he can get his hands on. It’s all part of Sight Machine, a Paglen-orchestrated performance that explores the rise of computer vision.
Proceedings of the National Academy of Sciences, Christa Brelsford, José Lobo, Joe Hand, and Luís M. A. Bettencourt
from
Most nations worldwide have recently committed to solving their most severe challenges of sustainability by 2030, including eradicating extreme poverty and providing universal access to basic services. But how? Rapid urbanization is creating the conditions for widespread economic growth and human development, but its consequences are very uneven. We show how measures of sustainable development—identified by residents of poor neighborhoods—can be combined into a simple and intuitive index. Its analysis reveals that challenges of development are typically first addressed in large cities but that severe inequalities often result as patterns of spatially segregated rich and poor neighborhoods. A new systematic understanding of these processes is critical for devising policies that produce faster and more equitable universal sustainable development.
Jacqueline Poh, founding chief executive of Singapore’s Government Technology Agency (GovTech), doesn’t have long to chat. The day before we meet, she was swapping notes with UK government digital strategists. The day we meet, she’s flying to Amsterdam. … Poh lists a series of government-backed “enablers”, rolled out over the past decade, that have helped Singapore to achieve top status. They include a unique digital identity – SingPass – used by 3.3 million citizens for speedy access to government online services. Faster payment methods – including phone to phone – are under development. Crucially, Singapore’s 5.5 million population has access to the national high-speed optical fibre broadband network “which is cheaper than most countries and a lot faster”.
We’ll provide a dataset of roughly 10 million tweets from a variety of users. Your challenge is to build a algorithm that can predict entire messages with only the first letters of each word. Registrations close on May 15.
Tokyo, Japan DSAA2017 is the 4th IEEE International Conference on Data Science and Advanced Analytics. Special Session on Evolving Networks @ DSAA 2017 (EvoNets) is intended to attract researchers who are actively engaged in theoretical, technical and application oriented aspects of Evolving networks.
Deadline for special sessions’ papers submissions is May 25.
The National Science Foundation (NSF) has announced a call for Big Data Spoke proposals through the program solicitation “Big Data Regional Innovation Hubs: Establishing Spokes to Advance Big Data Applications (BD Spokes).” The South Big Data Hub is accepting requests for Letters of Collaboration from PIs in our region. Deadline for to request a Letter of Collaboration is June 19.
“At Metamarkets, we ingest more than 100 billion events per day, which are processed both realtime and batch. We store received events to our Kafka cluster and the stored events in Kafka are processed by both Samza and Spark for real-time stream processing and batch processing, respectively.”
Microsoft Research, Social Media Collective, Dan Greene
from
I recently published a pair of articles with Katie Shilton exploring how mobile app developers help each other learn what privacy means and how to build that abstract value into their software. Katie and I analyzed hundreds of forum conversations about privacy among iOS and Android developers, and compared the different development cultures and privacy values that arose around each platform.
We wanted to explain a little more about what led us to the development and release of Solarcast and some of the challenges we faced along the way. With the deployment of our Pointcast network we realized that a recurring problem with placing sensors was access to both power and internet for the devices we wanted to deploy in the field. We also found that complicated configuration requirements for devices required a Safecast team member to be physically involved with each installation which made deployment slow. We know from our bGeigie deployment that the easier we make the process of getting data from the device to us, the more data we get. With the intention of adding air quality sensors we wanted to rethink all of these issues and see where we would go if starting from square one.
The idea to have a totally wireless, solar powered, auto-configuring device that could be dropped anywhere and forgotten and it would just work was born. We wanted this to be very simple, and initially called the project Simplecast, though realized that the solar aspect of this really made it standout out and shifted to that instead. It started with a breadboard and an idea…
This project turns Raspberry Pi 3 into an intelligent gateway with deep learning running on it. No internet connection is required, everything is done locally on the Raspberry Pi 3 itself. At DT42, we believe that bringing deep learning to edge devices is the trend towards the future. It not only saves costs of data transmission and storage but also makes devices able to respond according to the events shown in the images or videos without connecting to the cloud.
Papers about deep learning ordered by task, date. Current state-of-the-art papers and papers useful for getting started are labelled. For each paper there is a permanent link, which is either to Arxiv.org or to a mirror of the original paper in this repository.