Original addressDeepMind moves to TensorFlowDeepMind Steering TensorFlow
Friday, April.Posted by Koray Kavukcuoglu, scientist, Google DeepMind
At deepmind, we conduct state-of-the-art research on a wide range of algorithms, from deep learning and Rei Nforcement Learning to systems neuroscience, towards the goal of Building artificial General Intelligence. A key factor in facilitating rapid progress are the software environment used for the. For nearly four years, the Open Source torch7 machine Learning Library have served as our primary-PLATFO RM, combining excellent flexibility with very fast runtime execution, enabling rapid prototyping. Our team have been proud to contribute to the open source project in capacities ranging from occasional bugs fixes to being Core maintainers of several crucial components. On the DeepMind project, towards the goal of building artificial intelligence, we have created art-level research on deep learning, enhancement of learning to the nervous system, a series of algorithms. An important factor contributing to the rapid progress of the project is the software environment used in the study. For the last 4 years, open source software Torch7, the machine learning Library, has been our primary research platform, combining the perfect flexibility and very fast runtime execution to ensure rapid modeling. Our team is proud to have contributed to the open source project, which has evolved from the occasional bug fix to being the core maintainer of several key modules. With Google ' s recent open source release of tensorflow, we INITiated a project to test it suitability for our environment. The last six months, we had re-implemented more than a dozen different projects on TensorFlow to develop a deeper UN Derstanding of its potential use cases and the tradeoffs for the. Today We is excited to announce that DeepMind would start using TensorFlow for all our future. We believe that TensorFlow would enable us to execute our ambitious the goals at much larger scale and an even faster Pace, providing us with a unique opportunity to further accelerate US-programme. For Google's recent open source TensorFlow, we initialized a Engineering to test its adaptability to our research environment. Over the last six months, we have re-implemented more than a dozen different projects on tensorflow to improve our understanding of its potential use cases and research intentions. Today we are very excited to announce that DeepMind will begin to use TensorFlow in all of our future studies. We believe that TensorFlow will ensure that we can carry out our ambitious research goals on a larger scale and at a faster pace, and provide a unique opportunity for us to accelerate our research process in the future.
As one of the core contributors of Torch7, I has had the pleasure of working closely with an excellent community of devel Opers and researchers, and it had been amazing to see all the great work that have been built on top of the platform and th E impact this have had on the field. Torch7 is currently being used by Facebook, Twitter, and many start-ups and academic labs as well as DeepMind, and I ' M pro UD of the significant contribution it has made to a large community in both and industry. Our transition to TensorFlow represents a new chapter, and I feel very excited about the prospect of DeepMind contributing Heavily to another great open source machine learning platform, everyone can use to advance the State-of-the-art. As to As a core contributor to Rch7, I am honored to be working closely with a community of very good developers and researchers looking at the great work that has been built on this platform, and the impact of these efforts on the field, all of which are very surprising. Torch7 has now been used by Facebook, Twitter, and many startups and college labs, just like DeepMind, and I'm very proud of its outstanding contributions in research and industry. We turn to TensorFlow to show a new chapter, and I am very excited and look forward to DeepMind will make an important contribution to another great open source machine learning platform, so that everyone can advance this artThe product of the operation grade.
TensorFlow Blog translation--deepmind turn TensorFlow
