Institute of Mathematical Statistics | XL-Files: XL is X (or LX)

December 13, 2022

Xiao-Li Meng writes:

Aging is a process that few look forward to, but like everything else in life, it is not without its silver linings. One of them is getting invitations to toast and roast your friends (or secret foes), at their expense. In the last month alone, I was invited to an LX, an LXX, and a, well, X birthday symposium. The Roman numerals seem especially useful here because some ages prefer to remain anonymous. I gather few birthday stars would mind the imputing X, a perfect ten, and a symbol especially dear to our profession. Of course, I also have a self-serving reason to invoke X. No, I don’t mind disclosing my age, especially as its Roman numeral is already embedded in my initials. (Incidentally, for those who make a living in the US, I assume we would all be happier to see MXL, my un-Americanized initials, than seeing its Arabic equivalence). Rather, it is because I have my own X anniversary to celebrate, and you, dear reader of the XL-Files, are invited. To thank you for your decade-long indulgence of my Chinglish, I offer—below—the XXXIII XL-Files choices for you to toast or roast. (Whatever you choose to do, I’d recommend it be well done. But if you prefer medium rare, well, be warned that food-for-thought poisoning is more than a thought.)

A Decade of XL-Files, from I to XXXIII

While you are busy cooking or savoring, an MDL (Monolayer Deep Learning) algorithm applied to the yearly tally of XL-Files, in the table below, reveals two periods lurking in the X years, thanks to the Roman numerals: those with V and those without it.

Annual tally of XL-Files columns

2013	2014	2015	2016	2017	2018	2019	2020	2021	2022
VIII	V	IV	IV	V	II	I	I	I	II

Just in case you are thinking, “Oh no, the aging XL is regressing into a Roman numerologist”, this finding is perfectly interpretable, without any overfitting. The V-years were exactly those where I was provided a team of over L members, responsible for ensuring the smooth operations of more than LV PhD programs year around; and the VC(V complement)-years started with my appointment as the founding Editor-in-Chief of Harvard Data Science Review (HDSR), where I only had (and have) the resources to recruit a team of II members. Whereas a graduate school will continue to sail or at least flow without a dean’s navigation, for a start-up, every individual’s involvement can be the difference between sink or swim, especially when launching anything of scale but without scalable resources. I therefore had to devote XL hours a week (sometimes I wonder if my parents foresaw that number when they named me, LX years ago) to HDSR, and that of course was in addition to my paid job (which reminds me annually of the Form MXL). Dropping the V from the annual count of XL-Files was therefore a sad but necessary coping step. It’s sad because composing XL-Files always entertains me, with my mind meandering, ten digits dancing, and laughing gear lubricating (Hint: it’s V o’clock somewhere…).

But again, every pain can generate gains, some of which are rather unexpected. The following remarks document one of those delightful occasions, though without the accompanying slides—or the abundant social lubricants—you need to exercise your imagination. While you are at it, could you imagine my delight of being effectively given a lavish birthday party without having to worry about getting roasted and toasted? I of course thank my coeval, Amy Brand from MIT Press, for that. Go ahead and imagine this is your free, and worry-free, celebration too, if you have the same amount of hair as me.

Xiao-Li’s Remarks on celebrating the 60^th anniversary of MIT Press

Yes, Harvard Data Science Review is published by the MIT Press. No, that’s not a sign that there is an ongoing secret negotiation to merge the two parent institutions, despite at least six historical attempts made by one of them. But it does signal their different choices. Harvard University Press chooses to publish only books, whereas MIT Press publishes everything, including journals. Publishing journals is becoming an increasingly exciting, yet challenging business, because of the open access revolution. Harvard Data Science Review is an example of both. The open access platform encouraged HDSR to have a very ambitious plan. It aims to be a popular magazine, inspired by Harvard Business Review; a premier scholarly journal, like Harvard Law Review; and a cutting-edge educational publication, given education is the ultimate mission of a university. We have four categories covering perspectives, impact, education, and research, along with seven columns, reaching out to industry, government, students, teachers, and the general public. HDSR is extremely flexible in article styles and content, as long as they are data-science-related, from an interview with the MIT president (I’m sure MIT students are thrilled to hear their president declaring, “Our curriculum is awfully hard”), to featuring a collection of articles like a recent one on online teaching with two courses from MIT, and their strategy for benefiting from the technology developed in the gaming industry (and I know this will relax many parents out there who will suddenly realize that their children have been spending so much time playing online games merely to prepare themselves for getting into MIT).

