60,000 Tree Roots Grow... In 3 Seconds!

Did you know that there are unbelievable research papers on how to simulate the life of virtual trees? We can write a computer program to simulate how they move by interacting with them with a mouse. If we go too hard, we can simulate how they break. And if we feel like it, we can even burn a tree down. Or simulate not one burning tree, but even an entire forest. Dear fellow scholars, this is Two Minute Papers with Dr. Zsolnai. Now for a more peaceful simulation. We can simulate the life of complete ecosystems for a thousand years as well. This is a story of human ingenuity. Yes, all of these were handcrafted techniques. No AI of any sort was used to make this. And I have good news for you, fellow scholarship. We even have more papers in this area. And I have found some new ones that are beyond amazing. For instance, the previous works showed what happens above the ground. But a great deal of a tree's life also happens underground. As this poplar tree develops over time, so do its roots as well. Now take a careful look at the shape of the roots. And now have a look at this pine tree growing. Wow. The poplar went vertical. The pine went horizontal. Now, why does this happen? Well, poplar trees are rapid growing but relatively short lived trees. And they primarily grow in areas where water is deeper down in the soil. And with these roots, they can reach them. Now, the pine tree sees the world entirely differently. Its roots are horizontal, thus maximizing water and nutrient absorption near the surface. As an additional bonus, it gets stabilized in unfriendly environments where the soil is looser. Very smart. Now, what about the oak? The oak is a wise tree. And what does a wise tree do? Of course, it diversifies its beds. Its deep taproot system gives it access to deep water sources. But it also has widespread horizontal roots for maximum nutrient absorption, thus giving it great longevity. Great choice. Hmm. I feel a bit like David Attenborough here. However, the story does not end here. A good tree root simulation has to take into consideration how smart and adaptable these little plants are. You see, environmental stimulus matters a great deal in their development. For instance, when growing around the slope of a hill, it will need a different kind of root system to survive. Loving it. So, is that it? Is that all this simulator can take into consideration? Well, not quite. Now hold on to your papers and look at this, fellow scholars. Holy mother of papers. It simulates the effect of all of these parameters, and this is not even a complete list. For instance, it simulates how different kinds of soils exist and show a different growth resistance factor. And we can Also mix them up so it can simulate heterogeneous systems. What does that mean? Well, multiple soil types scattered across this virtual world. Absolutely beautiful. And the tree will grow as a tree should grow in reality, Hopefully. And as we change the distribution of soil types, so does this wise little tree adapt to its environment. It even learns to avoid regions of high growth resistance. And it grows around it, thus simulating asymmetric growth. So good. And now I wonder, how much computation does such a simulation take? I mean, we need to simulate these virtual trees for years and years. So does it take years and years of computation time in real life? Look at that. Not at all, fellow scholars. We can simulate such a system on the scale of months in, in a matter of seconds, simulating 60,000 little routes in less than three seconds and sometimes even interactively. Wow. Now, at this point I have one more question left. You know what it is? Are these simulations for show or is this how reality works? Well, let's see about that. When compared to a previous paper that digitizes tree root systems in real life, the this is the real deal. Now, does the simulator live up to these results? Look at that. My goodness. Clean Sweep. Loving it. Not only the kinds of structures are very similar, but I actually have to look real close to find the differences. It needn't be perfect, and yet it almost is. Wow. And it gets better. The root and shoot ratios for real trees are known, and the simulated oaks properties are very similar to the ones seen in real trees. And now imagine that the games and virtual worlds of the future could be with simulations that are this realistic. We could play in these worlds in the future. What a time to be alive. And once again, all this was done through sheer human ingenuity. No AI technique was used in this paper. And the paper contains so much more analysis. You simply have to take a look. A beautiful piece of work. Of course it is, because this is a Siggraph paper. What is that? Siggraph is the most sacred place for some of the best computer graphics papers ever. Getting a paper accepted there is, in my opinion, like winning the Olympic gold medal of graphics research. Huge congratulations. However, computer graphics is a field that is orders of magnitude smaller than than AI research. For instance, look at this. Goodness. This paper was seen by about 200 fellow scholars. And I fear that if we don't talk about works like this here on two minute papers, almost no one will. Here I try to be a good ambassador of computer graphics research and show everyone how amazing these papers are and can give to the world so much and it breaks my heart that so few people know about them. But good news, I have more papers like this up the sleeve if you wish to not miss out on them when they come in the future, make sure to subscribe and hit the bell Icon Experiment tracking, model evaluation and production monitoring for your deep learning projects and LLM apps. This is what weights and biases does and it is the best everyone is using it. Try it out now at WNB Me Papers or click the link in the description below.