Swisscom

Problem

The recent development of large language models (LLMs) has achieved remarkable success in natural language processing tasks, demonstrating impressive capabilities in complex reasoning, knowledge retention, and nuanced text generation. However, these models are often prohibitively large, requiring significant computational resources and energy consumption. This poses a major challenge for deploying LLMs in resource-constrained environments, such as edge devices, mobile phones, or low-power embedded systems. While LLMs have shown incredible abilities, their smaller counterparts, small language models (SLMs), often struggle to replicate these feats. The primary hurdle in developing capable SLMs lies in the inherent trade-off between model size and performance. As a result, there is a growing need to distill the vast knowledge and reasoning capabilities of LLMs into a more compact architecture without significant degradation in performance or frugal optimised multi LLM systems, that match each request to the most efficient sequence of model prompt chains.

Why hack?

Because developing capable small language models (SLMs) can enable a wide range of applications, from edge devices and mobile phones to low-power embedded systems. By participating in this challenge, you will be working on a real-world problem with significant impact, including: Enabling efficient natural language processing on resource-constrained devices Reducing energy consumption and carbon footprint Democratizing access to AI capabilities

Problem

All organisations need to learn to improve. In Swisscom and many other organisations we do this by creating course material, online course, video, powerpoints etc. Without lots of testing it is very difficult to validate the impact these courses will have.

Why hack?

As education material increases exponentially with AI we need a way to test the quality of these course.

Problem

At Swisscom we continuously collect information about its network infrastructure in real time. All this information is stored in a massive knowledge graph with more than 200 million nodes, capturing dependencies from network devices and services. While analyzing knowledge graphs with AI (embeddings) is well understood, there are still many challenges on dynamic evolving heterogeneous graphs. Such graphs play a key role in network forecasting, entity resolution or any form of interaction model (e.g. cells). Mining the gap between changes, interactions or any evolution of edges in real-time puzzles our engineers for years. Using the power of the Swiss AI LLM models creating natural human interaction interfaces to such complex graph analytics tasks Join our journey and spot temporal anomalies of structural graph changes.

Why hack?

Because detecting anomalies in such a huge, live-changing graph is a real-world problem with enormous impact. Better anomaly detection means: Faster root cause analysis of outages. More resilient networks. Millions of customers with more reliable service.