# Computer graphics Published articles for Computer graphics. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## GeoPT helps AI models simulate how objects respond to physical forces DevFeed: [GeoPT helps AI models simulate how objects respond to physical forces]() Original publisher: [Read original article]() Author: Alex Shipps | MIT CSAIL Published: 2026-08-10T19:25:00Z Content type: news Language: en Sources: [MIT AI News]() Topics: [AI Models](), [Simulation](), [AI Research](), [Computer Science and Artificial Intelligence Laboratory (CSAIL)](), [Robotics]() Tags: [3-d-imaging](), [ai](), [ai-models](), [artificial-intelligence](), [computational-fluid-dynamics-cfd](), [computer-graphics](), [computer-modeling](), [computer-science-and-artificial-intelligence-laboratory-csail](), [computer-science-and-technology](), [crash-simulation](), [design](), [electrical-engineering-and-computer-science-eecs](), [geometric-pre-training](), [geopt](), [haixu-wu](), [human-computer-interaction](), [kaiming-he](), [manufacturing](), [minghao-guo](), [mit-csail](), [mit-eecs](), [mit-schwarzman-college-of-computing](), [neural-physics-simulation](), [paper](), [physics](), [physics-aware-ai](), [physics-foundation-models](), [research](), [robotics](), [school-of-engineering](), [self-supervised-learning](), [simulation](), [surrogate-modeling](), [synthetic-dynamics](), [transformer-based-simulators](), [wojciech-matusik]() ### AI overview Researchers at MIT CSAIL and Tsinghua University developed GeoPT, a pre-training approach that uses 3D simulations of mechanical interactions to help AI models learn physics more efficiently. The article reports that models using the approach reached peak performance twice as fast and trained on up to 60 percent less data than leading models. ### Source excerpt "GeoPT" helps AI models understand the basics of physics so they can simulate how objects respond to things like wind and water more efficiently and accurately. ## Metaballs DevFeed: [Metaballs]() Original publisher: [Read original article]() Author: Stanko Published: 2020-02-06T00:00:00Z Content type: article Language: en Sources: [Stanko Tadić]() Topics: [generative art](), [creative coding](), [Graphics](), [coding](), [Algorithm](), [implementation](), [render]() Tags: [algorithm](), [art](), [coding](), [computer-graphics](), [creative-coding](), [exploration](), [generative](), [generative-art](), [graphics](), [implementation](), [projects](), [render]() ### AI overview A project write-up about creating Metaballs, a generative art project that combines spheres into organic shapes, slices them, and renders their outlines as vector files. The author describes exploring metaballs, learning about marching cubes and marching squares, using an existing solution, and building tooling for immediate visual feedback. ### Source excerpt I'm all about generative art these days. Today, I present you Metaballs, my latest creation. It is one of those projects that I visualized in my head and managed to pull it off exactly how I imagined it. Somehow I came up with an idea to generate spheres, smoothly merge them to get organic looking shapes, which then I would slice up and render the outlines.I ordered a pen plotter and that is the reason I'm focused on producing vector files. One of the final images looks like this (click on it to see more): Exploration # Not sure how I started thinking about slicing these blobby objects, but I had a clear vision of what I wanted to achieve. At this point I had no idea they are called metaballs. I asked on our company's #tech-creative-coding channel, if anyone is familiar with making something like this, and my colleagues pointed out the term metaballs and marching cubes algorithm. Definition # Wikipedia defines metaballs as: In computer graphics, metaballs are organic-looking n-dimensional isosurfaces, characterised by their ability to meld together when in close proximity to create single, contiguous objects. Which makes them perfect for what I had in mind - generating organic looking blobs. Implementation # The first try # Initially I wanted to implement it myself, but I didn't really know where to start. I had an idea to hack it by overlapping two spheres and calculate a curve between them. By rotating this curve, I would create a tube and merge everything together. It looks ok-ish for two spheres, but I wasn't completely satisfied. And it was impossible do it for more than two spheres. So I gave up early on this idea. Marching on # So I went back to the marching cubes algorithm. After some research I found a wonderful post, which very thoroughly explains implementation of marching squares.Marching squares are the same algorithm as marching cubes, but in 2D. Now I understood how algorithm works, but I gave up on implementing it myself. I was more interested in bei ## Math Dodger: A Flash game DevFeed: [Math Dodger: A Flash game]() Original publisher: [Read original article]() Author: Graham King Published: 2009-03-15T01:26:28Z Content type: tutorial Language: en Sources: [Graham King]() Topics: [math](), [ActionScript](), [Graphics](), [Code](), [Tutorial]() Tags: [code](), [computer-graphics](), [download](), [flash](), [game](), [math](), [misc](), [sdk](), [software](), [source]() ### AI overview A Flash game challenges players to steer a tank around mathematical blobs in Los Angeles. The article explains the controls, provides source code under the GNU General Public Licence, and notes that the game requires the Flex SDK and standalone Flash Player to build and run. ### Source excerpt "Dude, like, totally dodge those math blobs in L.A.!"