This repository contains agents for the SCML world.
To install this package just run:
>>> pip install scml-agents
There are two ways to submit agents to this repository:
- Participate in the SCML competition https://scml.cs.brown.edu
- Submit a pull-request with your agent added to the contrib directory.
Please note that torch does not yet fully support python 3.11. If you face issues installing (especially on a mac), try to use Python 3.10.
You can get any specific subset of the agents in the library using get_agents(). This function has the following parameters:
- version: Either a competition year (2019, 2020, 2021, ....) or the value "contrib" for all other agents. You can also pass "all" or "any" to get all agents.
- track: The track (any, collusion, std, sabotage[only for 2019], oneshot [from 2021]).
- qualified_only: If true, only agents that were submitted to SCML and ran in the qualifications round will be
returned
- finalists_only: If true, only agents that were submitted to SCML and passed qualifications will be
returned
- winners_only: If true, only winners of SCML (the given version) will be returned.
- bird_only: If true, only winners the bird award are returned (new in 2021)
- top_only: Either a fraction of finalists or the top n finalists with highest scores in the finals of
SCML
- as_class: If true, the agent classes will be returned otherwise their full class names.
For example, to get the top 10% of the Oneshot track finalists in year 2023 as strings, you can use:
>>> get_agents(version=2023, track="oneshot", finalists_only=True, top_only=0.1, as_class=False)
- First Place: Pedro Hrosz Turini and Jaime Sichman for QuantityOrientedAgent
- Second Place: Shota Kimata and Yuko Sakurai for CCAgent
- Third Place: Masato Kijima for KanbeAgent
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2023, track="oneshot", winners_only=True)
- First Place: Sadahiro Atsunaga for AgentSDH
You can get this agent after installing scml-agents by running:
>>> scml_agents.get_agents(2023, track="std", winners_only=True)
- Honorary Mention: Kazuki Komori for M5
You can get this agent after installing scml-agents by running:
>>> scml_agents.get_agents(2023, track="collusion", winners_only=True)
- First Place: Chris Mascioli and Amy Greenwald for PatientAgent
- Second Place: Takumu Shimizu for GentleS
- Third Place: Shiraz Nave, Amit Dayan, Sariel Turayfor AgentSAS
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2022, track="oneshot", winners_only=True)
- First Place: Ito Nobuhiro and Takanobu Otsukafor Lobster
- Second Place: Kazuki Komori for M5
- Third Place: Koki Katagiri and Tatanobu Otsuka for Artisan Kangaroo
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2022, track="std", winners_only=True)
- Honorary Mention: Kazuki Komori for M5
You can get this agent after installing scml-agents by running:
>>> scml_agents.get_agents(2022, track="collusion", winners_only=True)
- First Place: Assaf Tirangel, Yossi Weizman, Inbal Avraham for Agent112
- Second Place: Takumu Shimizu for Gentle
- Third Place (tie): Sagi Nachum for Agent74
- Third Place (tie): Yuchen Liu, Rafik Hadfi and Takayuki Ito for UCOneshotAgent
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2021, track="oneshot", winners_only=True)
- First Place: Kazuki Komori for M4
- Second Place: Mehmet Onur Keskin, Umit Cakan, Gevher Yesevi, Reyhan Aydogan, Amy Greenwald for CharliesAgent
- Third Place: Koki Katagiri for Artisan Kangaroo
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2021, track="std", winners_only=True)
- First Place: Kazuki Komori for M4
- Second Place: Mehmet Onur Keskin, Umit Cakan, Gevher Yesevi, Reyhan Aydogan, Amy Greenwald for CharliesAgent
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2021, track="collusion", winners_only=True)
- First Place: Masahito Okuno for SteadyMgr
- Second Place: Guy Heller, E. Gerson, I. Hen and M. Akrabi for Agent30
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2020, track="std", winners_only=True)
- First Place: Kazuki Komori for MMM
- Second Place: Ayan Sengupta for Merchant
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2020, track="collusion", winners_only=True)
A list of all agents accepted for the SCML 2022 qualifications round can be found at https://scml.cs.brown.edu/scml2022
A list of all agents accepted for the SCML 2021 qualifications round can be found at https://scml.cs.brown.edu/scml2021
This is a list of all the agents accepted for the SCML 2020 qualifications round.
Team | Identifier | Agent/Class name | Team Members |
---|---|---|---|
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
------------- | ------------- | ----------------------- | ---------------------------------------------------------------------------------------------------- |
|
|
|
|
You can get these agents after installing scml-agents by running:
>>> scml_agents.get_agents(2020, track="any")
If you are on Apple M1, you will need to install tensorflow before installing this package on conda using the method described here