microsoft / microsoft/CyberBattleSim
More realistic random domain generation required
Nobody has claimed this yet.
- Dominant language
- Jupyter Notebook
- Stars
- 1.8k
- Forks
- 286
- Avg merge
- 4h 35m
- Merged PRs (30d)
- 12
Description
The static toy environments are not suitable for training an agent for generalization. The chainpattern is solvable with a trivial strategy, which is not enough to demonstrate that a certain method is good enough.
The random environment in cyberbattle.simulation.generate_network is not realistic and deep enough. In a random network of 50 nodes, only 1 or 2 are actually hackable.
The random environment in cyberbattle.simulation.environment_generation seem to be very limited as well (usually, no other nodes discovered).
To train generalizing agents, a more complex, dynamic and realistic environment generation is needed.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading cyberbattle.simulation.generate_network and cyberbattle.simulation.environment_generation, then run the existing random environments to observe node discovery and hackability. Define measurable criteria for a more complex, dynamic, realistic environment that supports agent generalization; completion requires those generators to produce deeper, more consistently usable networks.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning, networking
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100