The goal, as far as I understand, is to escape from the current simple execution of neural networks and have them work out programs by themselves. The ultimate product of this line of research will be a Gödel machine. It is actively carried out by Google and Facebook employees, among others.
TODO: Important development: neural stored-program memory (Le et al. 2020)[1] approximates universal Turing machine.
- Neural Turing Machine (NTM) (Graves et al., 2014)[2]
- Stack-Augmented recurrent neural networks (Joulin & Mikolov, 2015[3]; Grefenstette et al., 2015[4])
- Grid-LSTM (Kalchbrenner et al., 2015)[5]
- Neural Random-access machines: Kurach et al. (2015)[6]
- Zaremba et al. (2015)[7]
- Neural programmer: Neelakantan et al. (2015)[8]
- Neural GPUs: Kaiser & Sutskever (2015)[9]
- Zhang et al. (2015)[10]
- Neural Programmer-Interpreters: Reed & de Freitas (2015)[11]
References[]
- ↑ Le, H., Tran, T., & Venkatesh, S. (2020). Neural Stored-program Memory. In ICLR 2020.
- ↑ Graves, Alex, Wayne, Greg, and Danihelka, Ivo. Neural turing machines. arXiv preprint arXiv:1410.5401, 2014.
- ↑ Joulin, Armand and Mikolov, Tomas. Inferring algorithmic patterns with stack-augmented recurrent nets. arXiv preprint arXiv:1503.01007, 2015.
- ↑ Joulin, Armand and Mikolov, Tomas. Inferring algorithmic patterns with stack-augmented recurrent nets. arXiv preprint arXiv:1503.01007, 2015.
- ↑ Kalchbrenner, Nal, Danihelka, Ivo, and Graves, Alex. Grid long short-term memory. arXiv preprint arXiv:1507.01526, 2015
- ↑ Karol Kurach, Marcin Andrychowicz, Ilya Sutskever. 2015. NEURAL RANDOM-ACCESS MACHINES Arxiv preprint
- ↑ LEARNING SIMPLE ALGORITHMS FROM EXAMPLES. 2015. Wojciech Zaremba, Tomas Mikolov, Armand Joulin, Rob Fergus
- ↑ NEURAL PROGRAMMER: INDUCING LATENT PROGRAMS WITH GRADIENT DESCENT. Arvind Neelakantan, Quoc V. Le, Ilya Sutskever
- ↑ Łukasz Kaiser & Ilya Sutskever. 2015. NEURAL GPUS LEARN ALGORITHMS
- ↑ Wei Zhang, Yang Yu, Bowen Zhou. 2015. Structured Memory for Neural Turing Machines Arxiv
- ↑ Scott Reed and Nando de Freitas. 2015. Neural Programmer-Interpreters