| 1 | Jun 10, 2010 7:05 PM | No opinion. All past year's ICML reviews have been fine with me. |
| 2 | Jun 10, 2010 7:06 PM | In general the first two options are fine but there is one very big caveat: there should be the ability for the authors to write in their own keywords. A certain (newish) topic did not have a keyword, and so papers on it keep getting reviewer mismatch. |
| 3 | Jun 10, 2010 7:24 PM | Follow the nips strategy this year. |
| 4 | Jun 10, 2010 11:27 PM | This year's model without assignment of areas being public. |
| 5 | Jun 10, 2010 11:36 PM | Why not use machine learning to assign papers? |
| 6 | Jun 12, 2010 12:09 AM | I suggest to assign two area chairs for each paper so in the second phase they need to discuss and reach a consensus on selection the additional reviewers. |