Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
altumintelligence.com
Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
1–10 of 10 posts
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#2Sounds like you need a neural network to help you write a better title for your post
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#3Someone who works in time series forecasting once told me that (to paraphrase) you know that if someone publishes a new method it's not all that good, because if it was they would have started a hedge fund instead.
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#4This title is like something out of Star Trek mumbo-jumbo. Surely a title made out of that summary at the top would make more sense.
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#5"This article does not exist" type headline.
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#6Someone who works in time series forecasting once told me that (to paraphrase) you know that if someone publishes a new method it's not all that good, because if it was they would have started a hedge fund instead.
It's not my area, but isn't forecasting financial time series stuff harder than forecasting for systems where the system components aren't actively trying to outsmart one another based on some of the same data? I.e. a method could be good enough to use in systems that aren't about the interaction of smart agents (e.g. what will my service's request volume look like next week), but fail to be useful in markets?
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#7Sounds like you need a neural network to help you write a better title for your post
Yeah absolutely, I want all my research papers to have snappy, attention grabbing headlines instead of clear descriptions of the subject they are addressing. How about:
"Leading temporal autoencoder expert reveals the No. 1 trick to deep sequence reconstruction youre NOT doing."
"The $$$ secret to temporal autoencoders that AI researchers are hiding from you."
"20 something invents impossible autoencoder technology in their garage that could change almost everything."
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#8Uh... no comparison with any other method?
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#9This looks like a standard seq2seq model.
Re: Multivariate Temporal Autoencoder for Predictive Reconstruction of Deep Sequence
#10Someone who works in time series forecasting once told me that (to paraphrase) you know that if someone publishes a new method it's not all that good, because if it was they would have started a hedge fund instead.
It's not my area, but isn't forecasting financial time series stuff harder than forecasting for systems where the system components aren't actively trying to outsmart one another based on some of the same data? I.e. a method could be good enough to use in systems that aren't about the interaction of smart agents (e.g. what will my service's request volume look like next week), but fail to be useful in markets?
Not my area either, but I worked on related problems. The impression I got is that it depends on what you are using the model for, i.e. are you trying to generate forecasts? prediction intervals? detecting anomalies? comparing to human forecasters? All models are wrong, but some are wrong in a useful way.
Also, a big caveat is that the output of those models, even basic ones, are supposed to come with prediction intervals, and not just as points. Libraries like statsmodels are massively guilty of making it hard to use the model like it's supposed to be used.