Summary
In this episode, Dr. Bryce Meredig and Prof. Buriak discuss:
- Prof. Buriak’s perspective on Open Data and its impact on scientific publishing.
- The influence of machine learning and AI on professional societies.
- The Buriak Group’s recent research applying ML and informatics to photovoltaic materials research.
- How Prof. Buriak utilizes Twitter and social media to grow her professional network and learn about new breakthroughs in the scientific community.
- How Prof. Buriak became actively involved in the scientific publishing realm.
“A journal is more than simply publishing scientific results, we also are the basis of community.” — Prof. Jillian Buriak
Speaker Bio
Prof. Jillian Buriak is a professor and the Canada Research Chair at the University of Alberta, where she has researched and co-authored over 100 papers on surface chemistry, nanoscience, synthetic materials, and inorganic nanomaterials.
In addition to her work as a professor, Prof. Buriak has been an editor for Science Magazine and ACS Nano and is currently the editor-in-chief of ACS Chemistry of Materials. She is a Fellow of the American Association for the Advancement of Science, the Royal Society of Canada, and the Royal Society of Chemistry in the United Kingdom.
Connect with Prof. Jillian Buriak:
Profile: @JBuriak
LinkedIn: Buriak Research Group
Transcript
Bryce Meredig: Welcome to DataLab, a Materials Informatics podcast with Bryce Meredig, Chief Science Officer at Citrine Informatics. All right, welcome to DataLab. Our guest today is Dr. Jillian Buriak. Dr. Buriak started her faculty career at Purdue University in 1997. She’s now a professor and Canada Research Chair at the University of Alberta where she has researched and co-authored over 100 papers on surface chemistry, nanoscience, synthetic materials and inorganic nanomaterials. In addition to her work as a professor, Jillian has been an editor for Science Magazine and ACS Nano and is currently the editor in chief of ACS Journal Chemistry of Materials. She’s a fellow of the American Association for the Advancement of Science, the Royal Society of Canada and the Royal Society of chemistry in the UK.
Bryce Meredig: Jillian, thanks so much for joining us. Welcome to the podcast.
Jillian Buriak: Thank you so much Bryce. Happy to be here.
Bryce Meredig: Well, I like to try to start these conversations out with a bit of a personal note about our guests. You’re somebody who is, especially for a scientist, I would say extremely successful on Twitter. You have over 4,000 followers. I’d be curious to hear an example of maybe a particularly interesting interaction you’ve had with someone else through Twitter that you don’t think would have happened otherwise.
Jillian Buriak: Oh sure, absolutely. Twitter has turned…part of the reason why Twitter and science is so useful is that you truly meet people. It sounds ridiculous. You can get used to people’s avatars and in a sense you get a sense of their personality. It really does feel like you get to know people. I know 4,000 followers, that’s Kim Kardashian-ish. She doesn’t have to exactly worry about me, but for science, it’s harder to develop these high numbers. There are fewer people, although that’s changing.
Jillian Buriak: One thing that is fun is that American Chemical Society meetings for instance is there a tweet ups, and so the tweet ups, people who are involved in Twitter and they’re going to be at the meeting, they all meet up. It’s been a great way to actually meet, put a true face as opposed to an avatar to these people, and so one person that I really value having met is Dr. Rubidium, so rubidium like the element. That’s Raychelle Burks. She’s in Austin, Texas. She’s very active on Twitter, very interesting, very passionate, funny, but also very biting at times with respect to social issues when she needs to be.
Jillian Buriak: I never would have encountered her, I think, and now she and I are friends, but thanks to Twitter, I know her and keep up with her on a daily basis. It’s very useful for that because science can be very isolating, and so Twitter has really crystallized different communities and link these individual communities together as well.
Bryce Meredig: I noticed that the chemistry community in particular seems to have adopted Twitter with some enthusiasm, and I would say insofar as the materials can be separated out from chemistry, I think the materials community lags maybe a little bit behind. You sit at the interface between those two worlds.
Jillian Buriak: Yes.
Bryce Meredig: Can you identify any reasons why chemistry in particular might have or more chemists might have adopted social media?
