drag to pan · scroll to zoom
Loss vs. step (log₁₀)

Learned Optimizer Race: 2D Contour View

Meta-learning is not only about learning a good starting point (MAML, Reptile) or a good embedding space (Prototypical Networks) — it can also learn the optimizer itself. This top-down contour-map simulator drops three update rules onto the same randomly-oriented, ill-conditioned quadratic loss bowl: plain SGD, momentum SGD, and a per-coordinate adaptive rule that mirrors what "learning to learn by gradient descent" research found a meta-trained optimizer converges to. Drag the condition-number slider to make the valley narrower and watch hand-designed SGD zig-zag while the learned rule drives straight down; pan and zoom over the contour map, toggle the gradient-descent field, and watch the live loss-vs-step chart update in real time.