Extending research into evolutionary #AutoML, we present an approach that evolves algorithms from scratch—using only basic mathematical operations—rediscovering fundamental ML techniques & showing the potential to discover novel algorithms. Read more ↓ goo.gle/3edHPBc
To support research into cloud computing and cluster scheduling, we are releasing a new 29-day trace of the Borg cluster management system, which captures scheduling decisions and resource usage for Google’s computational fleet. Learn more & get the data ↓goo.gle/3aKfUad
Our paper on learning data layouts in storage for big data analytics will appear at #SIGMOD2020. We propose a data structure called a qd-tree (built using deep RL) to layout data blocks in a workload-guided manner. Easy to integrate with DB as well! Paper: badrish.net/papers/qdtree-si…
Today, we present new research on offline #ReinforcementLearning to create a data-driven paradigm for real-world applications, along with the release of the DQN Replay Dataset to facilitate research in this exciting area. Learn more at goo.gle/2K5Viyo
The head of research @Waymo, Drago Anguelov, will be hosting a Reddit AMA this Thursday, April 16 at 11am PDT.
Ask them some tough questions → redd.it/g18xad