Jillian Buriak: The one thing that I do not see a lot of are engineers in general. I mean, there are some. Apologies if any of you are listening engineering tweeps. I wonder, I don’t really know why I also encounter some skepticism when I’m talking with people at meetings for people who are engineers and say, “Oh, Twitter, it’s such a waste of time.” I don’t know if it’s simply just a matter of time because chemists were very skeptical as well four, five years ago. Even today, some chemists are still skeptical. Then actually, I do see them quietly logging on and keeping an eye on what’s going on.
Jillian Buriak: Maybe it’s a matter of time. I think it just takes a while for these ideas to spread, and maybe four more of those meetings, like I mentioned, where people have all the answers or all the insights before everybody else, thanks to things they learned on Twitter.
Bryce Meredig: Well, that’s a good segue into I think the way that probably many people who have gotten to know you through Twitter, which is your role as the editor of Chemistry of Materials. I’m interested in knowing a little more about how you become actively involved in serving as a journal editor and how that fits in with the portfolio of all of your professional activities.
Jillian Buriak: It’s interesting. I never set out in life to be an editor. It sounds… I mean, can you imagine? I guess it’s like a lot of professions is that you can have a stereotype or an image of what that might be. For me, it sounded like there was a lot of red pens and a lot of corrections and a lot of circles on a page, and that sounded pretty boring. I never applied to be an editor. Back in 2004, the editors in chief of science actually approached me and asked me if I would be a part of their what’s called bored of reviewing editors. It’s called BORE, so B-O-R-E. It’s funny.
Jillian Buriak: What the bored does is they help the editors in chief decide which papers are going to go up for review. We would get seven to 10 papers a week. We turn around your comments in 48 hours. What was so fascinating about it is that you would see things, they’re not even fresh off the press. This is the pizza going in the oven. I mean, you’re seeing everything really fresh. What people think is very exciting and worthy of publishing this top journals, so you are seeing things before everybody. That then became in and of itself such an education.
Jillian Buriak: Again, it provided those sparks, those ideas. You could see trends. You could see new areas of science emerging long before they came out in paper form or in a publication in a journal. It became in my mind almost this this essential thing to doing good science that seeing the papers, seeing what was coming out of people’s labs so early on in the process and then also learning how the publishing process works and seeing peer review first hand.
Bryce Meredig: That’s great. I’m interested in hearing from you what you enjoy about the editing process, and also at a human level, when you’re talking about journals like Science or Chemistry of Materials, getting published in these journals can really shape someone’s career.
Jillian Buriak: Yes.
Bryce Meredig: How do you think about that when when you’re serving as an editor, when you’re making some of these very influential decisions?
Jillian Buriak: With science, when you’re on the board of reviewing editors, you don’t actually interface directly with authors. Where that changed for me was in 2009 when I started at the American Chemical Society Journal, ACS now, and actually, I was pretty burns out. I have to admit, after five years with Science, I actually went home and said to my husband, “You know, I think I’m done with this editing thing.” Turning papers around in 48 hours, there was always this demand on your time. If you got a paper on a Friday, that means you have done on a Sunday.
Jillian Buriak: It constantly was weighing on everything I did, vacations, et cetera. I said after five years, “You know, I’m really done.” Then one day, actually it was minus 40, and I was outside running, and Paul Weiss called. I talked to him for about a minute and a half before my phone battery died because it was so cold, but Paul had launched this journal, ACS Nano, which is now this fabulous journal, in 2007. When I finally got to a workplace and warmed up my battery, he gave me a spiel that really changed the way I see journal editing. He said, “Here’s what we’re doing at ACS Nano. We’ve a very personalized approach. We want to serve as the core or the home base of this whole nano community.”
Jillian Buriak: “And here’s what you’ll do. You have total editorial freedom. You interface directly with the authors.” I actually went home and said to my husband, “I think I’m gonna do this editing thing again.” He said, “But you said you’d never do it again.” I said, “Well, I guess Paul Weiss has convinced me otherwise,” and I am forever grateful for that. I was with ACS Nano for five years. If you ever have a chance to hear Paul Weiss speak, you really have to take advantage of that opportunity. He is just this tornado of energy and ideas. What he has done with ACS Nano is to transform a journal from being this place where you almost coldly or almost feels anonymously in a way.
