NeurIPS 2020

Node Embeddings and Exact Low-Rank Representations of Complex Networks

Meta Review

The referees found this to be a very well-written paper with interesting and valuable insights (both methodological and algorithmic) for low-rank embeddings in graphs. The proposed low-rank embedding function is also simple to interpret and opens the door for future research. The only outstanding question is how well these types of embeddings perform on downstream tasks not involving graph reconstruction, but the authors appropriately discuss this as an open direction. This paper should be accepted.