Neurips Disability In Ai Panel Transcript

37 minute read

This is a transcript of the panel discussion at NeurIPS 2019 on disabilities in AI.

Disability in AI

Panel discussion aiming at raising awareness, supporting and advocating for disabled people.

Panelists

Kat Heller

Hi, and welcome to the disability here. So, here we’re gonna hopefully take some questions and have some answers about how things stand in relation to disabilities and figure out maybe what we can do in order to improve the situation with how inclusive we can be towards the disability community. So, let’s start out by talking a little bit about the challenges that people face who are disabled and how far we think we’ve come as a conference and where we might need to go.

Oh, sorry, yes, let’s introduce ourselves and then answer the question. Thank you.

Hugo

I’m Hugo Larochelle and was program chair this year, was because I’m done today, thank you. I happen to be color-blind but I don’t think that’s why I’m on this panel but I guess that’s my little sort of way in which I can relate to some maybe some people here. And I know we paid attention to that this year thanks to Maria in particular. And, yeah, that’s it.

Negar

Hello I’m Negar Rostamzadeh. I’m a research scientist at ElementAI and the next question was about the challenges.

Kat

challenges… and how well we’re doing and where you think we need to go. We could just introduce ourselves.

Emtiyaz

My name is Emtiyaz Khan but you can just call me “MT”, and I work in Tokyo at Rickon Institute and my experience with disability is, I mean… I think it’s limited but my parents, my father had been paralyzed for six years and he just passed away recently so I’ve had some experience and he was in India. So, but it’s in a very different context but I hope I still can understand what we discuss here and give some feedback.

Kat

I’m sorry to hear about that.

Hanna

Hey, I’m Hanna Wallach. I’m the general chair of this year’s conference and my duties are not done, still ongoing. I’m a senior principal researcher at Microsoft. My, let’s see, my connection to disability, I guess, my sister had a spinal cord injury when she was 18, so she uses a wheelchair. I, suddenly out of nowhere, went partially deaf in one year about three years ago, and it never came back and so I actually got my first hearing aid just last week , which is very exciting, so I’m testing it out at the conference. And my mother had a very major stroke a couple of years ago and needs a live-in carer and sort of has a lot of, yeah anyway.

3:01 Constantino

Hi, my name is Constantino Dasklakis. I’m a professor of computer science at MIT. My work is on algorithms and probability and but recently I’ve been interested in learning from bias data and I feel like this is connected to the topic and you know, in general I’m excited about the opportunities that AI research brings to the topic of you know, accessibility and inclusion.

Kat

So what, sorry, going back to the question: what challenges do you think that we have that could be addressed here or through AI algorithms?

3:49 Constantino

Yeah first of all should say, yeah I’m like a newcomer to the community but I can speak more generally. So, accessibility is a major obstacle for people to be included in, you know, doing research and coming to the conference and being included in the, you know, scientific and progress and innovation. So, also like, well disabilities in, let’s say in contrast to other, I think you know, features of humans that cause bias and unfairness is a higher dimensional feature. So, sometimes we don’t measure it properly and sometimes people don’t want that feature which harder to measure and as a result we were facing also a, you know, data collection issue. So, let’s say from outs, you know, to relate to what I was talking about earlier, you know, like when you’re doing statistics if you can have data it opens up an interesting challenge for how to be fair, to treat, you know, to treat, fairly. Suppose you don’t observe it properly. That’s one way. But, of course, there are many more practical ways in which AI can help. And I believe we have seen some ways in this conference that can be also assisted with technology. For example, you know, like you know, like automatic captioning could be a way we could help. You know, like, for people who have like visual impairments or different perception of color, we could combine AI research with video processing, to show content in a way which is more helpful to those people. And I feel like there are many opportunities both more practical and more scientific in which we can help directly in more practical ways but also from a AI bias and fairness perspective, yeah combined statistical tools to help those people.

6:30 Hanna

Okay, I guess maybe I’ll talk a little bit more about sort of the conference itself and sort of some things that we could be doing here to make NeurIPS a bit more inclusive along this dimension. And I should say before I rattle off a long list of things that I think Mary Ellen who is over in the corner over there has done a fantastic job of trying to pull together a lot of this stuff this year working with Kat and working with Maria and so my long list of things should not be taken as a criticism of their hard work but simply these are things that we should be thinking about as we come back to this space next year, as we move into different spaces in the future.

