Skip to content

Commit

Permalink
Add the SkinnerMT and SkinnerDB papers
Browse files Browse the repository at this point in the history
  • Loading branch information
@paul356
paul356 committed Mar 12, 2024
1 parent ecd5715 commit c127527
Show file tree
Hide file tree
Showing 2 changed files with 13 additions and 9 deletions.
14 changes: 8 additions & 6 deletions _org/2024-03-01-march-papers.org
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,10 +8,12 @@ nav_order: {{ page.date }}
---
#+END_EXPORT

|----------------------------------------------------------+--------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------+---------------------------------------------|
| Title | Authors | Synthesis | Publisher | Keywords |
|----------------------------------------------------------+--------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------+---------------------------------------------|
| OpenTuner: An Extensible Framework for ProgramAutotuning | Jason Ansel, Shoaib Kamil, Kalyan Veeramachaneni, Una-May O'Reilly, Saman Amarasinghe, etc | This paper present the configuration learning framework OpenTuner which uses a meta learning method called *multi-arm bandit with sliding window and area under the curve credit assignment* to ensemble serveral simple search methods. This software is provided as a python module publicly available. It can be used in configuration search problems in wide areas. | PACT 2014 | Multi-armed Bandit Problem, Area Under the Curve Credit Assignment, OpenTunner |
| | | | | |
|----------------------------------------------------------+--------------------------------------------------------------------------------------------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-----------+---------------------------------------------|
|-----------------------------------------------------------------------------------------------------------+--------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------------+--------------------------------------------------------------------------------|
| Title | Authors | Synthesis | Publisher | Keywords |
|-----------------------------------------------------------------------------------------------------------+--------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------------+--------------------------------------------------------------------------------|
| OpenTuner: An Extensible Framework for ProgramAutotuning | Jason Ansel, Shoaib Kamil, Kalyan Veeramachaneni, Una-May O'Reilly, Saman Amarasinghe, etc | This paper present the configuration learning framework OpenTuner which uses a meta learning method called *multi-arm bandit with sliding window and area under the curve credit assignment* to ensemble serveral simple search methods. This software is provided as a python module publicly available. It can be used in configuration search problems in wide areas. | PACT 2014 | Multi-armed Bandit Problem, Area Under the Curve Credit Assignment, OpenTunner |
| SkinnerMT: Parallelizing for Efficiency and Robustness in Adaptive Qery Processing on Multicore Platforms | Ziyun Wei, Immanuel Trummer | This paper present SkinnerMT which is based on SkinnerDB. SkinnerMT uses the same reinforcement learning approach as in SkinnerDB to search for the optimal join order. This paper combines parallel exploration of join orders with parallel data processing. | VLDB 2022 | Parallel Query Execution, Adaptive Data Partitioning |
| SkinnerDB: Regret-bounded Query Evaluation via Reinforcement Learning | Immanuel Trummer, Junxiong Wang, Ziyun Wei, Deepak Maram, Samuel Moseley, Saehan Jo, etc | This paper present SkinnerDB. SkinnerDB employs the UCT algorithm (a reinforcement learning algorithm) to find the optimal join order for a query. In order to reduce the overhead to store intermediate join state this paper uses a Multi-way Join Algorithm which generates at most one tuple as the intermediate result. SkinnerDB quickly switches between different join orders to find the optimal join order for the query on the go. This system doesn't use prior experiences and fully depend on the feedback of DBMS to drive the learning process. This learning process is repeated for each query execution. | SIGMOD 2021 | Join Order Learning, Query Optimization, Reinforcement Learning |
| | | | | |
|-----------------------------------------------------------------------------------------------------------+--------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+-------------+--------------------------------------------------------------------------------|

8 changes: 5 additions & 3 deletions _posts/2024-03-01-march-papers.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,8 @@ tags: [Query Optimization, Learned Query Optimizer, Machine Learning for Databas
nav_order: {{ page.date }}
---

| Title | Authors | Synthesis | Publisher | Keywords |
| OpenTuner: An Extensible Framework for ProgramAutotuning | Jason Ansel, Shoaib Kamil, Kalyan Veeramachaneni, Una-May O'Reilly, Saman Amarasinghe, etc | This paper present the configuration learning framework OpenTuner which uses a meta learning method called **multi-arm bandit with sliding window and area under the curve credit assignment** to ensemble serveral simple search methods. This software is provided as a python module publicly available. It can be used in configuration search problems in wide areas. | PACT 2014 | Multi-armed Bandit Problem, Area Under the Curve Credit Assignment, OpenTunner |
| | | | | |
| Title | Authors | Synthesis | Publisher | Keywords |
| OpenTuner: An Extensible Framework for ProgramAutotuning | Jason Ansel, Shoaib Kamil, Kalyan Veeramachaneni, Una-May O'Reilly, Saman Amarasinghe, etc | This paper present the configuration learning framework OpenTuner which uses a meta learning method called **multi-arm bandit with sliding window and area under the curve credit assignment** to ensemble serveral simple search methods. This software is provided as a python module publicly available. It can be used in configuration search problems in wide areas. | PACT 2014 | Multi-armed Bandit Problem, Area Under the Curve Credit Assignment, OpenTunner |
| SkinnerMT: Parallelizing for Efficiency and Robustness in Adaptive Qery Processing on Multicore Platforms | Ziyun Wei, Immanuel Trummer | This paper present SkinnerMT which is based on SkinnerDB. SkinnerMT uses the same reinforcement learning approach as in SkinnerDB to search for the optimal join order. This paper combines parallel exploration of join orders with parallel data processing. | VLDB 2022 | Parallel Query Execution, Adaptive Data Partitioning |
| SkinnerDB: Regret-bounded Query Evaluation via Reinforcement Learning | Immanuel Trummer, Junxiong Wang, Ziyun Wei, Deepak Maram, Samuel Moseley, Saehan Jo, etc | This paper present SkinnerDB. SkinnerDB employs the UCT algorithm (a reinforcement learning algorithm) to find the optimal join order for a query. In order to reduce the overhead to store intermediate join state this paper uses a Multi-way Join Algorithm which generates at most one tuple as the intermediate result. SkinnerDB quickly switches between different join orders to find the optimal join order for the query on the go. This system doesn't use prior experiences and fully depend on the feedback of DBMS to drive the learning process. This learning process is repeated for each query execution. | SIGMOD 2021 | Join Order Learning, Query Optimization, Reinforcement Learning |
| | | | | |

0 comments on commit c127527

Please sign in to comment.