Learn how Meta is making its open-source Language Model for Multi-Domain Applications (LLaMA) available to developers worldwide despite lawmaker inquiries. Read more here.
Meta's I-JEPA uses self-supervised learning to learn from images like humans and animals, without needing labeled training data. Tests show better performance with less computing resources. Open-sourced for non-commercial use.
Discover how Facebook-owner Meta is moving beyond generative AI to create machines that can conceptualize abstract ideas and replicate human rationality.
University of Kansas scientists have developed an AI detection tool with 99% accuracy, able to identify scientific texts generated by AI such as ChatGPT. This tool could defend scientific integrity by detecting when AI-generated text is presented as work from a human mind. The tool was specially designed to recognize the difference between AI and human writing, focusing on scientific writings in peer-reviewed journals generating better accuracy than general-purpose detectors. The team made their AI-detecting code accessible to researchers emphasizing the great opportunities for people to get involved in AI and make useful contributions.
Explore the evolution of tech policy from Obama's optimism to Harris's vision at the Democratic National Convention. What's next for Democrats in tech?