This week’s list of top data news highlights covers February 10, 2024, to February 16, 2024, and includes articles on using virtual reality to train surgeons and improving cybersecurity using AI.
1. Reducing Compute Requirements
The Linux Foundation has introduced a new AI model architecture, receptance weighted key value (RMKV), which merges the parallelization capabilities of transformers with the efficient inference of recurrent neural networks. RMKV can reduce the computational demands of training large language models (LLMs), potentially reducing GPU requirements by up to 100 times.
Grace Iordanov, a 17-year-old based in Redmond, Washington, has developed Blink Communicate, an app that leverages computer vision and AI to enable individuals who are paralyzed or only able to communicate through blinking to better express their needs and desires. The technology interprets blinks into predefined phrases and then displays the phrases on the screen.
Google Cloud has partnered with the Environmental Defense Fund to tackle methane emissions, a significant contributor to global warming. The initiative plans to use satellite imagery and artificial intelligence to track emissions back to their sources, enabling companies, researchers, and public sector organizations to identify and curtail emissions.
4. Strengthening Cybersecurity
Google has launched a series of free AI tools to strengthen cybersecurity defenses for organizations. This initiative includes an open-source AI resource for improved malware detection and a white paper on AI’s role in cyber defense.
VirtualiSurg, a French health tech innovator, has blended virtual reality (VR) with medical education to deliver surgical training. Its VR simulator, SurgiSim, allows surgeons to practice procedures in a highly realistic digital environment, enhancing skills safely and efficiently.
Slack has launched Slack AI, an intelligent assistant designed to help navigate communication at work. It helps users efficiently catch up on missed Slack conversations by offering summaries of threads and channels and an enhanced search function.
7. Advancing Materials Science Research
Researchers at the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory have used neural networks to discover new materials that exhibit superior toughness and durability, suitable for various engineering applications. The neural networks enabled the team to test and improve materials through simulation without resorting to expensive trial-and-error methods typically used in material science research.
OpenAI has unveiled Sora, an AI tool capable of converting text prompts into videos up to one minute long. Sora can create complex scenes, intricate character interactions, and precise details in subjects and backgrounds.
The Ohio Department of Transportation is introducing a new traffic warning system, set to be installed at 13 high-risk highway sites, that will alert drivers of upcoming traffic jams using cameras and message boards. This initiative aims to reduce rear-end crashes by at least 16 percent by providing real-time traffic updates to drivers and traffic apps.
Google introduced its new large language model (LLM), Gemini 1.5 Pro, which processes complex inputs ranging from lengthy texts to extensive video and audio content. Gemini 1.5 Pro is designed to tackle intricate tasks like analyzing books with hundreds of pages, summarizing films, and translating uncommon languages.
Image credits: OpenAI
