web analytics

Learn AI With Kesse | Newest Trends in Artificial Intelligence

We answer questions about artificial intelligence and bring you what's new in the AI World.

DataGemma AI Open Models Connecting LLMs to Googles Data Commons

3 min read

The rise of generative AI has brought remarkable advancements, but it’s not without challenges. One significant issue is the phenomenon known as “hallucination,” where AI confidently presents false information.

DataGemma aims to mitigate this by grounding large language models (LLMs) in real-world data from Google’s Data Commons. This development promises to enhance the factual accuracy of AI responses, making them more reliable and informative.

Data Commons: A Vast Repository of Trustworthy Data

Data Commons is a publicly available knowledge graph that houses over 240 billion rich data points across numerous statistical variables. This vast resource pulls data from trusted sources like the United Nations, World Health Organization, Centers for Disease Control and Prevention, and Census Bureaus.

This unified database provides a wealth of reliable, public information on various topics ranging from health to economics. Users can interact with the data in their own words using an AI-powered natural language interface, exploring questions like the correlation between income and diabetes in U.S. counties.

How Data Commons Tackles Hallucination

By integrating Data Commons within Gemma, DataGemma models aim to ground generative AI experiences in factual data. These lightweight, state-of-the-art open models, built from the same research as the Gemini models, are now available to researchers and developers.

DataGemma employs two distinct approaches to enhance LLM factuality: Retrieval-Interleaved Generation (RIG) and Retrieval-Augmented Generation (RAG). RIG queries trusted sources and fact-checks against Data Commons, while RAG incorporates relevant context beyond training data, enriching the model’s responses.

DataGemma’s Unique Methodologies

RIG enhances Gemma 2 by retrieving statistical data from Data Commons when generating responses. This proactive approach minimizes the risk of hallucinations by ensuring the information is accurate and reliable.

In contrast, RAG uses a long context window to absorb and include more contextual data before generating responses. This results in more comprehensive and informative outputs, reducing the AI’s chances of presenting incorrect information.

For example, when asked about the increase in renewable energy use, both RIG and RAG methodologies ensure that the response is backed by authoritative data from Data Commons.

Promising Results of DataGemma

Initial findings with RIG and RAG show promising enhancements in the accuracy of AI handling numerical facts. This improvement suggests fewer hallucinations in various use cases, benefiting researchers, decision-makers, and curious minds alike.

Ongoing research aims to refine these methodologies, scale up the work, and integrate the enhanced functionality into both Gemma and Gemini models through a phased, limited-access approach.

Future Directions and Aspirations

DataGemma’s open model seeks to facilitate broader adoption of these techniques, grounding LLMs in factual data. The goal is to make AI more reliable and trustworthy, ensuring it serves as an indispensable tool for everyone.

Researchers and developers can get started with DataGemma through quickstart notebooks for both the RIG and RAG approaches, enabling them to explore the potential of these advanced methodologies.

By continuously refining these methods, DataGemma aims to foster informed decisions and a deeper understanding of the world through accurate information grounded in real-world data.

Getting Started with DataGemma

Researchers eager to explore DataGemma’s capabilities can access quickstart notebooks for RIG and RAG, providing a hands-on approach to understanding these advanced methodologies.


DataGemma stands at the forefront of addressing AI hallucination by anchoring LLMs to real-world data from Google’s Data Commons. This innovative approach promises to enhance the accuracy and reliability of AI-generated responses.

As research progresses, DataGemma is set to become an essential tool for providing accurate, data-driven insights, ultimately empowering users with truthful information and fostering informed decisions.

About The Author

Leave a Reply

Your email address will not be published. Required fields are marked *

We use cookies to personalize content and ads and to primarily analyze our geo traffic sources. We also may share information about your use of our site with our social media, advertising, and analytics partners to improve your user experience. We respect your privacy and will never abuse your information. [ Privacy Policy ] View more
Cookies settings
Accept
Decline
Privacy & Cookie Policy
Privacy & Cookies policy
Cookie name Active

The content on this page governs our Privacy Policy. It describes how your personal information is collected, used, and shared when you visit or make a purchase from learnaiwithkesse.com (the "Site").

Kesseswebsites and Advertising owns Learn AI With Kesse and the website learnaiwithkesse.wiki. For the purpose of this Terms and Agreements [ we, us, I, our ] represents the owner of Learning AI With Kesse which is Kesseswebsites and Advertising. [ You, your, student and buyer ] represents you as the user and visitor of this site. Terms of Conditions, Terms of Service, Terms and Agreement and Terms of use shall be considered the same here. This website or site refers to https://learnaiwithkesse.com. You agree that the content of this Terms and Agreement may include Privacy Policy and Refund Policy. Products refer to physical or digital products. This includes eBooks, PDFs, and text or video courses. If there is anything on this page you do not understand you agree to reach out to us via email [ emmanuel@learnaiwithkesse.com ] for explanation before using any part of this site.