  • So, okay, there’s a bunch of steps to get up onto this stage, that’s not great.
  • There are a bunch of steps to get onto the stage in the main conference center as well for the opening remarks.
  • The microphones are often on a stand that again is not at a particularly easily customizable height.
  • The podiums on the stages again are also at a height that is not, that is potentially problematic.
  • The poster sessions are often very crowded and the posters are quite high up.
  • Poster sessions are for me personally often problematic because I have tinnitus so continual ringing in my partially deaf ear. And then sometimes if I’m in very loud environments, it gets kind of painful, hyperacusis. And so poster sessions can sometimes make that happen. Also, just, it takes a huge amount of just mental energy for me to listen in that kind of a noisy scenario. And so after those poster sessions, I’m then absolutely exhausted and have to sleep for more than usual which is not really super feasible at a conference.
  • Well, I guess we’ve talked a little bit about captioning which would be helpful.
  • I know there have been conversations around making the LaTeX source of papers available and posters available so that those can be sort of presented in different ways that might be more accessible to people who have particular needs.

Let’s see. I could probably list a whole bunch of other things but maybe I’ll stop and hand it over to somebody else.

9:05 Emtiyaz

But, wow, I just have one thing to say actually after all that. So, I actually really think that if we are thinking about inclusiveness here and we’re, you know, for these kind of issues. I mean I said, like I’m here and I’m trying to think about this because I’ve had some personal experiences but it’s so limited that I just don’t feel qualified at all to say, you know, anything meaningful. And I think it’s very important that as a community we listen to other people that we create a platform where we can understand and you know, what is needed, right. So, the kind of thing you just listed. I mean, I would not even think about this and I don’t know where that platform is. It’s not only about disability, it’s also about well-beingness that you know, other things. So I think we really need to do a good job and this has to be one of the, you know, big points of the organization, that you know, we put this thing, so we hear people.

10:05 Hanna

And everybody, literally everybody has different needs and so we need to be listening to everybody to see what it is that we could be doing better about in this regard.

Emtiyaz

Yeah, exactly. So that’s all I would like to say, thanks.

Negar

So, basically, 30% of people under the age of 65 has kind of either disability or chronic illnesses so it’s a huge population and their needs need to be accommodated in different ways because each person as both Hanna and Emtiyaz mentioned has different needs. So one idea, like continuing what MT mentioned, could be like to send a survey, an anonymized survey about kind of disabilities that people have and kind of problems that you’re dealing with and how we can accommodate it based on them and I think that would be very helpful for the idea of like having the conference more accessible for everyone. And the other side is like how to make the whole community more accessible to everyone including people with disabilities. So the fact is that most disabilities are invisible. Disabilities are non-binary so people have sometimes like experiencing some issues but sometimes they are totally fine and they may actually, you may not be able to understand from looking at the person. That like if the person is dealing with some sort of disability at the moment or like is in good health condition so the idea is to listen and trust people who are communicating these needs with you.

Rachel Thomas actually has two really nice articles about disabilities in AI and generally in the tech community.
[The tech industry is failing people with disabilities and chronic illnesses]

And the idea of like disabilities is not binary it’s coming from her article and actually companies who try to include people with disabilities better and accommodate them better have financially more success. There’s a study which was published in Accenture and they actually studied this. Like, I don’t know about the details, though. And the other factor is that again, like in terms of like treating this as a binary measurement about like disability lives. People expect people who have chronic conditions or disabilities when they are in disability lives to be totally off. Like it’s like a binary thing. Like I mean you are in disability, you shouldn’t work. Or, you gave a talk so you can come back to work then. So, these things should be also addressed and I think like that idea of like having maybe a survey and send it to people in the community in an anonymized way could really accommodate these needs and we can then amplify their voices.

13:15 Hanna

I don’t have much to add, except that I was glad to be on this panel largely to listen. So I feel like I am, much like you, not so much aware of the challenges that some of our participants are facing and to me it’s an opportunity to learn more about these so.

13:38 Kat

So, I think one of the things we saw coming into this conference was the trade-off between what we can do as a community and what people are paying for having as and paying for themselves as an individual in order to accommodate their disability. And I guess maybe one of the questions is: Can we get around this idea that you might need to provide something on your own or if you do need to provide something on your own, to what extent should the conference be involved in that providing of something on an individualized level.?

