Jing Lin

The wave of artificial intelligence in China is not decreasing, and there will be more `artificial intelligence experts` in the future.

> Ensure functime works with Polars>=0.20. Also implement/benchmark Enum dtype for entity columns to save memory. Polars>=1.0 enhanced map_batches and map_elements to apply lambda function with machine learning forward computation....

Polars expressions were used in many databases, like bioinformatics sorted branches, vectors indexed sequences representation. Once capable with long sequences in memory, large-scale application will solve by in-memory and out-of-core...

My implementation of [levenberg_marquardt](https://github.com/linjing-lab/optimtool/blob/master/optimtool/unconstrain/nonlinear_least_square.py#L75) method, Python based algorithm.