hrzn

11 ms·

1.
Isomorphic Labs, the secretive AI startup spun off from Google DeepMind(cnbc.com)
3 points by hrzn 1y ago|0 comments
2.
Isomorphic Labs, Google's A.I. Drug Business, Raises $600M(nytimes.com)
1 points by hrzn 1y ago|0 comments
3.
by hrzn 2y ago
A few years back I made a modest attempt at writing a concise yet readable sudoku solver in Python - in about 29 lines: https://github.com/hrzn/sudoku/blob/master/sudoku.py Could have been made shorter a
4.
by hrzn 2y ago
We likely overestimate AI's short-term impact, and there might even be a financial bubble about to pop. But I also think we underestimate the long-term impact. We're building absolutely amazing capabilities faster than many would
5.
Nano Prot GPT: Nano GPT on Protein Sequences(github.com)
1 points by hrzn 4y ago|0 comments
6.
by hrzn 4y ago
> Python had nothing comparable This is in part why we built Darts. Now I think we can say the situation is quite different. Darts offers many things offered by the R forecast package, and then some (for instance the ability to train ML
7.
by hrzn 4y ago
I would recommend Darts in Python [1]. It's easy to use (think fit()/predict()) and includes * Statistical models (ETS, (V)ARIMA(X), etc) * ML models (sklearn models, LGBM, etc) * Many recent deep learning models (N-BEATS, TFT, et
8.
Darts: User-Friendly Modern Machine Learning for Time Series(jmlr.org)
1 points by hrzn 4y ago|0 comments
9.
Darts: User-Friendly Modern Machine Learning for Time Series(jmlr.org)
5 points by hrzn 4y ago|0 comments
10.
Darts: User-Friendly Modern Machine Learning for Time Series(jmlr.org)
1 points by hrzn 4y ago|0 comments
11.
by hrzn 5y ago
Note that the actual EV of a Cauchy random variable is undefined...
12.
by hrzn 5y ago
You can install darts with conda using conda install u8darts-all We are working on improving our support of conda based installs.
13.
by hrzn 5y ago
You can try installing it using pip install 'u8darts[torch]' This will give you everything besides Prophet and pmdarima.
14.
by hrzn 5y ago
I would say that compared to Greykite, Darts really attempts to unify a wide variety of forecasting models under a common simple and user-friendly API. There are many differences, but for instance, AFAIK there's no deep learning model
15.
by hrzn 5y ago
In short: yes, at least anomaly detection. That's one of the main things on our to-do list.
16.
by hrzn 5y ago
Oh, nice, I didn't know that. I've edited my reply, thanks!
17.
by hrzn 5y ago
We do have ensemble models. Something like model = NaiveEnsembleModel([model1, model2, ...]) model.fit(my_series) prediction = model.predict() Will return an average prediction. Look at RegressionEnsembleModel for an ensemb
18.
by hrzn 5y ago
Thanks for the feedback, I absolutely agree about the need for easy-to-use tools for dealing with time series. This is exactly the motivation that prompted us to work on Darts initially. I like your suggestions of adding comparison to the f
19.
by hrzn 5y ago
> Or is it just turning every column in my pandas dataframe into a series to pass into the covariates array? Basically if you have a multivariate series represented as a pandas dataframe with several columns, the way to go is to create y
20.
by hrzn 5y ago
You're welcome, glad you like it and thanks for the feedback :)
21.
by hrzn 5y ago
In some cases Darts is wrapping around existing models (like Prophet, or statsmodels-based models for instance); in other cases we wrote our own implementations, so it's really a mix.
22.
by hrzn 5y ago
Hi! I'm one core developer (and creator) of the library. Thanks for all the comments. I just wanted to highlight a couple of things that we think are quite cool about Darts: * It makes using all sorts of forecasting models (from ARIMA
23.
VeGANs: Various existing GANs (and other generative models) in PyTorch(github.com)
2 points by hrzn 5y ago|0 comments
24.
by hrzn 5y ago
This is supported but only by neural-nets models, which are fit using SGD, hence naturally not requiring the whole dataset in memory. Other models like ARIMA do need the full series loaded in memory. The models that work on multiple time se
25.
by hrzn 5y ago
Both - some models are wrapped (like ARIMA & ETS around statsmodels, Prophet around fbprophet) and we write others ourselves (RNNs, TCNs, N-Beats, ...). Basically we take a pragmatic approach here, we do whatever is best to use a given
26.
by hrzn 5y ago
For training forecasting models on multiple time series (and potentially large datasets), you can take a look at Darts [1] and the blog post [2]. [1]: https://github.com/unit8co/darts/ [2]: https://medi
27.
by hrzn 5y ago
For those interested in time series library, we are developing Darts [1], which focuses on making it easy & straightforward to build and use forecasting models. Out of the box it contains traditional models (such as ARIMA) as well as re
28.
by hrzn 5y ago
A lot of these issues also have to do with the publishing system. There is no incentive in publishing stuff that don't work even though they were reasonable things to try. This indirectly pushes a lot of bad/flawed/incomplete
29.
VeGANs: A Python library to train various existing GANs(github.com)
2 points by hrzn 5y ago|0 comments
30.
by hrzn 6y ago
Even if these "noticeable" sequences make up only a tiny fraction of all possible sequences, the probability that they show up at least once in a while among all of the world lotteries is actually high. In general the probability

More ›