Businesses today can make leaps and bounds to revolutionize the way things are done with the use of Large Language Models (LLMs). LLMs are widely used by businesses today to automate certain tasks and create internal or customer-facing chatbots that boost efficiency.
Challenges with dynamic adaption of LLMs
As with any new hyped-up thing that dropped to the market fast, there are a few challenges with traditional, pre-trained LLMs that do not have retrieval functionality. For example, think of a chatbot that has been trained on a massive amount of data in 2020 but if you ask it a simple question about what the weather is today in your area, or the ongoing score of a football game, it’s unable to give you a proper answer. You may have already come across this problem.
This kind of LLM simply does not know present data and is missing the relevant, timely context required to answer this question correctly. You are lucky if it admits to not knowing the answer but it may just hallucinate and give you an incorrect answer.
According to NASA, Saturn has 146 moons in orbit as of June 2023. Although this traditional LLM does point out that more moons may have been discovered since then, it is unable to give you an accurate, updated answer.
Which brings us to some of the core challenges we face with LLMs:
- It may not have a proper source or be able to cite one
- It can get outdated easily
- It can struggle to answer complex or challenging questions and may end up hallucinating answers
This is where Retrieval Augmented Generation (RAG) comes in. In the most basic of terms, Retrieval-augmented generation (RAG) is an AI framework for improving the quality of LLM-generated responses by grounding the model on external sources of knowledge.
For instance, in the example below, the app draws on McDonald’s annual reports to answer the question of how much total revenue was earned by franchised McDonald’s outlets in the year 2022.
Achieving business agility using RAG
RAG-equipped chatbots absorb their information from a variety of sources, including databases, documents, and the internet, to provide accurate and contextually relevant responses. This is particularly useful when users have complex or multi-step queries. Using a RAG system contributes significantly towards making the business more agile, especially if the company has a customer-facing chatbot.
Here are a few examples of how businesses are embracing RAG-based systems today:
- Financial services companies are using RAG to generate personalized financial advice and risk assessments.
- Healthcare companies are using RAG to generate personalized treatment plans and answer patient questions about their medical conditions.
- E-commerce companies are using RAG to generate personalized product recommendations.
- Customer support teams are using RAG to generate personalized customer support responses and answer customer questions more quickly and efficiently.
With RAG, the core challenges faced with outdated LLMs are mitigated by augmenting new data to the LLM to help it give more updated and context-aware responses. This is done in two ways:
- Training the model on new data by constantly searching the web
- Uploading documents to enhance the LLM module and give it more information to base its responses on
For example, consider a chatbot being used in a medical context. An LLM without RAG may struggle to provide a good answer to the question “A patient is calling in to dispute a hospital charge. What documents and evidence do they need to attach to their claim?”. A RAG-based system, however, would be able to retrieve some relevant information from internal data uploaded by the hospital. An LLM-powered system can also be prompted to not answer questions that are not contained within the trusted information.
H2O.ai’s generative AI tools are based on a RAG-centric approach and can be used to easily create internal chatbots that are context-aware, updated, and customized toward your business use cases.
In addition to improving accuracy, RAG can also help to reduce bias in question answering systems. Pre-trained LLMs can be biased towards the data that is already baked into the model. For example, a pre-trained LLM that is trained on a dataset of news articles might be biased towards certain topics or perspectives.
RAG can help to reduce bias by retrieving information from a variety of sources, including sources that are known to be unbiased. This allows RAG-based systems to provide more objective and unbiased answers to questions.
Now, as a business, it is expensive to keep continuously fine-tuning a large language model to constantly include new data and become more context aware. You may even have internal data that you are cautious about exposing. Using a RAG approach, businesses can leverage their own internal data for generating precise, context-aware responses based on internal trusted information, and without incurring the substantial operational costs associated with continuously fine-tuning the model. Even with the fast-paced, constantly evolving nature of LLMs, the added benefit of the RAG approach is that companies just don’t need to worry about fine-tuning newer LLMs that come out. With this approach, you can simply upload your data at question/query time, making it instantly context-aware.
H2O.ai uses our own tools like h2oGPT, H2O LLM Studio, and H2O LLM Data Studio to achieve new levels of productivity within the company, and they can be used by other businesses to do the same. The most enticing benefit being that you can still hold ownership of your own internal data while creating a customized and updated chatbot to provide updated, specific, and context-aware responses to user prompts.
As the world becomes increasingly interconnected and reliant on data-driven decisions, the need for powerful and innovative AI solutions has never been more critical. At H2O.ai, we've been at the forefront of AI and machine learning for the last decade, providing you with the tools and platforms to harness the power of data. Today, we're thrilled to unveil the public H2O GenAI App Store. If you're eager to see what's available, head over to genai.h2o.ai.