Jillian Buriak: Submit your paper, it goes into this box, and then somehow something happens, and a few months later, you get a letter back saying, “Oh, your paper’s been accepted, rejected, it needs revision, whatever.” He really changed that by treating every paper, and we’re talking thousands of papers, in this very personalized way. When you submit a journal to ACS Nano, you’ll get back a little note from Paul or from from the associate editor that’s handling it saying, “Hey, thanks so much for this paper.” That’s not written by a computer. When there’s a decision coming and it’s a good one, you’ll get a little note from Paul that says, “Hey, good news, it’s coming shortly,” that sort of thing. He is out on the road all the time travel… I think he travels 300,000 miles a year, meeting people, putting a face to the journal. That’s what we have done too. They have a very active Twitter account.
Jillian Buriak: When the associate editors who were all practicing scientists are out giving talks, that’s advertised. You can go and talk to the editor. You can go and see the editor. The editor themselves, they’re all scientists. Five years of learning from Paul was so great, and I’m so grateful for having had that opportunity. In 2013, when ACS asked me if I would consider being editor in chief of Chemistry of Materials, I was really scared and nervous. Chemistry of Materials is a journal that has been around for 30 years, but I really felt strongly that Paul’s vision, Paul’s approach towards putting a human face on journal editing is the way to go.
Jillian Buriak: I brought that to Chemistry of Materials. We have very active Twitter account. Our editors, of course, they’re all practicing scientists. We really encourage people to meet them, to talk to them in person, to hear them talk about their own science. This helps to embody a community. We serve as a focal point, I hope, of scientists who want to get not only their science out, but we also invite editorials and people who want to talk about maybe upcoming trends or sometimes it’s even things that seem almost funny. Like, our second most highly downloaded paper right now is an editorial by a scientist in Sweden who doesn’t like the use of the word functional before materials, because he says all materials have a function.
Jillian Buriak: That may sound funny, but actually it helps us to think about what are field means and what a material is and what a material does, and how we can use materials. These are intriguing because a journal is more than simply publishing scientific results. It’s that we also are the basis of community.
Bryce Meredig: Well, I think you just provided a good reminder to me that actually, I owe you one of those editorials which I’m looking forward to writing.
Jillian Buriak: Oh yes.
Bryce Meredig: I also wanted to just make the observation about Chemistry of Materials from my perspective coming from the materials informatics community, the journal has really become I think a go-to outlet for some of the top-work in materials informatics, this confluence of materials, science, chemistry and machine learning. Is that something that you’ve promoted deliberately, or have you just noticed the natural change in some of the papers that have been submitted to the journal?
Jillian Buriak: We’ve been really lucky in hiring and bringing on Kristin Persson as an associate editor. Kristin is at University of California, Berkeley. She is one of the co-founders of the Materials Project or just the materials informatics project also at UC Berkeley. We have 22 associate editors with a huge range of expertise going all the way from biological materials through to inorganic nano materials to organic nanomaterials to computational methods and theory, and now the materials informatics. It’s really important too to scientists and authors to know that when you have a paper in a particular area that your paper is going to be handled properly.
Jillian Buriak: I mean, you’re really looking for competence. You’re looking for someone who knows the area, and so when people look to, “Okay, I have this paper where we’ve looked at 30,000 possible materials for this application of water splitting or batteries or whatever it is, where are we going to submit that? Who can handle this? Who’s going to know who the top people are who could review it, peer review it? Who will have the expertise to then handle or manage or interpret those peer reviews when they come in?” Because sometimes, peer reviews very often are not so cut-and-dried.
Jillian Buriak: It’s difficult to know how to proceed with them. When people see Kristin’s name, they say, “Wow, that’s the person I want to handle this paper.” Also on our editorial advisory board, we’ve got Gerbrand Ceder, and so people say, “Okay.” People like Gerbrand and other computational people who can also manage these sorts of papers, like, Rem Seshadri at University of California, Santa Barbara, or Jean-Luc Bredas at Georgia Tech, and so they say, “Hey, that place has the expertise and the competence to handle those papers.” They come to us. Kristin I know is also out there actively saying, “Send you papers to us,” and encouraging authors.
Jillian Buriak: When she sees an interesting talk being presented by someone, she’ll go up and ask them, “Hey, have you published that yet? Well if not, you know, consider sending it to us, and we’ll make sure that it’s handled properly.” I think that is really the critical thing. It’s people. It all comes down to people.
