Google DeepMind is pinning artificial intelligent agents against each other to see how they cooperate. To research this, the team is using a game known as Prisoner’s Dilemma to test its willingness to compete and cooperate.
According to the research, at times the agents worked peacefully together, but were less cooperative in complex situations or when it benefited them.
“The new notion of sequential social dilemmas allows us to model how rational agents interact, and arrive at more or less cooperative behaviors depending on the nature of the environment and the agents’ cognitive capacity,” the DeepMind team wrote in a blog. “The research may enable us to better understand and control the behaviour of complex multi-agent systems such as the economy, traffic, and environmental challenges.”
More information is available here.
Apache Spot becomes a modern open-source solution for cybersecurity
Cloudera and Intel have announced their open-source solution Apache Spot can now provide a comprehensive cybersecurity hub for forensic, telemetry and contextual data. Apache Spot is an open common data model designed to provide visibility into cybersecurity threats, and it was launched in September 2016.
Other features include log management, compliance, reporting and advanced NetFlow analytics.
“Cobbling together cybersecurity applications without a common architecture creates data silos that limit the effectiveness of advanced analytics that can be used for faster threat detection, investigation and remediation,” said Tom Reilly, CEO of Cloudera. “The Apache Spot project is breaking down data constraints to open up new analytics capabilities, enabling organizations across industries to reduce risk.”
Yahoo open-sources TensorFlow on Spark
Yahoo has found that deploying distributed deep learning would allow it to gain more insight from massive amounts of data. In order to address some of the limitations of TensorFlow programs, Yahoo decided to open-source TensorFlow on Spark (TFoS), its latest framework for distributed deep learning on Big Data clusters.
The framework enables distributed TensorFlow execution on Spark and Hadoop clusters, and it is designed to work along with Spark SQL, MLlib and other Spark libraries. Other TensorFlow programs can be modified to work with TFoS, and developers at Yahoo who use TensorFlow have already migrated TensorFlow programs for execution with TFoS.
More information on this project can be found here.
3T MongoChef becomes Studio 3T
MongoDB’s tool 3T MongoChef has been renamed to Studio 3T by 3T Software Labs, which has been developing the MongoChef IDE since 2013.
“The old name didn’t do our development environment justice,” said 3T cofounder Tomasz Naumowicz. “At first we set out, in lean startup style, to make a quick-and-dirty GUI. But over three years we have developed it into something much more useful than just a GUI.”
Naumowicz said that the latest version (4.5) gives developers more of a studio feel to the IDE, and it lets them do things like aggregation query building, data in-place editing, and SQL export.
Developers can check out Studio 3T here.