This article discusses a study where a group of scientists used algorithms of trained artificial intelligence to identify a new antibiotic abaucin, which destroys a deadly type of superbugs.
This article discusses the introduction of a multimodal neural network ImageBind by Meta, which can process 6 types of data at once, including text, pictures, sounds, depth maps, heat maps, and IMU signals.
This article discusses an experiment where advertising created by neural networks was able to deceive marketing experts and was almost indistinguishable from human-created content. The experiment measured the creativity of texts and the potential to stimulate emotional reactions.
This article discusses MinD-Video, a neural network model created by researchers from the universities of Hong Kong and Singapore that creates a high-quality reconstruction of images seen by a person based on the activity of their brain.
This article discusses the collaboration of several neural networks to generate a short sci-fi film. The text-to-video model Gen-2 was responsible for the visual, the script was written in collaboration with the chat bot ChatGPT, the voiceover was synthesized by a text-to-speech algorithm from the ElevenLabs neural network, and background music was created with Mubert artificial intelligence.
This article discusses the work of artist Chris Kashtanova, who used several neural networks, including ChatGPT-4, GMT-4, Midjourney, Kaiser AI, Maubert, and RunwayML, to generate an image, video, and soundtrack of the human soul.