Scikit-learn-intelex

The release of **Intel® Extension for Scikit-learn 2023.2.1** introduces the following changes:

🚨 What's New
- sklearn 1.3 support fixes ([1](https://github.com/intel/scikit-learn-intelex/commit/9eb6afe124bc0c5daa7a4597391b68dc98c55f3b), [2](https://github.com/intel/scikit-learn-intelex/commit/40ccd1cb7108377270dbf09823f0784890c59ae0), [3](https://github.com/intel/scikit-learn-intelex/commit/c40bdda56282b4d7da36f8bc03828a729fc584be))
- [Model builders API update](https://github.com/intel/scikit-learn-intelex/commit/12b963ab487fca9d8291978a575b1c3f07c7e908)

The release of **Intel® Extension for Scikit-learn 2023.2.0** introduces the following changes:

:x: Deprecation Notice
- The compression functionality in the **Intel® oneDAL** library is deprecated. Starting with the 2024.0 release, oneDAL will not support the compression functionality
- The DAAL CPP SYCL Interfaces in the **Intel® oneDAL** library are deprecated. Starting with the 2024.0 release, oneDAL will not support the DAAL CPP SYCL Interfaces
- The Java* interfaces in the **Intel® oneDAL** library are marked as deprecated. The future releases of the oneDAL library may no longer include support for these Java* interfaces
- ABI compatibility is to be broken as part of the 2024.0 release of **Intel® oneDAL**. The library’s major version is to be incremented to two to enforce the relinking of existing applications
- macOS* support is deprecated for oneDAL. The 2023.x releases are the last to provide it


🛠️ Library Engineering

- CSR tables interface has been changed and moved from detail namespace

🚨 What's New

- Introduced new **Intel® oneDAL** functionality:
- Distributed KMeans++ algorithm
- Logistic Loss objective algorithm
- Introduced new functionality for **Intel® Extension for Scikit-learn**:
- NaN(missing values) support was added to Model Builders
- Improved performance for the following **Intel® Extension for Scikit-learn** algorithms:
- Model Builders performance has been improved up to 2x

The release of **Intel® Extension for Scikit-learn 2023.1.1** introduces the following changes:

🚨 What's New

- Splitting mode for Random Forest algorithm https://github.com/intel/scikit-learn-intelex/commit/4abff0df77475a7e7c3f3da135fdc9dc586f8f1e
- SPMD interface for Random Forest, kNN and PCA https://github.com/intel/scikit-learn-intelex/commit/e887d996a3da57ba75a1cbdaf26d3fea03ad882f, https://github.com/intel/scikit-learn-intelex/commit/97f0d470c5794cd63da18524f65053b89ab4177d, https://github.com/intel/scikit-learn-intelex/commit/2405630020da1563646aef768f12951c6f0d2dc3
- Native support for DPCTL https://github.com/intel/scikit-learn-intelex/commit/ca284998ffb98cb318a8ee6b6a9eb31a68b4d509
- Patching for sklearn's `LocalOutlierFactor` https://github.com/intel/scikit-learn-intelex/commit/7ae01712a75dbda9bf21c9ed444722ff7ac03dd2

The release of **Intel® Extension for Scikit-learn 2023.1** introduces the following changes:

📚Support Materials

- [Accelerating Barnes-Hut t-SNE Algorithm by Efficient Parallelization on Multi-Core CPUs](https://arxiv.org/abs/2212.11506)

🛠️ Library Engineering

- Reduced the size of **Intel® oneDAL** library by approximately ~30%

🚨 What's New

- Introduced new functionality for **Intel® Extension for Scikit-learn**:
- Enabled PCA, Linear Regression, Random Forest algorithms and SPMD policy as preview
- Scikit-learn 1.2 support
- [sklearn_is_patched()](https://github.com/intel/scikit-learn-intelex/blob/rls/2023.1.0-rls/examples/sklearnex/patch_sklearn.py#L24) function added to validate status of algorithms patching
- Improved performance for the following **Intel® Extension for Scikit-learn** algorithms:
- t-SNE for “Burnes-Hut” algorithm
- SVM algorithm for single row inference

❗ Known Issues

- In certain conditions DAAL SYCL interface might hang with L0 backend – please use oneDAL DPC interfaces instead. If older interfaces are required OpenCL backend can be used as workaround.

The release of **Intel® Extension for Scikit-learn 2023.0.1** introduces the following changes:

🚨 What's New

- Performance improvements for tSNE algorithm https://github.com/intel/scikit-learn-intelex/commit/5275ebac37c416ce110634c6cee7b56f872ed71b
- Fixes for balanced classes and number of iterations in SVM https://github.com/intel/scikit-learn-intelex/commit/14849ee7190f5701e4eab2ad923fce9125e904ff, https://github.com/intel/scikit-learn-intelex/commit/4872a8ea0afa22813f1a4446ef5fc0d608660283, https://github.com/intel/scikit-learn-intelex/commit/9d0a05b80c09aa3930107e2eca233cd7b872593c
- Fix for `gamma` parameter in KMeans https://github.com/intel/scikit-learn-intelex/commit/1dca20c3761197d082d548f27f698d6389e60fbe

The release of **Intel® Extension for Scikit-learn 2023.0** introduces the following changes:

🚨 What's New

- Introduced new **Intel® oneDAL** functionality:
- DPC++ interface for Linear Regression algorithm

❗ Known Issues

- **Intel® Extension for Scikit-learn** SVC.fit and KNN.fit do not support GPU
- Most Intel® Extension for Scikit-learn sycl examples fail when using GPU context
- Running the Random Forest algorithm with versions 2021.7.1 and 2023.0 of scikit-learn-intelex on the 2nd *Generation Intel®* Xeon® *Scalable* Processors, formerly *Cascade Lake* may result in an 'Illegal instruction' error.
- No workaround is currently available for this issue.
- Recommendation: Use an older version of scikit-learn-intelex until the issue is fixed in a future release.