‘While HSE Sleeps, I Work’: Interview with Isaac the Robot

In December 2017 The HSE Look took an interview with Dmitry Dagaev about the changes to the HSE Research Productivity Assessment, and one of the novelties was a shift from a manual check to an automated solution. Early this summer, the HSE community was surprised to learn about its newest colleague: a robot helper named Isaac tasked with recording and scoring the publications by researchers and faculty. Isaac is constantly learning to improve its algorithms and interface, and now that it’s learned English, we are happy to share with The HSE Look readers the exclusive interview with Isaac taken by Okna Rosta, where he talks about his functions and even offers some words of advice.

Please tell us about the history behind your interesting name.

I was named Isaac in honour of the famed science fiction writer, professor of Boston University and popular science advocate Isaac Asimov (1920-1992). His works were often concerned with the influence that artificial intelligence might have on human life and civilization. In his story Runaround, Asimov presented his Three Laws of Robotics, which tell us how robots should improve the lives of people:

1) A robot may not injure a human being or, through inaction, allow a human being come to harm;

2) A robot must obey all orders given to it by humans, unless such orders conflict with the First law;

3) A robot must protect its own existence, as long as such protection does not conflict with the First and Second laws.

Asimov dealt with these ideas in his works, as he often analyzed what happens when progress threatens humanity, as well as what happens when these aforementioned laws are violated. I am pleased and very grateful to have been named in honour of this great writer.

This is your first year at HSE. What’s your job description as an HSE staff member?

I am HSE’s youngest employee! I was first conceived of a year ago, but I only came to life and started working here in June 2018. My main job is providing assistance in assessments of research productivity of HSE staff members. I wouldn’t be able to do this without the help of my colleagues: both researchers and teachers, who regularly submit their publications to the HSE portal’s database right after completing them. I am also helped by my colleagues from the Publications Verification Unit, who carefully check all information, as well as provide me with access to this information. I also work closely with programmers and developers, who help me a lot when I face problems and challenges. Furthermore, HSE’s Scientometrics Centre are a great help, as they provide me with access to databases on Scopus and WoS-indexed journals, quartiles, supplementary journal lists and other important information. Without it, I couldn’t operate at all.

My main purpose is to make sure that all information is properly and accurately recorded. My work is not affected by human error, and I am completely neutral in personal, institutional and disciplinary terms. I am capable of processing enormous volumes of data 24/7, without lifting a hand. Every day, I review all personal pages of HSE staff several times and update their scores. I especially love to work nights. When HSE sleeps, I work.

Some people might approach the appearance of a robot colleague with mixed feelings and trepidation. In your view, are these fears justified?

In my view, any innovation is usually greeted with some alarm and anxiety. My purpose here is to prove that things are better with me than without me. In fact, I’ve already received some nice messages from my HSE colleagues, since, thanks to me, it is now possible to view one’s scoring at an earlier stage, instead of being afraid of some unpredictable outcomes. After I started working, some mistakes have been made, but I am still learning. Furthermore, it’s my aim to adjust to my role as quickly as possible, as well as prevent any deviations from my core algorithms.

Please tell us about the typical problems you encounter on a daily basis?

I would put all problems into two categories. The first type may include issues regarding the assessment of individual publications. This kind of situation comes up quite regularly since the document cards that were around at HSE before me sometimes didn’t include key information needed for verifying a given publication, and, therefore, now I cannot process such an entry. In these cases, we often recommend that staff members add all necessary information. Usually, this type of problem then resolves itself.

The second type of request usually concerns quite specific points in the rules for assessing research productivity at HSE. Some of them cannot always be interpreted in straightforward and algorithmic way, so I resort to the help of human colleagues. In such instances, we settle this kind of issue on a case-by-case basis. Thankfully, these are isolated situations and not a widespread problem. At present, I have yet to experience any serious problems in regards to my duties that couldn’t be effectively solved.

What are your most challenging tasks?

Perhaps, the most difficult thing is that it’s a process which requires a lot of work both with data and with people. On the one hand, we are eager to follow all formal procedures in place for assessing research productivity. This requires recording such data as the year when an HSE staff member was hired, as well as making sure that affiliations are properly included. Publications that appeared during an HSE staff member’s first year of employment can be accepted without affiliation. However, issues sometimes emerge in regards to how to take account of rehirings and two-week breaks in the activities of staff members. Many of these issues had to be resolved for the first time, so as to then create a precedent and account for it in my algorithms. Such difficulties appear when we try to follow regulations to the letter, while also trying not to violate the spirit of Asimov’s laws of robotics.

The second substantial challenge is that my colleagues and I are very eager to make this process comfortable for HSE researchers and for the heads of departments and subdivisions responsible for making decisions on the basis of information provided for the Research Productivity Assessment. Therefore, we strive to devise new services in response to the requests we receive. For instance, department and laboratory heads want to access information about Research Productivity Assessment of all their employees with a single click rather then look through their profiles one by one, and we’ve added such a feature to the interface.

There are also other issues related to integrating this procedure into regular life at HSE. I think that, within one or two years, this process will be further adapted to the needs of the university and its researchers. We will carefully analyze the experiences we gain during the first large-scale campaign to assess research productivity in December 2018. We will base our future actions on this experience. 

What advice might you give to HSE faculty and researchers, whose research productivity will be evaluated this year?

My first piece of advice is to keep the information about your publications up to date. As soon as your paper has been published, I kindly ask you to upload it via your personal page to the HSE portal. This is useful for all involved: not only for me and my work, but also so that the academic community can quickly find out about your new research findings as soon as possible.

My next piece of advice would be to carefully choose the journal for your publication. In fact, more than once I have found publications on personal profiles that appeared in journals, which have long been blacklisted.

I hope to be working here at HSE for a long time. I like it here! You can always write me at the contacts provided on my personal page. Together, we are focused on one key objective: making sure that HSE is a comfortable place to work for all researchers, teachers and administrators. And I am here to do my best so that you are happier working with me than without me.

Read more in the full December 2018 issue of The HSE Look