Comments (9)
Hey Christoph, apologies for not updating you on this but I lost track of the issue.
The warnings and emails stopped when I set the app password for the user as you mentioned.
from suspicious_login.
From the graphs the classifier believes that it has a recall of 100% and precision of 81%, when that is actually false given that no real suspicious login has happened (yet at least).
FYI during training we simulate suspicious requests by using randomized IPs and assigning IPs from other one user to another user (shuffle).
How can I ensure that the app recognises legitimate logins over suspicious ones? There are no links in the warning emails asking whether that was a legitimate login or not. How does the app learn which suspicious logins are actually suspicious?
No, currently the app uses unsupervised learning, there is no user feedback.
from suspicious_login.
Do you have many users that access the instance from the same IP? I have to check the logic but maybe we shuffle the training data in a way that a user gets assigned an IP for the suspicious data set that it would otherwise also use as their legit login address. This is just a suspicion and I have to check the details.
from suspicious_login.
Yes, most users access the instance from the same IP. But most users don't generate so many warnings daily/weekly/monthly when they change their location and therefore IP. Usually after one warning/email and the model learns the new pattern.
But this one user that is a bit more dynamic than others (but the set of IPs is usually the same) keeps generating warnings for the same IPs over the span of a few days or even months. Sometimes the same IP is seen as suspicious for a few days in a row.
from suspicious_login.
Does the affected user use app passwords for the connected clients?
from suspicious_login.
No they don't
from suspicious_login.
Ok, it would certainly help if they did because app password logins are never considered suspicious
from suspicious_login.
hey @eibex any update?
from suspicious_login.
cool and thanks for the update :)
from suspicious_login.
Related Issues (20)
- training error. Not quite sure what im doing. HOT 1
- More Information link should open in new tab when clicked
- [Bug]: Caught unknown error during IPv6 background training HOT 2
- Update app to support NC v24 HOT 7
- First (legitimate) login attempt always fails (silently) HOT 8
- [I18N] Gettext escaping issue
- [feature] Whitelist IPs HOT 2
- Misleading german translation in machine learning chart legend HOT 4
- Memory consumption
- Log spam `jsonSerialize() should either be compatible with` HOT 1
- Setup psalm
- ValueError: random_int(): Argument #1 ($min) must be less than or equal to argument #2 ($max) HOT 10
- New login location detected email messages HOT 1
- App icon is inverted on app management
- Replace lint.yml with split workflows
- Dependency Dashboard
- Update rubixML HOT 15
- App not passing integrity check in v26/25 HOT 3
- Too verbose logs when model not found HOT 1
- Unit tests don't execute against php 8.2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from suspicious_login.