How to Ground LLM’s to minimize hallucinations

Language models have made significant advancements in recent years, with models like GPT-3 and GPT-4 showcasing impressive capabilities. However, one persistent challenge that arises with these models is the occurrence of hallucinations—instances where the model generates plausible-sounding but incorrect or nonsensical responses.

In this talk, we will explore strategies to ground language models to minimize the occurrence of hallucinations. By grounding LLMs, we aim to enhance their reliability and ensure that the generated outputs align more closely with factual accuracy and logical coherence.

We will discuss various techniques and approaches that can be employed to address hallucinations effectively. These may include fine-tuning the models on domain-specific data, incorporating external knowledge sources, leveraging human-in-the-loop feedback, and implementing robust evaluation mechanisms.

Furthermore, we will delve into the underlying causes of hallucinations and examine the limitations of current language models. By understanding these factors, we can develop targeted strategies to mitigate the occurrence of hallucinations and improve the overall performance of LLMs.

Join us in this talk as we explore practical methods to ground language models and minimize hallucinations. Discover how these techniques can enhance the reliability and trustworthiness of LLMs, making them more suitable for real-world applications across various domains. Together, let’s unlock the full potential of language models while ensuring their outputs align with factual accuracy and logical coherence.

I’m a Developer not a Data Scientist: How do I leverage LLM’s?

As a developer, you may feel intrigued by the potential of Large Language Models (LLMs) like GPT-3 and GPT-4 but wonder how to effectively leverage them without extensive data science expertise. In this talk, we will guide developers on harnessing the power of LLMs and incorporating them into their projects.

LLMs offer incredible capabilities for natural language processing, content generation, and more. However, you don’t need to be a data scientist to tap into their potential. We will explore practical strategies and tools that empower developers to work with LLMs efficiently and effectively.

Join us as we discuss techniques for data preparation, model selection, and fine-tuning LLMs to suit your specific needs. We will explore user-friendly libraries, APIs, and pre-trained models that simplify the integration process, allowing developers to focus on their core expertise.

Additionally, we will address common challenges developers may encounter when working with LLMs. By understanding these challenges and learning best practices, developers can navigate potential pitfalls and optimize their use of LLMs.

Through hands-on demonstrations and real-world examples, we will showcase how developers can leverage LLMs in various applications. You will leave this talk equipped with practical knowledge and resources to start incorporating LLMs into your projects immediately.

Join us to discover how, as a developer, you can unlock the power of LLMs and enhance your applications with cutting-edge natural language processing capabilities. Embrace the potential of LLMs without the need for extensive data science expertise and take your projects to new heights.

Creating your own ML powered Holographic Assistant Workshop

I’ve been fascinated with Mixed Reality since the introduction of the HoloLens.  There is something magical about combining Artificial Intelligence with Holograms.  In this workshop, I will show you a practical example of how to mix holograms and AI to create your own holographic assistant.

For this workshop, I will skip the theory and go directly to application using off-the-shelf tooling and models.  I hope to inspire the audience to look closer at the capabilities of off-the-shelf machine learning by tapping into the power of ML without building any custom models.

This talk will make use of a lot of different tools and technologies all available, accessible, and inexpensive.  I will use C#, Unity, a Holographic Display, Azure Cognitive Services, and Azure Open AI to create an interactive experience pushing the envelope of current tools and techniques.

Each attendee will create their own assistant using their own laptop, Azure subscription, Visual Studio, and Unity.  At the end of the workshop attendees will have the opportunity to demonstrate their assistant on a holographic display but will also build a working assistant that runs on 2D displays.

  • See an end-to-end solution created around LLM’s
  • Inspire the attendee to do more than create another text based chat bot
  • Help the attendee to walk away with their own working assistant

Stop playing with ChatGPT, let’s get serious about LLM’s

ChatGPT has taken the mainstream media by storm and everyone is talking about it. From tech enthusiasts to skeptics, everyone is buzzing about this groundbreaking AI.   With such a big impact in the consumer space many enterprises are attempting to incorporate ChatGPT into their products and workflow.

Be warned though, ChatGPT is a great demo and a fun toy to play with but it is just that, a toy.  OpenAI themselves has even published the warning: “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers.”. In other words, ChatGPT should NOT be used for production usage in most scenarios.

The good news is that ChatGPT is built on top of the Foundation models GPT 3.5 and GPT 4.0. The true power lies in these Foundation models, which, when wielded by skilled hands, can deliver exceptional results for your production needs.

Don’t miss out on this game-changing technology. Embrace the future of AI and discover how the remarkable Foundation models, can elevate your organization to new heights. Let’s delve into this this together!