September 1st, 2023

GUEST POST – EmbedElite Meets Semantic Kernel: A Game-changer for Context Retrieval

Lukas Schmyrczyk
Co-Founder

Introduction

We are EmbedElite, proud winners of the recent Semantic Kernel Plugins Hackathon. A common challenge for enterprises is the extensive effort in gathering and preparing the necessary context which isn’t inherent to LLMs. Recognizing this, we at EmbedElite, set out on a mission to democratize this process, making context retrieval a more streamlined experience for developers worldwide.

The Challenge of Context Retrieval

In niche areas like legal court decisions, patent databases, intricate financial regulations, or the latest EU VAT rules, building applications with domain-specific context is an uphill challenge. Despite the data being public, the process of retrieval, preprocessing, and timely integration into LLM applications require considerable effort.

EmbedElite: Democratizing the Context Creation Process with Semantic Kernel

At EmbedElite, our mission is to provide access to domain-specific knowledge and streamline and democratize the process. Developers no longer need to repeatedly undergo the process of sourcing and refining data when they could plug into a ready-made solution. Our platform hosts a constantly updated marketplace for LLM assets maintained by domain experts, curating the very context that numerous developers are tirelessly searching for.

Semantic Kernel, an exceptional software development kit (SDK), is at the forefront of integrating large language models with leading programming languages.

With EmbedElite, developers can simply plug context-rich assets into their Semantic Kernel applications, reducing their solutions’ build time and enhancing their reliability. Not only does this approach free the developers from the burden of preparing and maintaining a context database, but it also ensures that they are working with credible, curated, and validated data, improving the overall quality of the application. Developers can now focus more on innovating and less on data gathering, resulting in faster and more reliable solutions built with Semantic Kernel.

EmbedElite and Semantic Kernel: A Winning Partnership

The EmbedElite plugin for Semantic Kernel makes context retrieval easy. With this tool in their arsenal, developers can dive straight into tasks like retrieving the latest EU VAT rules, understanding complex patent structures, or analyzing dynamic market trends. This solution transforms these labor-intensive tasks into achievable, streamlined processes, vastly reducing the time to deployment for developers and companies. The EmbedElite plug-in implements a native function that enables you to pass a query to EmbedElite to retrieve additional context. Under the hood, the native functions executes a HTTPS request to the EmbedElite API and returns the response including the context to the Semantic Kernel workflow that you can process further. Below you can see usage of our plug-in in action:

from plugins.EmbedElitePlugin.EmbedElitePlugin import EmbedElitePlugin

# ...

# Import the semantic functions

kernel.import_semantic_skill_from_directory("./plugins", "EmbedElitePlugin")

# Import the EmbedElite Plugin

embedElitePlugin = kernel.import_skill(

EmbedElitePlugin(), "EmbedElitePlugin"

)

# Call EmbedElitePlugin to fetch responses from EmbedElite API

response = embedElitePlugin["semantic_search_query"].invoke(

"What are the VAT rules in Germany if you sell services?"

)

print(response.result)

# ...

The Semantic Kernel Hackathon: A Validation of EmbedElite’s Potential

During the Semantic Kernel Plugin Hackathon, our team, consisting of Christoph Drayss, Mike Woodcock, and Lukas Schmyrczyk, demoed our plug-in. We chose the VAT rules embeddings hosted on the EmbedElite platform to answer tax related questions based on the latest VAT regulations in the EU. This practical demonstration led our team to win the community prize, endorsing the practicality and robustness of our collaboration.

Image EmbedElite Meets Semantic Kernel A Game changer for Context nocaps
Integrating SK with EmbedElite: User query -> Plugin utilization -> API communication -> Context retrieval

EmbedElite and Semantic Kernel: Shaping the Future

As we move forward, EmbedElite is committed to driving greater innovation in LLMs. From legal indices to evolving tax directives, to rapidly shifting market trends, our integration with Semantic Kernel opens opportunities to parse and understand complex contexts with unprecedented ease. The field of context-retrieval is an emerging research field. With our Semantic Kernel plug-in, we aim to contribute domain-specific knowledge to the SK ecosystem using the latest retrieval techniques like vector search using embeddings, hybrid search, knowledge graph retrieval, RAG algorithms and more.

Conclusion

From identifying real-world challenges to architecting a state-of-the-art plugin and witnessing its triumph at the Semantic Kernel Hackathon, our expedition with Semantic Kernel is an exciting and promising one. We invite you to be a part of this transformation, to chart new territories in the domain of contextual understanding in LLMs.

Join us on this journey with EmbedElite and Semantic Kernel. Together, we can redefine context retrieval, one domain at a time.

Learn more about EmbedElite on our website:

https://www.embedelite.com

You can find the EmbedElite SK plug-in on GitHub:

https://github.com/embedelite/embedelite-sk-plugin

Find the latest Semantic Kernel version on GitHub:

https://github.com/microsoft/semantic-kernel

Author

0 comments

Discussion are closed.