From Reactive to Proactive: How AI is Redefining Resilience in Modern Networks
summary
Prototype copilot uses a traffic simulator called CT SIM and a pre-trained ML model to predict which RAN cells will be overloaded, trained from data collected on over 80 base stations in Sweden; it outputs impacted cells, predicted cell loads, and automatically generates off‑peak remediation schedules that can be executed with PR level authorization. The System Behavior Analysis tool learns normal log sequences, detects timing shifts down to microseconds, groups duplicate faults, suggests root modules, and plugs into CI for early anomaly detection.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker intro
- 1:26Company overview and expertise
- 2:13Today's talk overview
- 3:19Network engineer co-pilot explained
- 5:51System workflow and demo setup
- 9:25Demo of co-pilot solution
- 14:06Development and validation of co-pilot
- 15:31Q&A invitation
- 15:51Smart log analysis tool introduction
- 16:12Traditional vs AI-based log analysis
- 17:14SBA tool domains and deployment
- 20:26Automated fault detection and grouping
- 21:08Conclusion and closing
related talks
10mThe Plan That Learns
26mProactive Failure Prediction Through Anomaly Detection
18mPredictive Network: How AI Forecasts System Failures
18mExtending Mobility - How AI Contributes towards the Future of Railway Transportation
11mAI-Native Network Operation
11mLightning Talk: Outsourcing Telco Understanding AI
19mTeaching AI to Fight Fires: Autonomous Reliability Agents at Meta Scale
26mManufacturing Alerts To ReActive Agents: Self-Evolving AI Agents on K8S for Predictive…