Interactive visualization shows a virtual creature learning to chase light with reinforcement learning
What happened
On November 14, a Hacker News front-page post featured an interactive reinforcement learning visualization. It uses a virtual creature called spark with only four neurons and three adjustable parameters to show the learning process. Spark randomly nudges its internal connections: if performance improves it keeps adjusting that way, if it worsens it reverses. In the demo, spark first spins in place because of random wiring, starts moving after 50 attempts, learns to avoid light at 250, rushes toward the light source at 500, and becomes a light-chasing expert after 2,000. The post plots this as a learning path in "brain parameter space" and shows hundreds of individuals with different initial wiring converging on the same goal.
Written by AI from the coverage · updated 2 days ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageA visual explanation of reinforcement learning
一篇交互式可视化文章用只有四个神经元、三个可调参数的虚拟生物 spark 演示强化学习:随机微调脑内连接,表现变好就沿该方向调整、变差则反向调整。spark 从随机接线时原地打转,经 50 次尝试开始移动、250 次学会避光、500 次冲向光源、2000 次成为追光专家。文章把这一过程画成“大脑参数空间”中的学习路径,显示数百个初始接线不同的个体最终都收敛到同一目标。
Heat over time
Heat now 3·Comparable peak 9(Nov 15 01:00)·Comparable change over 24 hours -51%
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.