FLAML VS darts

I really like the simplicity of this framework, and they hit on a lot of common problems found in other agent-based frameworks. Most intrigued by the RAG improvements.

Seems like Microsoft was frustrated with the pace of movement in this space and the shitty results of agents (which admittedly kept my interest turned away from agents for the last few months). I'm interested again because it makes practical sense, and from looking at the example notebooks, seems fairly easy to integrate into existing applications.

Maybe this is the 'low code' approach that might actually work, and bridge together engineering and non-engineering resources.

This example was what caught my eye: https://github.com/microsoft/FLAML/blob/main/notebook/autoge...

4. FLAML

Improved AutoML - Again, with collaboration from Microsoft Research, we used FLAML to update our existing AutoML solutions. What does this mean for you? You're using the latest techniques but all you need is a problem to solve and some data to get started.

For a complete tutorial, navigate to this Jupyter Notebook: https://github.com/microsoft/FLAML/blob/main/notebook/flaml_automl.ipynb

Microsoft Researchers have developed FLAML (Fast Lightweight AutoML) which can now utilize Ray Tune for distributed hyperparameter tuning to scale up FLAML’s resource-efficient & seamlessly parallelizable algorithms across a cluster.

Looks nice but I wonder if this is practical for non-tiny problems. The papers are a bit hard to follow but it looks like training is restarted with every new architecture choice. As for the library itself, the only large neural net example is a finetune of an NLP model that only searches over ADAM's optimizer params - which could be useful but it's a stretch to call that AutoML.

Don't use an LSTM. Get up to date with SoTA methods and read the papers in the field. LSTMs are not the way forward. Read the papers I suggested. It would be very useful to come to grips with both the Time Series Repository (https://github.com/thuml/Time-Series-Library) and Darts (https://github.com/unit8co/darts) as these are widely used for research and in industry.

As others have pointed out, Prophet is not a particularly good model for forecasting, and has been superseded by a multitude of other models. If you want to do time series forecasting, I'd recommend using Darts: https://github.com/unit8co/darts. Darts implements a wide range of models and is fairly easy to use.

The problem with time series forecasting in general is that they make a lot of assumptions on the shape of your data, and you'll find you're spending a lot of time figuring out mutating your data. For example, they expect that your data comes at a very regular interval. This is fine if it's, say, the data from a weather station. This doesn't work well in clinical settings (imagine a patient admitted into the ER -- there is a burst of data, followed by no data).

That said, there's some interesting stuff out there that I've been experimenting with that seems to be more tolerant of irregular time series and can be quite useful. If you're interested in exchanging ideas, drop me a line (email in my profile).

3. darts

Looking forward to checking this out! How does this compare with darts[1]?

[1] https://unit8co.github.io/darts/

Darts is an open-source Python library by Unit8 for easy handling, pre-processing, and forecasting of time series. It contains an array of models, from standard statistical models such as ARIMA to deep neural networks. https://unit8co.github.io/darts/

active support

The new Forecast menu, featuring the open-source Darts Time Series library, offers script-friendly functionality. It's also easy to use. Don't have any data to load yet? Enter through the Stocks or Crypto menus.

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, etc)

* Seamlessly works on multi-dimensional series

* Models can be trained on multiple series

* Many models offer rich support for probabilistic forecasts

* Model evaluation is easy: Darts has many metrics, offers backtest etc

* Deep learning scales to large datasets, using GPUs, TPUs, etc

* There's even now an explainability module for some of the models - showing you what matters for computing the forecasts

* (coming soon): an anomaly detection module :)

* (also, it even include FB Prophet if you really want to use it)

Warning: I'm probably biased because I'm Darts creator.

[1] https://github.com/unit8co/darts