Mutual Information Surprise: Rethinking Unexpectedness in Autonomous Systems

Description of Software and Data

The code and data in the zip file referenced above are a snapshot of the software and data that were used in the research reported in the paper "Mutual Information Surprise: Rethinking Unexpectedness in Autonomous Systems" by Yinsong Wang, Quan Zeng, Xiao Liu, Yu Ding.

The goal of this repository is to replicate the numerical experiments in the paper.

Computer and Software Environment

The following describes the computer hardware conditions and software environment on which the authors produce the results reported in the paper.

CPU: i9-13900KF
GPU: RTX 4090
Storage: SKHynix HFS001TEJ9X102N, Sequential Read: 6000MB/s, Sequential Write: 4000MB/s
RAM: DDR5 32Gb, 4400MT/s
OS: Windows 11

Dependencies

The code in this repository requires the following dependencies. The dependency version number corresponds to the version of the package with which the code was tested.

• Matplotlib 3.10.0
• numpy 2.3.5
• pandas 2.3.3
• scipy 1.16.3

Installation

Unzipping all the files into one folder should be enough. If there are directory errors, adjust BASE_DIR, DATA_DIR, SIMU_DIR, OUTPUT_DIR within the utils_dynamic.py file accordingly.

Reproducibility Workflow

To reproduce the results in Figure 1

• Code File: Figure1.py
  
• Output: The four plots in Figure 1 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 1 minute

To reproduce the results in Figure 2

• Code File: Figure2.py
• Output: The four plots in Figure 2 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 1 minute

To reproduce the results in Figure 3

• Code File: Figure3.py
• Output: The four plots in Figure 3 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 1 minute

To reproduce the results in Figure 4

• Code File: Figure4.py
• Output: The four plots in Figure 4 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 1 minute

To reproduce the results in Figure 5

• Code File: Figure5.py
• Output: The four plots in Figure 5 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 1 minute

To reproduce the results in Figure 6

• Code File: Figure6.py
• Output: The four plots in Figure 6 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 1 minute

To reproduce the results in Figure 7

• Code File: Figure7.py
• Output: The four plots in Figure 7 in one PDF file.
• Run Time at the Above-Specified Computer Conditions: 1 second

To reproduce the results in Figure 8

• Data File: pollution_results.csv
• Code File: Figure8_raw.py
• Output: The four plots in Figure 8 in separate PDF files.
• Run Time at the Above-Specified Computer Conditions: 8 hours

To reproduce the results in Figure 9

• Data File: phase1.csv; phase2.csv
• Code File: Figure9_raw.py
• Output: The plot in Figure 9 in one PDF file.
• Run Time at the Above-Specified Computer Conditions: 1 hour

To reproduce the results in Figure 10

• Code File: Figure10.py
• Output: The plot in Figure 10 in one PDF file.
• Run Time at the Above-Specified Computer Conditions: 4 seconds

To reproduce the results in Figure 11

• Data File: phase1.csv; phase2.csv
• Code File: Figure11_raw.py
• Output: The plot in Figure 11 in one PDF file.
• Run Time at the Above-Specified Computer Conditions: 1 hour

To reproduce the results in Table 1

• Code File: Table1.py
• Output: The numerical results in Table 1, saved in one csv file.
• Run Time at the Above-Specified Computer Conditions: 1 second

To reproduce the results in Table 2

• Code File: Table2.py
• Output: The numerical results in Table 2, saved in one csv file.
• Run Time at the Above-Specified Computer Conditions: 1 second

Note

Table 1 and Table 2 currently rely on our pre-generated simulation results. If one wishes to run it with new results after running Figure8_raw.py and Figure9_raw.py, it can be achieved by changing the SIMU_DIR in utils_dynamic.py to BASE_DIR/”outputs/generated_simu_data”.

Cite

To cite the contents of this repository, please cite both the paper and this repository using their respective DOIs.

Article: https://doi.org/10.1287/ijds.2026.0182
Software and Data Repository: https://doi.org/10.1287/ijds.2026.0182.cd

License

Copyright (c) (2026 Wang, Zeng, Liu, Ding)

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.