HN Explorer

11 comments

[ 4.8 ms ] story [ 32.8 ms ] thread
SubiculumCode
I do not understand how time series can be forecast without training on data from a relevant domain. Like, would these be able to predict EEG/fMRI timeseries?
BobSonOfBob
Would be good if the site had a couple of case studies
clickety_clack
If these worked we would have heard a lot more about them.
anshumankmr
How does next-token prediction work for time series data?
ChernovAndrei
There is no single answer, because there are multiple architectures for foundation time-series models, such as T5, decoder-only models, and state-space models (SSMs).

For Chronos-2 (the current state of the art in time-series modeling), the setup is almost identical to that of LLMs because it is based on the T5 architecture. The main difference is that, in time-series models, tokens correspond to subintervals in the real-valued (ℝ) space. You can check the details here: https://arxiv.org/pdf/2510.15821

kavalg
It looks like this is an SaaS with an open source client only right?
bvan
Isn’t this the ultimate black box? If a forecasting system is a black box, then you have no chance of understanding why its performance might deteriorate. Once that happens it essentially becomes a digital paper-weight.