Robot learning · the reading map

Paper Atlas

Every paper worth keeping, in one place — filed by what it is, wired to what it connects to, and mapped onto the projects and ideas actually being worked on.

Three trees

The whole library sorted into the structure you actually think in. Every branch lists its papers oldest-first, so you can read a line of work in the order it happened. ◆ marks a foundation paper.

Open datasets

The other half of the map. Every figure below is taken from the dataset's own paper, and every project link was checked over the network.

Timeline

The important papers on one vertical spine, newest at the top, each tagged with the methodology it introduced or exemplifies. Pick a methodology to follow just that lineage down the years.

The network

Papers as nodes, intellectual lineage as edges. Drag to pan, scroll to zoom, drag a node to pull the neighbourhood apart, click to open it. Colour is the tree; size is how load-bearing the paper is.

Reinforcement learning Imitation learning Hardware & systems
drag · scroll to zoom · click a node

Projects

What each body of work is reading. Click a project to filter the library down to it.

Research ideas

Take it with you

← jerrycheng.io · UMI catalog · Humanoids · Actuators