The Map
The map is a 2D visualization of the entire ML paper landscape. Each circle is a paper. Papers that are close together share similar content, topics, or themes.
Position
A paper's position is determined by its content similarity to other papers. Clusters naturally form around topics — you'll see regions for computer vision,natural language processing, reinforcement learning, and more. Region labels appear as you zoom in.
Color
Each paper is colored by its community in the citation network. Papers of the same color are more closely related through citation pathways. There are 20 distinct communities:
A paper's position reflects its content while its color reflects its citation community. Sometimes a paper appears in one topic cluster but has the color of another — this means it bridges two research areas.
Navigation
| Action | Desktop | Mobile |
|---|---|---|
| Zoom in/out | Scroll wheel | Pinch |
| Pan | Click + drag | Drag |
| Select paper | Click | Tap |
You can also use the zoom controls in the bottom-right corner of the map.