Between Linear and Sinusoidal: Rethinking the Time Encoder in Dynamic Graph Learning

Open in new window