Dear NIPS board members,

As researchers in machine learning and data science, we write as advocates of a welcoming and diverse machine learning community.

Recently, some disappointing behavior at NIPS 2017 has come to our attention, and has led one lab here to ban participation in NIPS 2018. We understand that steps are being taken to evaluate changes to the meeting in light of this behavior. In addition, the acronym of the conference is prone to unwelcome puns, such as the perhaps subversively named pre-conference “TITS” event and juvenile t-shirts such as “my NIPS are NP-hard”, that add to the hostile environment that many ML researchers have unfortunately been experiencing. 

These incidents have brought up uncomfortable memories for some of us whose career paths have been affected by unwelcoming or harassing behavior.  NIPS is at a time of significant expansion and change.  We believe that serious actions are essential for the continuing health of the community.

We are aware that the NIPS board has appointed Diversity and Inclusion chairs and is working towards a strengthened code of conduct to set behavioral expectations. As you know, templates for such codes of conduct are available from other conferences.  We support this move, and further hope that conference attendees will be actively encouraged to speak up when they observe social behavior that may make other attendees uncomfortable. We also fully expect that this code of conduct will specify consequences for harassment and are in keeping with recent policy changes from NSF that mandate reporting of harassment.

 We also recommend a more radical reform: rebranding of the meeting, including a change to the name of the conference.  Certainly the NIPS name has a long and distinguished history, and it is an unfortunate coincidence that it is vulnerable to sexual puns.  Changing the name at this time could serve as a powerful symbolic move which, in conjunction with other changes, would signal the Board's commitment to improving the culture of the field and making the conference a leader in inclusion.  

We also believe the shoddy survey on the name change is an embarrassment to the AI/ML community. It is not scientifically sound for a number of reasons:(1)  there is a huge sampling bias, the majority of men surveyed are men who have had no issues with the name, and the women are the survivors who have not yet fled the field. (2) Equal weight is given to neutral response (no change) and to the change and hence, false negatives and positives are given same treatment (3) For issues of bias and discrimination of a minority group, the opinion of the majority should not be the basis of decision making.

We appreciate that NIPS has hosted and prominently featured the Women in Machine Learning and Black in AI workshops, and we look forward to further efforts to welcome contributors to the field from a range of backgrounds. We hope that you will continue efforts that add to it rather than subtract from it.

Thank you for your concern and quick action.

PS: This letter is mostly re-purposed from the one that JHU ML researchers sent to the NIPS board. I added the shortcomings of NIPS name change survey based on tweets by Daniela Witten.