domains / families / research / colt-experiments

colt-experiments

separate

Use when deciding whether a COLT (Conference on Learning Theory) paper needs numerical content at all — COLT has no experiments requirement — and, when numerics genuinely help, designing small illustrative simulations that visualize a proved bound, a separation, or a phase transition without diluting the theory.

When to use

• deciding whether to include numerics in a COLT paper
• designing small illustrative simulations
• visualizing a proved bound
• visualizing a separation
• visualizing a phase transition
• creating log-log rate check figures
• plotting separation between models
• plotting threshold phenomena

When not to use

• decorative benchmark tables
• empirical evaluation for non-COLT venues
• compensating for proof gaps
• simulating outside the proved regime without labeling

Skill page

Source