[DOCS] Adds ML 6.3.0 release notes

Author: lcawl

docs/Versions.asciidoc

@@ -13,7 +13,9 @@ release-state can be: released | prerelease | unreleased
 :release-state:   unreleased
 
 :issue:           https://github.com/elastic/elasticsearch/issues/
+:ml-issue:        https://github.com/elastic/ml-cpp/issues/
 :pull:            https://github.com/elastic/elasticsearch/pull/
+:ml-pull:         https://github.com/elastic/ml-cpp/pull/
 
 :docker-repo:     docker.elastic.co/elasticsearch/elasticsearch
 :docker-image:    {docker-repo}:{version}

docs/reference/release-notes/6.3.asciidoc

@@ -21,16 +21,29 @@
 //[float]
 //=== New Features
 
-//[float]
-//=== Enhancements
+[float]
+=== Enhancements
+
+Machine Learning::
+* Synchronize long and short tests for periodicity {ml-pull}62[#62]
+* Improvements to trend modelling and periodicity testing for forecasting {ml-pull}7[#7] (issue: {ml-issue}5[#5])
 
 [float]
 === Bug Fixes
 
 Machine Learning:: 
 
-* Function description for population lat_long results should be lat_long instead of mean ({pull}81[#81])
-* By-fields should respect model_plot_config.terms ({pull}86[#86])
+* By-fields should respect model_plot_config.terms {ml-pull}86[#86] (issue: {ml-issue}30004[#30004])
+* Function description for population lat_long results should be lat_long instead of mean {ml-pull}81[#81] (issue: {ml-issue}80[#80])
+* Fix error causing us to overestimate effective history length {ml-pull}66[#66] (issue: {ml-issue}57[#57])
+* Clearing JSON memory allocators {ml-pull}30[#30] (issue: {ml-issue}26[#26])
+* Fix sparse data edge cases for periodicity testing {ml-pull}28[#28] (issue: {ml-issue}20[#20])
+* Impose an absolute cutoff on the minimum variance {ml-pull}8[#8] (issue: {ml-issue}488[#488])
+* Check accesses in bounds when clearing recycled models {ml-pull}79[#79] (issue: {ml-issue}76[#76])
+* Set forecast progress to 100% and status finished in the case of insufficient history (data) {ml-pull}44[#44]
+* Add control message to start background persistence {ml-pull}19[#19]
+* Fail start up if state is missing {ml-pull}4[#4]
+* Do not log incorrect model memory limit {ml-pull}3[#3]
 
 //[float]
 //=== Regressions