Bryce Meredig: What you’re doing must be working, because like I said, I just noticed organically over time some of the best work in our field is going to Chemistry of Materials. Of course given that it’s very interdisciplinary, there are a number of places folks could be submitting these materials informatics papers, but some of the top-work has found its way into your journal, which is something that’s been very exciting to see.
Jillian Buriak: Well, thank you. I shall pass it on to Kristin.
Bryce Meredig: I mentioned in your role as an ACS editor, you get a fair amount of exposure to the inner workings of the society in general. I’m interested to know from your standpoint with that hat on, how has ACS and maybe its journal’s been reacting to the emergence of data-driven science, machine learning in chemistry in particular?
Jillian Buriak: There is broad recognition that a lot has to be done, that we’re all on the very beginning or first steps of having true open data, true ability to be able to manage large data sets to make them available to communities through journals. We know that. ACS knows that, and I think that the… In fact, I don’t think, I know that the advantage of ACS publications and the American Chemical Society as a whole is that all of our editors are practicing scientists. Every single one of them is, and so you have your editors who not only do they see papers that are generating and contributing to large data sets, but their own papers are doing the same, and their own work also relies on these data places and open data and everything associated with that.
Jillian Buriak: No one knows better than these people, than these experts in the area who are also the editors to know how important that is, how this has to be done sooner than later and has to be done well, because they themselves are using and generating this data. That’s why ACS is uniquely positioned to do this.
Bryce Meredig: Open data is something that is a major emphasis for us here at Citrine as well. We have tried to be involved in the community largely through the professional societies, also journals to some extent to promote the idea of publishing more data, but it’s pretty clear I think that that in some cases or in many cases requires a culture change. What are some of the things that you think we could be doing as a community to make more data available to more scientists?
Jillian Buriak: That’s an interesting question. You ask of course. You ask 1,000 chemists, and you get 2,000 different answers. You’ll hear different things. A number of young people for instance are very, very passionate. Passionate almost is an understatement. Really, really passionate about open data, about open access. I get that. There are also many people particularly in chemistry who have found that the present system, submitting papers, publishing them, putting things in a PDF in the supplementary information, that’s worked fine all these years. Why change?
Jillian Buriak: There is conflict in this system. There is conflict of viewpoints. I don’t mean conflict in necessarily a negative way. Change only happens or change often usually happens through conflict of one of another. What you hear now is a lot of discussion. A lot of it is happening on social media actually, on Twitter, about how we do this. Another thing is for instance, there’s an editorial by Daniel Gezelter. He’s from Notre Dame. In J. Phys. Chem. Letters, that’s another ACS journal, really outlining what he thinks needs to be done to enable better platforms and how this needs to be done.
Jillian Buriak: Again, the journalists are also serving as the places where people’s viewpoints are being expressed really clearly and in a very constructive way. I think everything is in a state of flux. It’s unclear whether even for how journals are going to look in say 10 years. I mean, are they going to be hybrid? Are they going to be a subscription model? Are they going to be all open access? I suspect it’s going to be all open access within that time. How will data be managed? Who’s going to host the data?
Jillian Buriak: Who is going to pay for that? How is that going to be managed? None of this is clear at this point, but it is clear the direction this is all going. That’s interesting. We’ll see in 10 years what this all looks like, but it is also very, very obvious that the present system now is not sufficient to do this properly, so a lot of change will have to be implemented in order just to see this work properly.
Bryce Meredig: I think that’s right. I think we’re seeing of course change happening in publishing right now. In fact, one of the more recent pieces of news that generated a lot of excitement was the emergence of ChemRxiv, which of course is something something that ACS participates in. I’m curious what the reaction or the reception from your authors or your colleagues has been to ChemRxiv.
Jillian Buriak: It’s so interesting. ArXiv has been around, correct me if I’m wrong, but since 1988.
Bryce Meredig: A very long time, that’s right.
Jillian Buriak: This just become absolutely standard, if I can use that term, in the physics community for instance. ArXiv is open to science in all areas, but chemists just never ever, very rarely versus a synthetic organic chemist would never post their work in arXiv. Maybe a more physical person would, but they’d be the real exception. People said, “Oh no, chemists will never do that. Chemists are very happy with the way things are, with the standard peer review. Journals for instance like ACS are generally pretty fast, so why would you need that?”
