during the past 2 weeks, i have been experimenting with using neuroevolution to make the AI agents explore the grid, by keeping track of what grid cells the AI agent's raycasts hit, and giving them a score proportional to the ratio of total grid cell ... read more →

After my last reya-related devlog, i have made the decision to ditch training data entirely and use either: imitation reinforcement training or neuroevolution i have decided to use option 2 for now, although it may be limited for the complex task ... read more →

i am still in the first step of the project (not to say that it is minor, it arguably is 80% of the project), which is spatial memory. I had just began focusing my effort on it a few days ago, most of the last 2 weeks were focused on technical detail ... read more →

I have disappeared for about a month, largely because I was busy adjusting to another bloody school year, but also because I spent most of my free time reading dozens (I may have surpassed a hundred) of cutting-edge research papers on neuromorphic co ... read more →

I wrote a lua FFI for thimni, called limni. It is not nearly finished, as it lacks support for a few helper functionality and is restricted to 2D and 3D. It is a very simple binding (only about 250 LOC). I wrote a little example for it in LOVE. I ini ... read more →

REPO: https://codeberg.org/0x177/LSNN I have been thinking about ways to simulate spiking neural networks besides the typical implementations (particularly custom hardware) and found it very inconvenient that usually spiking neural networks allow for ... read more →

I have been working on implementing block-based optimizations for rqism, that divide the statevector into blocks depending on the type of gate to efficiently implement that gate directly, avoiding the complex linear algebra of quantum gate applicatio ... read more →

i hacked together a small demo for thimni. the player is a capsule shape that can shoot bombs to destruct the fractal controls: WASD: forward, right, left, backward space: jump left click: shoot bomb downloads linux: https://drive.proton.me/urls/0B ... read more →

a system where each cell tries to clamp its local entropy in a range. got the idea while working on my AdS/CFT emergent gravity simulation (WIP, unpublished as of the time i am writing this, probably never finishing it :3) algorithm i am losing my sa ... read more →

I wrote an implementation of collision with a menger sponge using thimni and it works surprisingly well! I added a benchmark for them and 100 random menger sponges colliding takes an average of 100 milliseconds. here is a link to the result read more →