The Apache MADlib team is pleased to announce the immediate
availability of the 1.17.0 release.
The main goals of this release are:
New features
- DL: Add optional params to madlib_keras_fit_multiple_model
(MADLIB-1397)
- DL: Fit and evaluate changes for asymmetric cluster config
(MADLIB-1393)
- DL: Make param search fit() function work with existing evaluate and
predict (MADLIB-1387)
- DL: ParamSearch: Add utility function for generating model selection
table (MADLIB-1375)
- DL: Predict changes for asymmetric cluster config (MADLIB-1394)
- DL: Preprocessor should evenly distribute data on an arbitrary number
of segments (MADLIB-1378)
- DL: Preprocessor support for asymmetric segment distribution
(MADLIB-1392)
- DL: Remove model_arch_table column from the output of
load_model_selection_table (MADLIB-1381)
- DL: Support DL predict without training on MADlib (MADLIB-1359)
- DL: Transfer learning for multi-model (MADLIB-1389)
- Kmeans: Add simple silhouette score for every point (MADLIB-1382)
- Kmeans: Select number of centroids in k-means (MADLIB-1380)
- PostgreSQL 12 support (MADLIB-1391)
Improvements:
- Assoc rules: Add option to set number of posterior in association
rules (MADLIB-1327)
- Correlation: Improve correlation and covariance memory usage with
large number of groups (MADLIB-1301)
- DL: helper function for asymmetric cluster config (MADLIB-1390)
- DL: Mini-batch preprocessor for images - performance issue
(MADLIB-1342)
- DL: Modify warm start logic for DL to handle case of missing weight
(MADLIB-1400)
- DL: Param search for multiple models on MPP architecture (MADLIB-1386)
- DL: performance improvements to fit transition function (MADLIB-1418)
- Docs: Enhance Installation Guides (MADLIB-1399)
- Graph: SSSP should not show vertices in output table that are
unreachable (MADLIB-1415)
- Knn - add zero check and output distance array (MADLIB-1370)
- LDA: Add stopping criteria on perplexity to LDA (MADLIB-1351)
- Summary: Last optional param in summary errors when NULL (MADLIB-1413)
- Summary: Summary function has dups for MFV for approximate results
(MADLIB-1412)
- SVM: Change default num_components for SVM to max(100,
2*num_features) (MADLIB-1384)
All release changes can be found here:
https://cwiki.apache.org/confluence/display/MADLIB/MADlib+1.17.0
You can download the source release and convenience binary packages
from Apache MADlib's download page here:
http://madlib.apache.org/download.html
Alternatively, you can download through an ASF mirror near you:
https://www.apache.org/dyn/closer.lua/madlib/1.17.0
----
Apache MADlib is an open-source library for scalable in-database
analytics. It provides data-parallel implementations of mathematical,
statistical and machine learning methods for structured and
unstructured data.
The MADlib mission: to foster widespread development of scalable
analytic skills, by harnessing efforts from commercial practice,
academic research, and open-source development.
We welcome your help and feedback. For more information on how to
report problems, and to get involved, visit the project website at
https://madlib.apache.org
----
Thank you, everyone, who contributed to the 1.17.0 release. We look
forward to continued community participation for the next release!
Regards,
Orhan Kislal
