Differentiable neural computer

A differentiable neural computer (DNC) combines the learning and pattern-recognition strengths of deep neural networks with the ability to retain information in complex data structures such as graphs in a computer memory. The memory can be retained indefinitely, while the DNC uses what it has learned to solve related problems.[1][2][3] DNC memory interactions are differentiable end-to-end, making it possible to optimize them efficiently using gradient descent.

DNCs were introduced by DeepMind in 2016, as an extension of neural turing machines, with attention mechanisms that control where the memory is active. The attention mechanisms are compared by the DeepMind authors to the functional capabilities of the mammalian hippocampus.[4]

They demonstrated, for example, how a DNC can be trained to navigate a variety of rapid transit systems, and then apply what it learned to get around on the London Underground. A neural network without memory would typically have to learn about each different transit system from scratch.

On graph traversal and sequence-processing tasks with supervised learning, DNCs performed better than alternatives such as long short-term memory or a neural turing machine.[4]

With a reinforcement learning approach to a block puzzle problem inspired by SHRDLU, DNC was trained via curriculum learning, and learned to make a plan. It performed better than a traditional recurrent neural network.[4]

So far, DNCs have only been demonstrated to handle relatively simple tasks, which could have been easily solved using conventional computer programming decades ago. But DNCs don't need to be programmed for each problem they are applied to, but can instead be trained. Furthermore, they can learn some aspects of symbolic reasoning and apply it to the use of working memory. Some experts see promise that they can be trained to perform complex, structured tasks[5][6] and address big-data applications that require some sort of rational reasoning, such as generating video commentaries or semantic text analysis.[1][7]

See also

References

  1. 1 2 Burgess, Matt. "DeepMind's AI learned to ride the London Underground using human-like reason and memory". WIRED UK. Retrieved 2016-10-19.
  2. ↑ "DeepMind AI 'Learns' to Navigate London Tube". PCMAG. Retrieved 2016-10-19.
  3. ↑ Mannes, John. "DeepMind's differentiable neural computer helps you navigate the subway with its memory". TechCrunch. Retrieved 2016-10-19.
  4. 1 2 3 James, Mike. "DeepMind's Differentiable Neural Network Thinks Deeply". www.i-programmer.info. Retrieved 2016-10-20.
  5. ↑ Graves, Alex; Wayne, Greg; Reynolds, Malcolm; Harley, Tim; Danihelka, Ivo; Grabska-Barwińska, Agnieszka; Colmenarejo, Sergio Gómez; Grefenstette, Edward; Ramalho, Tiago (2016-10-12). "Hybrid computing using a neural network with dynamic external memory". Nature. advance online publication. doi:10.1038/nature20101. ISSN 1476-4687.
  6. ↑ "Differentiable neural computers | DeepMind". DeepMind. Retrieved 2016-10-19.
  7. ↑ Jaeger, Herbert (2016-10-12). "Artificial intelligence: Deep neural reasoning". Nature. advance online publication. doi:10.1038/nature19477. ISSN 1476-4687.
This article is issued from Wikipedia - version of the 11/24/2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.