1. Personal Information We Collect

When you visit this Site, we automatically collect certain information about your device, including information about your web browser, IP address, time zone, and some of the cookies that are installed on your device. The primary purpose of this activity is to provide you a better user experience the next time you visit our again and also the data collection is for analytics study. Additionally, as you browse the Site, we collect information about the individual web pages or products that you view, what websites or search terms referred you to the Site, and information about how you interact with the Site. We refer to this automatically-collected information as "Device Information."

We collect Device Information using the following technologies:

"Cookies" are data files that are placed on your device or computer and often include an anonymous unique identifier. For more information about cookies, and how to disable cookies, visit http://www.allaboutcookies.org. To comply with European Union's GDPR (General Data Protection Regulation), we do display a disclaimer a consent text at the bottom of this website. This disclaimer alerts you the visitor or user of this website about why we use cookies, and we also give you the option to accept or decline. If you accept for us to use cookies on your site, the agreement between you and us will expire after 180 has passed.

"Log files" track actions occurring on the Site, and collect data including your IP address, browser type, Internet service provider, referring/exit pages, and date/time stamps.

"Web beacons," "tags," and "pixels" are electronic files used to record information about how you browse the Site.

Additionally, when you make a purchase or attempt to make a purchase through the Site, we collect certain information from you, including your name, billing address, shipping address, payment information (including credit card numbers), email address, and phone number. We refer to this information as "Order Information."

When we talk about "Personal Information" in this Privacy Policy, we are talking both about Device Information and Order Information.

Payment Information

Please note that we use 3rd party payment processing companies like https://stripe.com and https://paypal.com to process your payment information. PayPal and Stripe protects your data according to their terms and agreement and may store your data to help make your subsequent transactions on this website easier. We never and [ DO NOT ] store your card information or payment login information on our website or server. By making payment on our site, you agree to abide by the Terms and Agreement of the 3rd Party payment processing companies we use. You can visit their websites to read their Terms of Use and learn more about them.

2. How Do We Use Your Personal Information?

We use the Order Information that we collect generally to fulfill any orders placed through the Site (including processing your payment information, arranging for shipping, and providing you with invoices and/or order confirmations). Additionally, we use this [a] Order Information to:

[b] Communicate with you;

[c] Screen our orders for potential risk or fraud; and

When in line with the preferences you have shared with us, provide you with information or advertising relating to our products or services. We use the Device Information that we collect to help us screen for potential risk and fraud (in particular, your IP address), and more generally to improve and optimize our Site (for example, by generating analytics about how our customers browse and interact with the Site, and to assess the success of our marketing and advertising campaigns).

3. Sharing Your Personal Information

We share your Personal Information with third parties to help us use your Personal Information, as described above. For example, we use System.io to power our online store--you can read more about how Systeme.io uses your Personal Information here: https://systeme.io/privacy-policy/ . We may also use Google Analytics to help us understand how our customers use the Site--you can read more about how Google uses your Personal Information here: https://www.google.com/intl/en/policies/privacy/. You can also opt-out of Google Analytics here: https://tools.google.com/dlpage/gaoptout.

Finally, we may also share your Personal Information to comply with applicable laws and regulations, to respond to a subpoena, search warrant or other lawful request for information we receive, or to otherwise protect our rights.

4. Behavioral Advertising

As described above, we use your Personal Information to provide you with targeted advertisements or marketing communications we believe may be of interest to you. For more information about how targeted advertising works, you can visit the Network Advertising Initiative’s (“NAI”) educational page at http://www.networkadvertising.org/understanding-online-advertising/how-does-it-work.

You can opt-out of targeted advertising by:

COMMON LINKS INCLUDE:

FACEBOOK - https://www.facebook.com/settings/?tab=ads

GOOGLE - https://www.google.com/settings/ads/anonymous

BING - https://advertise.bingads.microsoft.com/en-us/resources/policies/personalized-ads]

Additionally, you can opt-out of some of these services by visiting the Digital Advertising Alliance’s opt-out portal at: http://optout.aboutads.info/.

5. Data Retention

Besides your card payment and payment login information, when you place an order through the Site, we will maintain your Order Information for our records unless and until you ask us to delete this information. Example of such information include your first name, last name, email and phone number.

6. Changes

We may update this privacy policy from time to time in order to reflect, for example, changes to our practices or for other operational, legal or regulatory reasons.

7. Contact Us

For more information about our privacy practices, if you have questions, or if you would like to make a complaint, please contact us by e-mail at emmanuel@learnaiwithkesse.com or by mail using the details provided below:

8. Your acceptance of these terms

By using this Site, you signify your acceptance of this policy. If you do not agree to this policy, please do not use our Site. Your continued use of the Site following the posting of changes to this policy will be deemed your acceptance of those changes.

Last Update | 18th August 2024

Save settings
Cookies settings