Jillian Buriak: Well, so there was a fair amount of skepticism when ChemRxiv was launched, and a lot of people are saying, “Well, yeah, maybe just a small subset of the population is going to use it.” Well, holy smokes, it’s been really interesting. You have a really active leader of ChemRxiv who is very active on Twitter as @Organometallica. You can follow him on ChemRxiv. He’s put this face to ChemRxiv, and all of a sudden, everyone says, “Hey, this is actually kind of neat.” I think a lot of it is coming from our students.
Jillian Buriak: Our students are saying, “Hey, I like this idea of getting my work out there quickly. I like the idea of perhaps getting some feedback before we submit this for peer review.” All the ACS journals have come on board with this now, and so literally, over the course of six months, I think that broadly speaking, people’s viewpoints have changed. That’s a fast change. I think if you’d asked me two years ago, I would have said, “Oh boy, it’s like moving a big rock. It’s going to be really slow.” Maybe it was like moving a big rock, but there was enough of an incline going downwards that that rock picked up speed, and now it looks inevitable.
Jillian Buriak: It’s been interesting. ChemRxiv has really opened people’s eyes to the idea of posting things, making them available and making them available right away.
Bryce Meredig: It’s been really exciting to see to your point how quickly ChemRxiv seem to go mainstream in the chemistry community, which I think should give us all hope about the ability for example to publish more data, to get more researchers using informatics space methods as well because those changes seem to be coming quickly as well.
Jillian Buriak: Yes.
Bryce Meredig: Of course, we couldn’t conclude this podcast without asking you a little bit about your group’s own research. In particular, pretty recently, your group published a paper called Optimizing Organic Photovoltaics via Design of Experiments and Machine Learning. Would you be able to tell us a little bit about that work and why it’s exciting for you?
Jillian Buriak: Absolutely. I’m now 51 years old, and I feel like you can teach an old dog new tricks. It’s through a collaboration with my colleague Arthur Mar. Arthur is an inorganic chemist. He’s a solid state materials chemist. Now, we work on materials for solar energy conversion and storage, so we’ve been working on organic photovoltaics, elementally bonded inorganic materials for solar cells like zinc phosphide and iron sulfide, these rock-like materials. I think there’s tons of it on the earth’s surface.
Jillian Buriak: They’ve got amazing absorption properties, but they don’t really work as solar cells. We’ve been working on things like this. How can we convert these rocks into high-efficiency solar cells? We’re working on batteries and silicon surface chemistry for nanolithography applications for computer-type applications and things like this, but we never overlapped with Arthur enough to actually collaborate. Two years ago, there was a new institute that was set up on campus called the Future Energy Systems Research Institute, so it focuses on energy. Arthur and I are part of this.
Jillian Buriak: We had started talking about, “Well, what could we do?” Now, we have a very large database generated in house in our own group on plastic solar cells. That’s organic photovoltaics. We’ve had solar cells over the years, 11 years of data. Solar cells that didn’t work. Solar cells that worked. Some that didn’t work so well. Arthur asked us, “Can we use that as a database and apply machine learning principles and see whether we can at least get some drive, some data or at least some directions from that towards more higher efficiency organic photovoltaics and plastic solar cells?”
Jillian Buriak: We said, “Hey, why don’t we try that?” It turns out that generating a database is not as easy as I had thought, because, and I’m sure all the data managers and people who sanitize and clean up data ar rolling their eyes, databases are never perfect, especially when you’re starting with the raw data. It was out of that that we learned. We started talking about, “Well, what else can we do? If we were to generate a new database from scratch, how would we do it, and what would we do that’s different?”
Jillian Buriak: That then led into this whole field of design of experiments where it sounds really boring, but it’s not where you use statistics before doing all the experiments, and that will allow you to go and change several variables at once, and allow you to optimize systems much faster than you could if you’re just doing single variable, single optimization type strategies. It only change one variable at a time is one thing that we’ve taught over and over again, but if you use something called design of experiments, you can actually change several variables simultaneously if you do it in a statistically correct manner.