However, most people’s initial reactions to this three-in-one and incredibly versatile platform range from “Xiao-Li, are you crazy?” to “How are you going to market this?” But the response from MIT Press leadership was nothing short of enthusiastic. My first conversation was with Nick Lindsey, the director of journals and open access, over a momentous lunch on October 19, 2018. Nick’s reaction, after hearing me briefly describing what we try to accomplish, was simply “We’d love to do this”. No questions asked. The reaction from Amy Brand, the Director and Publisher of MIT Press and a leading expert in scholarly communication and research information, was the same. We quickly got into discussing how HDSR can be brought to serve two purposes: to increase Harvard’s leadership in Data Science on the global stage, and to facilitate MIT Press’s adventure in the world of open access. Yes, I did have a very ambitious vision for HDSR, which was to help define and shape what data science is or should be. But as Thomas Edison said, “Vision without execution is hallucination.” I am therefore deeply grateful to Amy, Nick, and everyone at MIT Press, as well as the HDSR editorial office, for preventing me from hallucinating, and more importantly for making HDSR an award-winning journal within two years of its launching. Now, with over 800,000 unique users (though as a statistician, I will take that number with half as many grains of salt), it has been accessed from IP addresses originated from every country/region around the world (except for Vatican City—so please pray with me to cross that one off the list!).

But open access is extremely costly, especially when one wants to innovate and do it well. Take this article on the authorship of Beatles songs as an example. These interactive features cost us $18,000, and that’s just for one article. I worked quite hard to raise funds, to a point that I was telling my friends that I finally got to experience what it is like to be a startup; because I had to work on raising funds, recruiting people, and creating product—or papers, in this case. However, one friend asked me, “But Xiao-Li, did you have to mortgage your house to do this?” When I said no, he said, “Then that’s not a start up.” Indeed, I did not need to mortgage my house because of some generous donors who believed in the HDSR’s vision, just as MIT Press did. HDSR chose MIT Press as its publisher not because MIT Press has all the resources we need, because it does not. This is exactly why MIT Press needs all the support it can have to continuously innovate and to lead the scholarly publications in this era of open access. For those of you who are in the position to help, on behalf of all future hallucinators, I thank you for whatever you can do to support MIT Press.

I also have a piece of unsolicited advice, and it is free. During my years as Dean, I learned quite a bit about fundraising from Harvard’s development office. One piece of wisdom was that if I need money, I should ask for advice, but if I need advice, then I should ask for money. Now I am quite sure MIT alumni and friends would love to prove that this Harvard wisdom was ill informed. You can accomplish that by giving Amy money whenever she asks you for money. Of course, if you want to know how much Amy really values your advice, just give her money whenever she asks for advice, and see if she would refuse your money, and insist, “No, no, I really want your advice.”

Now, if for whatever reason Amy cannot take your money, please feel free to send that check to me. I will use the funds to celebrate the fifth anniversary of the happy marriage between MIT Press and HDSR, and of course you will be invited. On top of that, I will send you a special inaugural volume of HDSR, which you can sell on eBay to recover all your generosity because it comes with my signature, which is priceless. For everything else, there is MIT Press.

Thank you and Happy Sixtieth, to MIT Press and to Amy!

Xiao-Li’s Remarks on celebrating the 60th anniversary of MIT Press

Xiao-Li’s Remarks on celebrating the 60^th anniversary of MIT Press