SLM-Driven RAN Intelligence

post thumb
18 Feb 2025, Sophia-Antipolis, France, BubbleRAN & Khalifa University

🤖Automating Telecom network operations with specialized language models (SLM) for enhanced efficiency and performance

SLM-Driven RAN Intelligence

BubbleRAN, a leader in intelligent network solutions, and Khalifa University’s 6G Research Center, renowned for cutting-edge telecommunications research, today announced a pioneering collaboration to demonstrate SLM-driven RAN Intelligence at Mobile World Congress 2025 (Barcelona, March 03-06).

This first-of-its-kind showcase highlight how Small Language Models (SLMs) are redefining the role of GenAI in telecom network automation accelerating telecom network operations. This initiative is further strengthened by Aalto University’s research expertise, contributing to advancements in AI-driven Telecommunications.

The shift to SLMs: Smarter, Leaner AI for Telecom.

While Large Language Models (LLMs) excel in general-purpose tasks, their computational demands and latency challenges limit their usage in telecom applications. BubbleRAN and Khalifa University’s solution leverages specialized SLMs (1–7 billion parameters) tailored for the telecom domain to deliver resource efficiency, domain expertise, real-time decision-making.

Traditional AI tools struggle to contextualize the flood of data in modern networks. BubbleRAN Intelligent Telco Agent (MX-AI), powered by Khalifa University’s SLM technology, addresses this by

  1. Answering deployment specific questions
  2. Automating Troubleshooting
  3. Enabling Proactive Optimization
  4. Simplifying Operations

Use Case in Action

Operator asks:
“Why did throughput drop in Sector X?”
Agent responds:
Cross-references live data, pinpoints a misconfigured transmit power, and guides remediation – all in under 5 seconds.

This effectively turns the CLI tools into an intuitive and context-aware dialogue!

Leadership Insights

We are delighted to collaborate with BubbleRAN for this first-of-its-kind showcase at Mobile World Congress 2025 in Barcelona, the perfect platform for innovative technologies. Our initial pioneering work with Telecom domain specific LLM laid the foundations. Now, this joint demo on SLM-driven RAN intelligence represents the next big step—leveraging specialized, efficient models with reasoning capability to reshape the future of wireless communications with real-time adaptability and automation.Says Professor Merouane Debbah, Founder-Director, 6G Research Center, Khalifa University’s 6G Research Center

Our Intelligent Telco Agent acts as a controlled autopilot – empowering operators with actionable insights while retaining human oversight. This is how we make 6G sustainable. Added Navid Nikaein, CEO of BubbleRAN.

Innovation

Unlike generic chatbots, BubbleRAN’s solution is tailored for telecom: it not only replies to generic questions about the Open Radio Access Network (O-RAN) stack or 3rd Generation Partnership Project (3GPP) standard but it also integrates live network data for real-time to answer specific questions related to the current network deployment, anomalies and how to troubleshoot or how to optimize the network, all in a form of actionable insights.

Venue

The full demonstration will be displayed from 3rd until 6th of March 2025 at BubbleRANstand number 6E9 at the MWC Barcelona 2025.

About Khalifa University

Khalifa University of Science and Technology (KU) is a research university dedicated to the advancement of learning through teaching and research and the discovery and application of knowledge. Khalifa University strives to be a catalyst to the growth of Abu Dhabi and the UAE’s rapidly developing knowledge economy, an education destination of choice, and a global leader among research intensive universities of the 21st century.

For more information, please visit https://www.ku.ac.ae/.

Are you interested to know more about MX-AI?

  1. Yes > contact us using the quick contact form
Tags: