Skip to content

parmanu-lcs2/inform

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Disentangling Intrinsic Importance from Emergent Structure in Multi-Expert Orchestration

Accepted by Transactions on Machine Learning Research (TMLR)

OpenReview: https://openreview.net/forum?id=4W7sgat04A

INFORM Framework Image

Setup

Install all dependencies using the following command:

$ pip install -r requirements.txt

Replace all placeholder values like BASE_URL, API_KEY with your own values, across all files.

Directories

The source code is organized in the following directories:

  • ablations
  • analysis
  • eval
  • plotting
  • src

Training

Running the following script will produce artifcats in the checkpoints folder:

$  python src/train.py

Citation

If you find this work useful in your research, please consider citing:

@article{
      ghosh2026disentangling,
      title={Disentangling Intrinsic Importance from Emergent Structure in Multi-Expert Orchestration},
      author={Sudipto Ghosh and Sujoy Nath and Sunny Manchanda and Tanmoy Chakraborty},
      journal={Transactions on Machine Learning Research},
      issn={2835-8856},
      year={2026},
      url={https://openreview.net/forum?id=4W7sgat04A}
}

About

Source Code for "Disentangling Intrinsic Importance from Emergent Structure in Multi-Expert Orchestration"

Resources

License

Stars

Watchers

Forks

Contributors

Languages