14:22 Constantino

You know, I think that you know, like we live in a data rich era and technological era and I feel like we should be able to provide personalized treatment of people who are diverse so I think as a community we should do best effort to accommodate the, you know, the variation in the attendees and you know as we said earlier like disabilities are not measurable. They’re high dimensional, non-binary, high dimensional and there’s no such thing as, basically it’s variation. So I think we should do our best effort to use our technology and you know, design our spaces and the conference in a way that is as accessible as possible both at a practical level but also as a community to think about you know, how you can use our tools to achieve that.

15:20 Hanna

Yes, so I guess my personal opinion, I’m speaking now as me as opposed to general chair is that yeah, we should as a conference be paying for this. Of course, I say that looking at Marnie the treasurer over there who has way more authority to talk on this one than I do. This year, we put, and Marnie you should correct me if I get stuff wrong. I think it’s 30% of sponsorship money into a fund for sort of diversity and inclusion related things. This is called the, I think it’s the “Next Generation Fund” and I would love to see NeurIPS doing even more of that kind of thing so that people who do have particular needs for themselves we can just help make that happen or make that happen. But this is just my personal view and I’m not you know, speaking on behalf of the entire board but I would love it if we went that route just because it’s a great way to show that we as a conference really care about including people who want to be part of this community and we’re willing to do what we can in our position of power to make that happen.

Yep sorry just to summarize what Marnie said for people who are watching on the livestream and might not have heard that. Marnie said yes!, she is supportive of this but that it would be great to know what people need and that it’s often very difficult to make last minute arrangements especially at such a sort of [inaudible speaking] large conference venue and so people should reach out early to let us know what it is that they need so that we can try to provide it.

17:15 Kat

So my follow-up question was going to be, so what kind of maybe organized mechanisms might we be able to provide for people to do this, right? So my first thought is maybe we could include something in registration where everybody has to list off what kind of accommodation needs they have.

[inaudible speaking]

Yes, so my experience with surveys are that you don’t have… the best way to get the right kind of information is actually through registration, right. Like you can send out a link and ask people to fill it out but then you’re just sort of counting on people being motivated to respond to that like whereas if you ask at registration you definitely get everybody, right.

18:07 Audience member

If I may suggest something. My name is Alfred, I was the organizer for the Well-being in ML social (https://www.eventbrite.com/e/well-being-in-ml-neurips-social-registration-83250988949#) and we found it really useful when we allowed the community to submit stories of their experience regarding their well-being. But it could just as likely be about their disability so I feel like a survey eventually becomes just numbers whereas a story is something tangible, something people can relate to. In our case we also presented, we had this row of posters, so people can actually see real examples of their well-being experience. I feel like this is an opportunity, on top of generating numbers about demographics of disability here. Also, for people to see what it actually feels like. This also goes back to educating people about disability. And really like what we need here is also a cultural change. We shouldn’t focus so much on a technological solution. Just automatic caption or automated color correction. Like we should make people conscious when they make these videos to provide their own captions, to provide the appropriate colors. I don’t think… Because disability is a human problem and we shouldn’t treat it as a technological problem. That’s it.

19:38 Hanna

A hundred percent. That was great.

19:55 Emtiyaz

I think this is an amazing point. I think it’s not really about us. It’s actually about them and they need to be included and we have to do whatever we can to, so we need to be really inviting and we need to create and I think that’s the. So we can’t, we don’t need to be perfect in this. We don’t have to like do things ourselves. I think we shouldn’t be sitting here. Like I shouldn’t be sitting here. I think people who really you know, should be speaking for this and that’s the fact that I’m here I kind of feel like we haven’t done this right yet, so.

20:30 Hanna

And just to draw an analogy, so I in my own research, I do a lot of research on fairness and machine learning and I spend a lot of time telling people. In fact, this is probably 90% of my day job, telling people that fairness is not a technical problem. This is fundamentally socio-technical and if you try and address it using purely technical solutions you will miss most of what’s important. And it won’t actually address the real issues that’s going on there. So I think your comment was very much in line with that and really, really, really applies to this scenario as well.

[inaudible speaking off microphone by audience member]

Of course…tech can definitely be an enabler, I’m just repeating, but it’s not the solution.

21:15 Kat

So talking about maybe solutions and ways to raise awareness, do you guys have any ideas for ways that might be effective at raising awareness like locally where we hold a conference like within the community?

