diff --git a/.github/workflows/IntegrationTest.yml b/.github/workflows/IntegrationTest.yml
index c0cab2ae..6a82a93e 100644
--- a/.github/workflows/IntegrationTest.yml
+++ b/.github/workflows/IntegrationTest.yml
@@ -14,7 +14,7 @@ jobs:
       fail-fast: false
       matrix:
         package:
-          - {user: TuringLang, repo: AdvancedPS.jl, ref: releases}
+          - {user: TuringLang, repo: AdvancedPS.jl, ref: master}
           - {user: TuringLang, repo: Turing.jl, ref: master}
 
     steps:
diff --git a/README.md b/README.md
index 661476c4..9175b1bb 100644
--- a/README.md
+++ b/README.md
@@ -87,16 +87,15 @@ a = copy(ttask)
 @show consume(ttask) # 3
 ```
 
-Note: The [Turing](https://github.com/TuringLang/Turing.jl)
-probabilistic programming language uses this task copying feature in
+Notes: 
+
+- The [Turing](https://github.com/TuringLang/Turing.jl) probabilistic programming 
+language uses this task copying feature in
 an efficient implementation of the [particle
 filtering](https://en.wikipedia.org/wiki/Particle_filter) sampling
 algorithm for arbitary order [Markov
 processes](https://en.wikipedia.org/wiki/Markov_model#Hidden_Markov_model).
 
-## Disclaimer
-
-- This feature is still experimental and should only be used with caution.
 - From v0.6.0, Libtask is implemented by recording all the computing
   to a tape and copying that tape. Before that version, it is based on
   a tricky hack on the Julia internals. You can check the commit