ScienceWorld: An interactive text environment to study AI agents on accomplishing tasks from the standardized elementary science curriculum.
Project description
ScienceWorld
ScienceWorld is a text-based virtual environment centered around accomplishing tasks from the standardized elementary science curriculum. This code accompanies the paper ScienceWorld: Is your Textual Agent Smarter than a 5th grader?.
Examples
Read some of the playthrough transcripts available at: https://sciworld.apps.allenai.org/explore
Citation
@misc{scienceworld2022,
title={ScienceWorld: Is your Agent Smarter than a 5th Grader?},
author={Ruoyao Wang and Peter Jansen and Marc-Alexandre C{\^o}t{\'e} and Prithviraj Ammanabrolu},
year={2022},
eprint={2203.07540},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2203.07540}
}
Quickstart
Before running: You will have to have Java 1.8+
installed on your system (shipped with most linux distributions).
Install with pip:
conda create --name scienceworld python=3.8
conda activate scienceworld
pip install scienceworld
Run an example random agent, on task 13 (classification: place a non-living thing in a box), for 5 episodes:
python examples/random_agent.py --task-num=13 --num-episodes=5 --simplifications-preset easy
Run a user console where you can interact with the environment, on task 3 (change of state: melting):
python examples/human.py --task-num=3 --num-episodes=5
Web Server Demo
A web server demo is also available, that allows running a ScienceWorld user console that can be interacted with in a web browser.
To run the web server demo:
conda create --name scienceworld python=3.8
conda activate scienceworld
pip install scienceworld[webserver]
Run the web server:
python examples/scienceworld-web-server-example.py
Point your web browser to:
localhost:8080
ScienceWorld Design
ScienceWorld is written in Scala (2.12.9), and compiles using sbt
into a JAR file that is run with java
. For convenience, a python
API is provided, which interfaces using the py4j
package.
Ports: ScienceWorld is nominally run as a server, which interfaces to the Python API with py4j
through a port. The default port is 25335
. The actual port used will be 25335 + the thread number provided when a ScienceWorld class is instantiated.
Threads: ScienceWorld is designed to run many threads simultaneously, if desired. To do so, initialize one ScienceWorldEnv
object per thread, and provide a unique threadNum
for each thread. Don't forget to close down servers that you instantiate using the env.shutdown()
command. If you are spawning many threads (10+) at the same time, you may wish to add a short delay (5-10 seconds) after initialization to wait for all the servers to initialize.
Tasks
The subtasks and their associated subtask indices are listed below. Subtask indices are automatically assigned by sorting task names:
SubTask # | Task Name |
---|---|
0 | task-1-boil |
1 | task-1-change-the-state-of-matter-of |
2 | task-1-freeze |
3 | task-1-melt |
4 | task-10-measure-melting-point-(known-substance) |
5 | task-10-measure-melting-point-(unknown-substance) |
6 | task-10-use-thermometer |
7 | task-2-power-component |
8 | task-2-power-component-(renewable-vs-nonrenewable-energy) |
9 | task-2a-test-conductivity |
10 | task-2a-test-conductivity-of-unknown-substances |
11 | task-3-find-animal |
12 | task-3-find-living-thing |
13 | task-3-find-non-living-thing |
14 | task-3-find-plant |
15 | task-4-grow-fruit |
16 | task-4-grow-plant |
17 | task-5-chemistry-mix |
18 | task-5-chemistry-mix-paint-(secondary-color) |
19 | task-5-chemistry-mix-paint-(tertiary-color) |
20 | task-6-lifespan-(longest-lived) |
21 | task-6-lifespan-(longest-lived-then-shortest-lived) |
22 | task-6-lifespan-(shortest-lived) |
23 | task-7-identify-life-stages-1 (animal) |
24 | task-7-identify-life-stages-2 (plant) |
25 | task-8-inclined-plane-determine-angle |
26 | task-8-inclined-plane-friction-(named-surfaces) |
27 | task-8-inclined-plane-friction-(unnamed-surfaces) |
28 | task-9-mendellian-genetics-(known-plant) |
29 | task-9-mendellian-genetics-(unknown-plant) |
Baseline Agents
Coming soon!
Additional Documentation
Coming soon!
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file scienceworld-1.0.1.tar.gz
.
File metadata
- Download URL: scienceworld-1.0.1.tar.gz
- Upload date:
- Size: 7.3 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.8.0 pkginfo/1.8.2 readme-renderer/34.0 requests/2.27.1 requests-toolbelt/0.9.1 urllib3/1.26.8 tqdm/4.63.0 importlib-metadata/4.11.3 keyring/23.5.0 rfc3986/2.0.0 colorama/0.4.4 CPython/3.9.7
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 118a0ee2f4ea50b71083246c942b743fe44d0c592882c646af54261311f18899 |
|
MD5 | dd382bc81dbc75e6c43539f78ce74a04 |
|
BLAKE2b-256 | 5b810f1ae3897d93148cd96b29d816790125f45fd9f957158e6c940d03e883f0 |
File details
Details for the file scienceworld-1.0.1-py3-none-any.whl
.
File metadata
- Download URL: scienceworld-1.0.1-py3-none-any.whl
- Upload date:
- Size: 7.3 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.8.0 pkginfo/1.8.2 readme-renderer/34.0 requests/2.27.1 requests-toolbelt/0.9.1 urllib3/1.26.8 tqdm/4.63.0 importlib-metadata/4.11.3 keyring/23.5.0 rfc3986/2.0.0 colorama/0.4.4 CPython/3.9.7
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 782b3e8643aca544acb36ea2a00677298c7b4cbef1add13ad66a3dc3e872c87d |
|
MD5 | 4f38828261b58761058e4ddfb76d17be |
|
BLAKE2b-256 | 960c20f38f0e27f7f225c1e9e1d496f677877b9565e20d1c646e60ee964fdcdb |