21:35 Negar

I think it again goes back to the fact that like we need to know which kinds of disabilities we are dealing with because it’s very diverse, the need and education are extremely diverse. So each people with disability may experience totally different story. And like I really like the idea of having these stories being told by and I also think it’s part of that it’s the responsibility of people with disability in an anonymized way or like sharing their challenges and ask for a basic. And ask for something that they need but again part of the issue is that sometimes people may not be able to expect it. Like know exactly what they should expect to ask before coming to a conference. Like they may not expect to have a long line in front of the registration desk and not ask it in advance. So that’s why I think it might be a good idea to have maybe a few people with disabilities to like raise some of these points in advance.

Kat

So that’s like forums for communication and like during the conference.

Negar

Yeah.

Hugo

I’m not sure but I saw someone raise their hand. Did you want to, yeah..?

23:00 Audience member

So I was basically gonna say the same thing that you kind of started with. That we have a selection bias problem because everybody who is here, except I didn’t think. I mean the livestream is getting some other people, but people who are too afraid to come because they thought that there would be long lines and they didn’t know what they were getting into. They haven’t been able to see the venue to provide feedback. So I think before registration, right, posting on the website in the conference materials, right, like this is how we expect the conference to be, right. Like the venue is this big, the distance between the East building and the West building is this big. And this is how much time you will have on average to go between them, right. And then say that there are elevators but you know and maybe you know, provide estimates of distances and expected queue times, right. Like just to provide people and I mean I think this would be an estimate but to help them to know what to ask for, as far as accommodation.

24:00 Hanna

That’s great and yeah. Certainly that’s something we can very, in many ways, we can very easily do for next year’s conference, especially having been here in Vancouver once. We can very easily do that.

Kat

I’m sorry, so ok. So I’m also interested in sort of getting back to this issue of how of like how is it that we raise awareness in the rest of the communities sort of like while we’re here. Like posting videos or like posting other such things. Does anyone want to speak to that?

Hanna

I mean, I really like the idea of stories or narratives, or oops, sorry. I really like the idea of stories or narratives or whatever from people in the community talking about their experiences because I think it does sort of, I don’t know, bring a different perspective, or get people to think about things in a different kind of way. So I think that might be something worth trying.

Emtiyaz

And I think we should not wait until the next conference comes in. We should just start this, start like Twitter. Some of these things are happening already. And we should make this like it’s just a thing that is required, right. So lots of things like in Women in ML has been done very well, I think that many of us have been, you know, hearing this thing all the time, so this should just be there.

25:30 Hanna

And actually a few years ago WiML started doing profiles of researchers in the community and that was really cool because you got to know people and you got to know what they were working on and thinking about and stuff like that and so maybe doing profiles here if people want that could be useful as well.

25:55 Constantino

So another idea is that the participants of this conference, you know represent future educators and engineers so it’s really important to emphasize the topic and one, there are many mechanisms that we can use to do that but one very concrete one would be to you know like when you prepare your poster, you get some content accessibility guidelines for how to prepare your poster. That’s really, you know, giving some guidelines that then will make people think about what content they are preparing and how they should make it accessible. But, it’s a broader question. It’s not only about the conference but it’s about those engineers and those educators.

26:45 Negar

Actually something very positive about this year’s NeurIPS was like the fact that during the poster session it was this choice of like having a long stand chair and using it for people who wants to sit and present their poster. I think it was really great.

27:10 Emtiyaz

Sorry, were you finished? The color-blind, these I think some people have been tweeting about this and so like in my own group we had like, we were making the poster, and one of the co-author made the poster with like red and green and then I had to point this out and then it turns out our other co-author is color-blind and so he didn’t even say it until I said it. That you know, that we should make sure that maybe this follows this policy which is written here. And then the other person comes with a smiley face and says, hey you know I can’t tell the difference between the two lines and like you know, see this is very important. I so surprising and I was happy that day.

28:00 Hugo

I do want to give credit to Maria because she early on insisted that we include guidelines which apparently your student didn’t read. But we did, I mean, I think this is a thing we need to emphasize because that’s kind of new for this year but for keynote speakers, oral presenters and also everyone that had a paper with posters, Maria gave us some guidelines and support to include and messaging to contributors. And I think, you know, the more we do it, the more the message gonna spread.

28:30 Kat

Ok, so, I guess one of the things I think about personally is like wouldn’t it be nice to see this affinity group have a workshop like the other affinity groups do? And then what does that entail, right because there’s a lot of like research in the community around different disabilities. Some of the research is done by some researchers who actually have the disability that they are doing research on. So do you want like you know, do you think that’s a good idea? Who do you think should be presenting at this workshop and what do you think they should be presenting on, if it were to happen?

