The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process — mentioned by duvenaud

Fri, 14 Dec 2018 03:04:05 GMT

The closest thing to a continuous-time RNN that exists are Neural Hawkes Processes [https://arxiv.org/abs/1612.09328]. They use a different model where observing the system necessarily changes its state, which is natural in some settings but doesn't fit in others. On the other hand, their model scales today and ours doesn't :)

Paper and discussion history. A mention is not an endorsement.

The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process — mentioned by stochastic_monk

Mon, 13 Nov 2017 00:24:50 GMT

So the author quotes Judea Pearl, whose causal calculus led to PGMs playing a critical in computer science and science as a whole, and somehow uses that to support the idea that probability theory is useless? That is absurd and out of context. Take a look at Kaldi or any state of the art platform for sound to speech and NLP analysis, or take a look at deep belief machines/networks. They all roughly extend probabilistic graphical models with deep architectures and some nonlinearity to do what couldn't be done before. (See also the Neural Hawkes Process [https://arxiv.org/abs/1612.09328].) The…

Paper and discussion history. A mention is not an endorsement.

The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process · HackerPapers discussions

The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process — mentioned by duvenaud

The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process — mentioned by stochastic_monk