Author here: I was tech chair for ICCV 2021 which had 6k+ submissions; I had "root access" to the whole process, so I saw (and scrubbed) its barnacly underbelly. I'm now a postdoc-entrepreneur at Mila with Hugo Larochelle and Chris Pal working on meta-science, multi-agent games, and AI for science.
Academic machine learning is changing and needs better policies. We need to get this right because society needs better ways to build consensus and seek truth.
This blogpost shares some meta-science analysis I did on academic ML and a few policy ideas.
My hope is to engage in discussion about academic ML policy. I believe that the only way we make progress on this is to sample and refine ideas together.
martincsweiss•10m ago
Academic machine learning is changing and needs better policies. We need to get this right because society needs better ways to build consensus and seek truth.
This blogpost shares some meta-science analysis I did on academic ML and a few policy ideas.
My hope is to engage in discussion about academic ML policy. I believe that the only way we make progress on this is to sample and refine ideas together.