[inaudible speaking]

29:20 Negar

Yeah, actually we organized, like Hugo and some other researchers in Montreal founded Montreal in AI Symposium and we organized it this year and one of the keynote speakers had disability, visual impairment. She actually gave a really great talk about the way, to like, with research, how to address this issue like from technology side.

30:00 Hugo

Yeah, there are people that effectively we could invite and I think knowing more about because that’s a field that I don’t know all that well and it would be like a really original, also just a workshop, period.

Kat

I might personally be one of the people who has something that they are doing research on, so it’s a very self, selfish problem, yes.

Good to know.

30:25 Emtiyaz

I think a very good thing that, sorry maybe this is going a little bit off from what you said but I mean for us to continue doing these things, I think we need to, you know, there’s this word that, we need to bring in the allies. We need to kind of have a network of people, I mean with disability, without disability, it doesn’t matter. But we need to have this thing that happens in machine learning and over time, I think this should be, you know, we should do this. I don’t know who, to try to create this.

31:00 Kat

So I guess I could say following on what I just said, right. So, so I can say, right, what I would like to see happen, like eventually, is of course the goal is like normalize, like disability is in AI and to bring more people here. And I guess one of the fears I could have about what I just said is like, it’s not like you look at me and see anything like super abnormal and so that if I were to say I have this and present technological solutions would that make people feel good or bad? And what do you guys think should be happening?

Negar

So actually like I think one of the challenges with like I mean it’s a really great idea to have this affinity group but one of the challenges is that, most like I’m saying we have 30 percent disability rate but we are not seeing visually this number and most people are not comfortable with sharing in public and there is no need like if they don’t need accommodations, to talk about that. So that’s why I think I don’t know actually what would be the best way solution for that. But it could be also hard to expect everyone to come along for that. Thanks.

32:30 Audience member

So, just in terms of these affinity groups, I think that it would be a good idea to. You know, I would definitely support this if something like this happened. But at the same time I think that one of the things that has made the other affinity groups work so well is that they’ve all been started and led by people from within the community that they’re kind of serving. And so I think it could potentially be problematic if you know, NeurIPS as an organization just decided, we’re going to have this affinity group and everybody with disabilities should come show up here. I think it would be better if this was something where there was kind of the community coming on its own and deciding that this was necessary and making it happen.

33:22 Hanna

I think that’s a really good point.

Kat

So, okay, maybe we should go back to the idea of like collecting better data statistics, right. Because statistics are really not collected well in terms of just sort of like in our community in general what’s going on and just sort of more getting our head wrapped around because the disability rate is really high and we don’t see it. How might we go about getting sort of like a broader view?

Emtiyaz

Like maybe we should just ask people here? Like maybe people in the audience can already speak. Maybe you know more things that we haven’t talked about.

34:24 Joe (audience member)

Hi, my name’s Joe and so mostly you’ve been talking about kind of physical disabilities. My disability is cognitive. In fact, I wouldn’t really call it a disability. It’s just that there are barriers in life and I’ve dyslexia, ADHD, dysgraphia (http://www.ldonline.org/article/12770/). And so my, I’ve learned a lot about advocacy and community as a parent and seeing the challenges my children have had and I don’t think I would have, like applied for their grant to come here without having had that experience. But my big takeaway from that experience is this emerging model of cognitive diversity, right. So this idea that we should have like these very kind of homogeneous skill sets. And certainly for the, you know people like me or my children or my wife is a tutor so we see a lot of these kids. And you know when you look at the IQ profile you see you know these big discrepancies between processing speed and verbal ability or you know spatial ability. So the thing I would like to advocate for is like seeing the talent, you know an understanding that people have a lot to give but often there’s some barriers that stop that. So, you know, being disgraphic, it’s kind of challenging writing papers. Also, you know, my path was I left school at 15 to make video games so I have no formal education and that was not a, in my previous career, making video games, that wasn’t really an issue for me. But now I think I have some gaps and it might be nice to go back to that so the challenge is more for me is like how to access education and you know how to contribute as writing papers. I think that’s it.

36:35 Hanna

I think that’s really important in one area that I worry about with this regard as our reviewing process which has these incredibly tight timelines and very strict requirements on what has to be done when, etc, etc and all this kind of thing. And I am worried that we are potentially missing out on people who might be really good reviewers and bring some interesting opinions to the community. Because there is this sort of, you know, “this week you do this! In three days time you do that. Next day you do that!” All of these kinds of things as well.