Jillian Buriak: That then led Arthur and I to start talking. We said, “Well, if we were to redo this database, let’s also do design of experiments,” and so we applied it towards plastic solar cells, these organic photovoltaics, these solar cells that you can potentially print like newspapers and be able to make really inexpensive rolls of these things. You could do lightweight. You could carry them anywhere. The problem is that they have so many different components. It’s that they’re very hard to optimize, and so this is the perfect system for design of experiments, and the application of machine learning type approaches once you have the data.
Jillian Buriak: Once we got talking, suddenly, you realize you’re in what might my students like to call the science vortex. As you’re talking so long, you don’t realize that it’s 8:00 PM, and you haven’t eaten dinner, and everyone’s wondering where you are. You haven’t even checked your cell phone, and that’s when you realize you’ve got a really neat project. That happened to us and led to that paper, but more importantly, that paper has changed the way we think about science, how we do research for everything.
Jillian Buriak: It’s been really neat. Not only can you discover new things, but you can also discover new ways of doing things that can actually be profoundly different. It goes back to that ChemRxiv thing about change can happen quickly, and it can also happen unexpectedly.
Bryce Meredig: I’m glad you mentioned or at least alluded to the challenge at least at the beginning of getting data organized, getting data to a point where it’s useful for machine learning, but I think it sounds like what you’ve found is that if you pay that startup cost, if you will, it puts you in a good position to just do data-driven science later. What we found anyway is that once people, students, other researchers adopt best practices around data, they find it easier to do going forward than to not do that going forward.
Jillian Buriak: Yes. That is so critical. Those best practices, those messages have to get out now because really what you have is you’ve got to 100 years of what we call modern scientific publishing. Just the sheer numbers of papers that people have published on, for instance, organic field effect transistors, I’m just making examples, organic solar cells, anything that’s been optimized, so much of this is buried in experimental sections in journals all in different formats, all with very subtle differences in the way people do experiments.
Jillian Buriak: It’s very difficult to compare experiment A to experiment B when they’re done in different groups. It’s not clear whether 100 years of data actually can be converted into a database, and so it’s almost like for a lot of this, we have to start from scratch. As you said, getting out those best practices is so important. We can’t waste any more time.
Bryce Meredig: We at Citrine like to call that stopping the bleeding. The first thing to do is to get your data under control going forward, and then over time, you can reach deeper, deeper into archives to digitize data or make it a minimal to machine learning.
Jillian Buriak: That’s true. I agree entirely.
Bryce Meredig: Well, it looks like we are coming up on the hour here. I wanted to ask you also, I assume you’ll be at the ACS meeting in Orlando?
Jillian Buriak: Oh yes. I go there every single meeting. Yes.
Bryce Meredig: Okay, great. I figured that’d be true, but I will be there as well.
Jillian Buriak: Cool.
Bryce Meredig: It would be nice to have the opportunity to meet in person if that makes sense.
Jillian Buriak: That would be great. Maybe it’s too late to be able to run one of these sessions or I’ll even call it a crash course. It’d be really great to have you there.
Bryce Meredig: Oh yeah. Well, that’s something we’ve started to do with a couple of the other professional societies, TMS and ACerS in particular.
Jillian Buriak: You have mentioned that.
Bryce Meredig: We’re finding out that there’s a lot of pent-up demand for teaching and education in these methods. A lot of students are asking for it. A lot of faculty are asking for it. We’re doing our best to play a role in promulgating the best practices and just tools of the trade as much as we can.
Jillian Buriak: I agree with those best practices. I hate to pester, but I mean, that editorial is part of it. I mean, that editorial by Daniel Gezelter on open data from 2015 actually is it’s really highly downloaded. These things do get noticed. We promote it. It would get out there.
Bryce Meredig: Absolutely. Jillian, thanks so much again for joining us on the podcast today. If listeners are interested in learning more about your research and your work, what’s the best way for them to do that?
Jillian Buriak: Thank you very much. You can just put Buriak Group into Google, and that should come up. Luckily, having a last name that’s not that common, that tends to work. You can follow us at Jburiak on Twitter as well.
Bryce Meredig: Great. Thanks again for your time. Thanks for listening. Please subscribe and rate our podcast on iTunes, Stitcher, or wherever you get your podcasts. Listen to past episodes or learn more about our guests and submit questions and suggestions at citrine.io/podcast. That’s citrine.io/podcast.