Close Menu
Versa AI hub
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

What's Hot

The Secretary of the Ministry of Information will attend the closure of the AI ​​Media Content Training Program

May 18, 2025

Building cost-effective enterprise RAG applications using Intel Gaudi 2 and Intel Xeon

May 18, 2025

Firecrawl is recruiting AI agents to transform content creation

May 17, 2025
Facebook X (Twitter) Instagram
Versa AI hubVersa AI hub
Sunday, May 18
Facebook X (Twitter) Instagram
Login
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools
Versa AI hub
Home»Research»Meta introduces OMOL25 and UMA, a new open AI tool for molecular research
Research

Meta introduces OMOL25 and UMA, a new open AI tool for molecular research

versatileaiBy versatileaiMay 15, 2025No Comments3 Mins Read
Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
#image_title
Share
Facebook Twitter LinkedIn Pinterest Email


summary

META has released OMOL25, the largest open data set of AI-driven chemistry to date, and introduced UMA, a universal AI model designed to predict the chemical properties of molecules and materials.

OMOL25 contains data from over 100 million high-precision molecular calculations. This exceeds the size of the previous open dataset in the field. According to Meta, it took over 6 billion hours of supercomputing time to generate the dataset. The dataset covers a wide range of molecules, ranging from small organic compounds and biomolecules (such as protein fragments and DNA segments) to metal complexes and electrolytes. It also includes charged and spin states, multiple conformations (spatial arrangements), and chemical reactions.

The goal is to help AI models learn how molecules work in a wide range of scenarios. OMOL25 can be used openly for drug discovery, battery material development, and Catalyst Research. In addition to energy and force values, the dataset includes additional details such as charge distribution, orbitals, and other chemical properties.

The OMOL25 dataset is available for embracing your face.

advertisement

Decoder Newsletter

The most important AI news is news directly in your inbox.

✓Every week

✓Free

✓ Cancel anytime

UMA: Universal AI Models of Molecules and Materials

Alongside the OMOL25, Meta is launching UMA (Atoms Universal Model), a new AI model trained on OMOL25 and other datasets. UMA can predict chemical properties at the atomic level, and according to Meta, it is done much faster than traditional approaches.

Unlike previous methods that require specialized models for each task, UMA handles a wide range of applications, from molecular simulations (such as drug discovery) to materials and catalytic research. Built on modern graph neural networks, it combines speed and precision using a “mixed linear expert” architecture. In benchmark tests, UMA achieved results previously only possible with specially tweaked models.

UMA allows you to complete simulations and calculations that once took several days in seconds. Meta says researchers can be screened for thousands of new molecules for potential as drugs and battery materials before lab integration is needed.

The UMA model can be used with a face hug.

A new way to speed AI simulation

Traditional AI models typically require a large amount of training data to generate new molecular structures. Meta is currently introducing a new method called “Admint Sampling,” which allows AI models to learn and propose new structures even when no real examples exist.

recommendation

Consistency and illogical: Studying research reveals unstable inferences in AI language models

Consistency and illogical: Studying research reveals unstable inferences in AI language models

The technique is based on concepts from stochastic control theory, and Meta’s team states that it is particularly suitable for simulating molecules. Adjunctive sampling allows rapid search of many structural variants in several calculations.

In early testing, this method was able to generate molecular structures that are not only consistent but often outperform classic software performance, especially for molecules with many flexible components.

Models, code, and other information is available through Face and Github embrace.

Despite these advances, Meta points out that there are still open challenges. Some areas of chemistry, such as polymers, specific metals, or complex protonation states, are not yet fully covered. AI models also need to improve their ability to predict charge, spin, and long-range interactions.

author avatar
versatileai
See Full Bio
Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
Previous ArticleNSFWlover announces game-changing NSFW AI images
Next Article ALLSCI opens access to AI-powered Discovery platforms
versatileai

Related Posts

Research

UC awards $18 million to expand AI’s ambitions and impact in science

May 15, 2025
Research

Morocco advances AI in transit with new research partnership Morocco advances AI in transit with new research partnership

May 15, 2025
Research

Using AI, researchers predict the location of virtually any protein within human cells | MIT News

May 15, 2025
Add A Comment
Leave A Reply Cancel Reply

Top Posts

The UAE announces bold AI-led plans to revolutionize the law

April 22, 20253 Views

The UAE will use artificial intelligence to develop new laws

April 22, 20253 Views

New report on national security risks from weakened AI safety frameworks

April 22, 20253 Views
Stay In Touch
  • YouTube
  • TikTok
  • Twitter
  • Instagram
  • Threads
Latest Reviews

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

Most Popular

The UAE announces bold AI-led plans to revolutionize the law

April 22, 20253 Views

The UAE will use artificial intelligence to develop new laws

April 22, 20253 Views

New report on national security risks from weakened AI safety frameworks

April 22, 20253 Views
Don't Miss

The Secretary of the Ministry of Information will attend the closure of the AI ​​Media Content Training Program

May 18, 2025

Building cost-effective enterprise RAG applications using Intel Gaudi 2 and Intel Xeon

May 18, 2025

Firecrawl is recruiting AI agents to transform content creation

May 17, 2025
Service Area
X (Twitter) Instagram YouTube TikTok Threads RSS
  • About Us
  • Contact Us
  • Privacy Policy
  • Terms and Conditions
  • Disclaimer
© 2025 Versa AI Hub. All Rights Reserved.

Type above and press Enter to search. Press Esc to cancel.

Sign In or Register

Welcome Back!

Login to your account below.

Lost password?