37:10 Negar

Thanks for raising this point. I think it’s very interesting and I was thinking about like just coming up with ideas like maybe having like an iterative survey. Like having an initial survey anonymized from people in the community specifying their needs but then also iterating over that and like making those kind of questions or those kind of needs publicly, then other people can look through them and see if they also feel the same kind of need and then we can get to maybe a better measurements of like maybe I mean someone is not thinking about a specific issue but someone else is bringing that up and say “oh yeah, that’s a very interesting point and like I’m also facing the same problem.”

38:30 Emtiyaz

Yeah, with this survey thing I am worried that we might have just one data point per issue, you know. It’s just very hard to kind of do this. I mean the story idea is great and then we have to understand what’s required. And I think it’s just getting the overall sense of you know, what is required to be done and how we can make people feel that we care about this thing and we want, you know such things to be discussed here.

[inaudible speaking]

Audience member

… this dyscalculia, is that what it’s called, I think? They’re, a lot of them are one a spectrum, right, so like by lumping them all together in one category, you’re missing out on a lot of the high dimensional information. And sometimes in combinations thereof, that need to be looked at in a different way.

39:15 audience member

So plugging data and getting information is very important but sometimes like people don’t respond to these surveys, right, so we end up with crappy data. So, let’s ask the question “why do people not want to respond to these surveys and that is like, I don’t know what is the expected reward out that, right. And it’s the similar thing with disclosure. So I also happen to have one of these like you know, hard-to-see disabilities and practically in the draw process I have the option for like asking for accommodations but I have no idea if the expected gain from providing the accommodations are more likely to help or the information I, you know, leads or is likely to hurt, right. So, in these settings like I very frequently decide not to disclose and I very firmly like get it’s cut at the lead stage of the interview . And the lead code is just like a programming test where it’s very difficult to succeed if you have you know certain issues. So yeah, that is basically my information and I think one way to like go around this and kind of like what you said to show people that we actually care about this and we’re actually trying to fix the problem is to provide resources, right. So providing scholarships or like opportunities to participate in research specifically if we see that there’s something to be gained that’s a signal that you know this is legitimately something you care about and not just a diversity flag, you know purchasing those things. But also you’re more likely to get better higher quality data which will kind of like improve the problem.

40:48 Hanna

Yeah, and I think this goes back to what you were saying, Kat, about incorporating this into registration and thus making that link to resources and accommodations that can be made by NeurIPS for people really explicit so then the reason why you’re disclosing this is not just for some miscellaneous nebulous reason, who knows why. But it’s that you are disclosing this so that you can get access to those resources that NeurIPS can and wants to provide.

Emtiyaz

Yeah, so it might be something like sorry. It might be something like travel awards for example right. You could say that if you indicate this then you’re more likely, you make this case about There could be these kinds of things that can be made.

41:40 Audience Member

Thank you. First of all, thank you for putting together this panel. This is really great. First step is obviously talking about it. I know that from past experience working for museums, curatorial sort of stuff, a very low hanging fruit which shows that you’re more committed than just saying like diversity flag stuff is having more visible minority accessibility. So for like the keynotes, having like speech impaired, sorry the speech impaired, hearing impaired like sign language access is a really great way to say like “hey we actually care about this. We’ve hired someone who is going to translate all these things.” It’s just a really easy low-hanging fruit. I’m also curious, as to another thing that like museums actively do is reach out to activist groups who have experience trying to reach out to community and providing recommendations and saying don’t do that. That’s a great way to alienate everybody. Do this. This is a great way to like make people feel included. And not to knock all of us in this room, but I feel like going out to a panel like that will probably be a much better use of money and lead to like better long-term engagement for these sorts of things, so that’s all I had to say. I’m questioning like whether you have all thought about doing that, like reaching out to a panel like that.

43:00 Hanna

I think that’s really important and it relates to Jen’s comment earlier about this not gonna be successful if it’s a bunch of people from, you know, the NeurIPS foundation saying “we’ve decided to do this thing for people!” So I think that’s a really good point and definitely something that is worth doing as we move forward. As for, I think we did have some kind of interpreter service this year so maybe you want to talk about that, Cat.