What is the H2O GenAI App Store?
H2O has long been committed to supporting the development of custom, internal app stores for businesses. We believe that AI Solutions are only helpful if they are in the hands of businesses and allowing data science teams to rapidly build and deploy custom apps for their end users is a big part of that mission. This philosophy has enabled organizations to create tailored solutions using our powerful Machine Learning, AutoML, and Generative AI technologies. Today, we're proud to announce the public H2O GenAI App Store, a platform filled with 10+ open-source Generative AI apps crafted by the makers at H2O.ai.
The H2O GenAI App Store is the go-to destination for those looking to understand the possibilities of Generative AI. You can visit today to explore the apps, or make an account to start using them today. Topics range from apps that help you plan your weekly meals to apps that help you understand the financial situation of public companies with new apps coming over time.
The code behind each app can be found in our open source repo. This repository is teeming with apps created using 100% Python, leveraging the dynamic capabilities of H2O Wave and H2O's Generative AI products. Whether you're an AI enthusiast, a developer, or someone eager to leverage AI for specific use cases, the H2O GenAI App Store is designed with you in mind. We encourage data scientists to submit their own apps to the GitHub if you would like to host them in the public GenAI App Store!
Meet the GenAI Apps
We’ll now walk through the apps that are available in the store and repository today. For the curious minds and developers among us, the source code for these apps is available on our GitHub repository. The open-source nature of these applications ensures that you can dissect, modify, and contribute to the ever-evolving landscape of Generative AI.
Retrieval Augmented Generation
These apps talk to one or Large Language Models to combine the power of language generation with information retrieval. They generate contextually relevant responses by retrieving information from documents within h2oGPTe and then generating human-like text to provide informative and contextually accurate responses.
Ask H2O enables users to ask open-ended questions to H2O’s product documentation.
The Study Partner app can run on any collection of documents and helps you learn more about the content by generating topics on the collection and then quizzing you with custom study questions on the topic of your choice. Users receive immediate feedback and tips on how to improve. In the H2O GenAI App Store, we have documents about the H2O Products, but this app could run on any collection of documents.
BrLawGPT is designed to streamline the study and interpretation of legal documents, particularly initial petitions in the Brazilian legal system. This application simplifies interactions with PDF documents, extracts valuable information, and enhances productivity across various use cases.
Chat with specific company SEC 10-Ks to extract information regarding potential risk positions and investment opportunities using the Financial Research app which comes pre-canned with historic 10-Ks.
The RFI Assistant simplifies the process of answering Requests for Information (RFIs). In the H2O GenAI App Store, this app is contextualized with H2O product information, but could run on any collection data to help anyone answering an RFI.
AI for Life
These apps can help the average person with their day to day life. While many people are using ChatGPT and LLMs for daily tasks today, wrapping these use cases in a custom UI makes it even easier for the average person to get value out of Generative AI without becoming an expert at Prompt Engineering.
Grow a thriving vegetable garden by getting expert AI help for growing the plants you want in your specific sub-climate. TomatoAI can help you with specific topics like “What is succession planting and how do I do it successfully in the tundra?” and with open ended questions you might have.
Do you cycle? And if not, do you want to start? Then pedal on over to the Cycling Training Plan app which allows you to share information about your goals and then creates a customized plan to help you achieve them.
Selling your home? Save time crafting the perfect listing to reach your target buyer by providing details and having AI write the listing for you in the Home Listing app.
Combining the power of H2O open source technology with open-source model Whisper, the Transcribe and Summarize app will do exactly that: take any public mp3 file and transcribe, summarize, and get the sentiment.
Why spend 3+ hours meal planning when you could spend under 3 minutes for the entire week? Weekly Meal Planning allows you to customize which meals you would like, how many people you have, and whether you want cooking instructions. It then customizes a prompt for you which you can edit yourself - maybe you only want Cuban cuisine - before sending to a Large Language Model for a customized meal plan.
If you’re a python developer, or want to start writing your own GenAI apps, the Clean Code app is for you! It offers assistance in writing amazing and easy-to-follow Python code.
Contribute to the Future
If you're excited about the possibilities that Generative AI opens up, we invite you to contribute to the development of GenAI apps. We welcome your Pull Requests (PRs), whether you want to enhance existing apps or create something entirely new.
The H2O GenAI App Store is not just a platform; it's a community of AI enthusiasts, developers, and innovators. Together, we can shape the future of AI-driven applications and empower businesses and individuals to achieve more with the power of Generative AI.
