Realistic 3D heads, Open-source beats GPT, AI designs cars & clothes, Acrobatic robots, Movable 3D models
Welcome to the AI Search newsletter. Here are the top highlights in AI this week.
DeepSeek V3 is a new open source AI model with 685 billion parameters that excels in tasks like coding and math. This is a strong competitor against established models such as GPT-4o and Claude 3.5. DeepSeek V3 showcases high performance while being cost-effective to develop. Read more
Do you prefer to watch instead of read? Check out this video covering all the highlights in AI this week:
PartGen is a new AI tool that generates 3D objects by breaking them down into meaningful parts, making it easier to create and edit 3D models. It uses advanced techniques to analyze images or text inputs, segmenting the objects into parts and then reconstructing them in 3D, even filling in missing details. Read more
Automating animation colorization just got easier! AniDoc, a new AI model, can automatically colorize line art videos using a single reference image, making animation production more efficient. It can handle varying sketches, sparse input, and multiple characters, and adapts to different styles and backgrounds. Read more
uPix is an AI Selfie Generator that allows users to turn into anyone in just one click. Select from a vast array of templates, ranging from superheroes to business portraits, and even anime characters. Try it out today!
Articulate-Anything is a system that can automatically create interactive 3D objects from text, images, or videos. It uses Vision-Language Models to generate Python programs that can be used to create digital twins of objects, allowing for more realistic simulations and robotic training. This system can be used in various applications such as AR/VR, animations, and robotics. Read more
Researchers have developed a robot that can give massages using traditional Chinese medicine techniques. The robot uses four different modes to mimic the movements of a human massage therapist, including palm punches, vibrations, kneading, and finger techniques. Read more
Scientists used machine learning to develop perovskite solar cells with near-record efficiency. They created an algorithm that helped them find new materials to improve the efficiency of solar cells. The algorithm was trained on a dataset of molecules and helped the researchers discover a new material that resulted in solar cells with efficiencies as high as 26.2%. Read more
xAlerts is a powerful tool that helps you track the activity of your favorite accounts. Stay informed from the latest activities of investors, celebrities, influencers, athletes, and more. Try it for free!
Self-supervised machine learning can adapt to new tasks without retraining. This approach, called Contextual Self-Supervised Learning (ContextSSL), uses context to dynamically adapt representations to be invariant or equivariant to different transformations, eliminating the need for repetitive retraining. Read more
ChatGarment is an AI system that can create, edit, and simulate 3D garments from images or text descriptions. It uses a large vision-language model to understand both images and language, allowing it to generate sewing patterns and create 3D garments that can be animated and simulated. This technology has the potential to revolutionize workflows in fashion and gaming applications. Read more
FaceLift is a new AI system that can take a single photo of a person's face and generate a 3D model of their entire head. It uses a multi-view latent diffusion model to generate novel views of the head, and then uses a GS-LRM reconstructor to create a 3D Gaussian representation. This allows for high-quality novel view synthesis and 3D facial animation. Read more
Researchers have developed a new method called HSfM that can reconstruct 3D humans, scenes, and cameras from sparse uncalibrated images. HSfM combines human mesh recovery and structure from motion techniques to jointly estimate 3D human pose and shape, scene point maps, and camera poses. This method enables capturing people interacting with their environment and improves camera pose and scene reconstruction. Read more