43:35 Kat

Yeah, I don’t know a lot about it but I know that there were interpreters who were hired for various things. They were there for a limited number of things. We tried to ask particular people who needed them where they were going to be and where they needed it. And then other people who had actually gotten interpretive services also, like separately, also offered their services during the things that they were attending but it seemed very much like the people who needed the services, the services were really there for those people at the things that they said that they were going to so I think getting a hold of exactly who needs the services and where the services are needed are important if we’re gonna have sort of, a small, a small amount of interpretation, yeah.

44:35 audience member

Right, so this was the first year that we heard from the, or I on the board heard from the community about the need for sign language interpretation. It seems obvious but we were trying to figure out how to provide this service in the best way possible. And how to respond and so we were looking for local sign language interpretation. We found Vancouver, it turned out to be Vancouver United States and not Vancouver Canada and so we, you know, we quickly regrouped and found somebody local. I don’t know how well that worked out. We also had the closed captioning in the talks but it wasn’t immediately obvious how to get that closed captioning and to how to turn it on. I don’t know how well that turned out. So, looking for feedback because this year was a beta test year and we want to make it better next year because we do need to provide that kind of accessibility for hearing impaired. And also for other kinds of issues, we need to provide that and find out, you know, what worked, what didn’t, and what we need to do.

46:10 audience member

So, actually, one of my PhD students is one of the ones who has sign language interpreters and I’m sure they would be happy to give you guys feedback if you reached out to them and I’m sure they would be more than happy to tell you about their experience. So that would, I think be a great thing to do. Another thing, I just I actually want to add this before. And so one thing that I think people should be, it sounded a little bit like people were saying “well, if you check off a box that you have a certain disability then you can apply for a certain scholarship or something like that based on your disability.” I’m not sure that I think, you have to be a little bit careful about that kind of thing. Some people might feel like okay about that. Other people might feel that, well, you know the conference should be inclusive. You shouldn’t have to write an application for it to be included, you should be included by default. And so I think, just about how it might come across to somebody who feels like, do they have to do something special, or are they part of the community without having to do something special.

47:15 Hanna

Absolutely. And framing and phrasing is going to be very always important. But it’s important to take that kind of thing really seriously.

47:24 Kat

I also think maybe we can do more, I don’t know what the cost is, but we can do more like as a community to make things global instead of so dependent on individual resources. Either through the conference or through their university or whatever, right. So for example, I’ve definitely seen situations in which talks were both closed captioned and there was a sign language interpreter standing there right next to it so we’ve got four tracks and to just do that for every track. Like every track in every track. I think it’s something that maybe should be looked into. I know there were also issues for students in the same situation at the poster sessions and maybe we could like also try to think globally about what solutions to that might be.

48:10 audience member

So I think this lady asked for data on the closed captioning. I think the quality probably could use improvements so when it was livestreaming a lot of the captioning was actually incorrect. And I know because I use it, you know. And so I don’t know what the technology is, maybe it’s not like machine learnable yet, but maybe you can have like a you know, a first pass that’s a little bit lower quality but then correct it later because people would definitely benefit from that.

49:00 audience member

So, I want to make sure we’re aware that there is like American Disabilities Act for those in the US and even some states have their own individual rules that are even more stringent. Canada also has very specific rules about those with disabilities. One of the things we can do without actually surveying our population is recognize that we do in fact know information about what types of things are considered disabilities already look to see what those address. Most of our employers are having to actually already look at these things for employees in the first place so rather than wait for somebody to identify. We should be looking to those things that we already know that we can do or probably should be doing because we’re not employees, but we’re members of a community just like employees or members of an organization. So from that perspective, if we already look at what those things are and then it’s by publicizing and making it clear to our communities at large that these things are available, it might then work. Again, the whole thing about the make sure they know about the dimensions of how long it takes to get someplace. They have more information. They can then make determinations whether or not those that would be, maybe it’s streaming, might actually come because they know there’s actually opportunity for them to participate in an active way because we’ve already set the stage for them. So I’m just saying one of the things that we can actually do, as a body, is actually go ahead and actively look at those things and put those into place and then communicate openly and often that those things are in place. Just two cents.

50:40 Kat

Okay, so I think that we’re over an hour so I will close it out. Thanks for coming. I really appreciate it. And thank you to all of our panelists who did a great job.

[applause]

References

Updated:

Leave a Comment

You May Also Enjoy

Psf Fellow

less than 1 minute read

2022 PSF Fellow Announcement