Hi,
I’m starting to engineer a solution for the following. A few architectures come to mind, but I’m not sure what would be the best. If this forum doesn’t know, I’m not sure who will
Problem is: I’m receiving a stream of {ID, data}
where ID
is a user ID, and data
is a list of a few hundreds floats (e.g., data.shape = (200,)
). I would like to model the data
sequence of each user over time so that I can spot “anomalies”.
Ideas. The first ideas that pop up are:
- “classic” multilayer LSTM, interleaved with various dropouts and a single FC layer at the end for multi-class prediction
- Using LSTM autoencoders
- The famous “taxi prediction” (https://arxiv.org/pdf/1508.00021.pdf) often discussed here.
- Recurrent highway networks?
- …
Obviously, I’d handle the online learning by:
- predict
ID
givendata
, and look at the logprob of the “correct”ID
- then, train/update the model by providing
ID
anddata
on a batch of the last few{ID, data}
points.
Am I missing any recent superfancy arch?
Thanks!
ps: after part1v1, part2v1 i’m now following part1v2. It’s never enough!