MI Initiative

Its not Mission Impossible.

The Malaysia-Italy Centre of Excellence for Mathematical Sciences (MICEMS) was established on 10 March 2016 (see this post) and it was officiated by the Deputy Minister of Higher Education in the presence of the Italian Ambassador to Malaysia. At the time, I was carrying out the duties of the Director before Prof. Akma was appointed as the second director. of INSPEM The centre was initiated out of an existing collaboration with Politecnico di Torino (Polito - see also here) with MoU signed much earlier on September 26, 2013. When the idea of the centre was mooted by our Deputy Vice-Chancellor (Industry & Community Relations), I help wrote the proposal paper of MICEMS. To be clear here, at the time I wasn't part of the existing collaboration then but I knew it was a good opportunity and wished that we can make this a reality. Discussed some of the ideas with Prof. Lamberto Rondoni and one of the model we had discussed was of course the well-known International Centre for Theoretical Physics (ICTP) in Trieste. Along the way, some of these ideas had changed mainly to accommodate the administrative and financial realities of both universities. Personally, I had told Prof. Rondoni that within my personal capacity, I will try to see things work through for the centre. It is worthwhile to remember some of the hopes of the centre that we had mentioned to the university within five years of the proposal paper for MICEMS:

• To conduct at least three simultaneous research projects shared between our researchers and the researchers of Polito.
• To establish relationships with industries in Turin and other cities of Italy.
• To initiate an international research grant proposal with Polito.
• To plan research programmes and invite renowned international researchers to the centre.

In a way, my future visit to Polito is hoped to address some of these (with my sincere hope that my visit gets endorsed by the Ministry). On my visit's agenda, they are:

• To discuss possible research with Prof. Rondoni (DISMA) and Prof. Bianca Cerchiai (DISAT) respectively on geometric aspects of chaotic systems and quantum information systems.
• To visit Marco Pizzi at Eltek Group, an industry does research on many things including (I was told) semiconductor lasers.
• To visit SmartData@Polito, Data Science and Big Data Research Centre in Polito.

There are many other places I would love to visit of course, like University of Turin and Institute for Scientific Interchange (ISI). But for five-day visit, there is already much to do at Polito itself (including things not listed in the above). The main purpose of the visit is very much to chart MICEMS' future with our Polito partners and hopefully extend the scientific cooperation beyond whatever existing collaboration. In part, just about a week ago, we organized an international seminar on Recent Trends in Mathematical Physics as a MICEMS activity (see pic below) with guest mathematical physicists from within the country and abroad. My personal thought is that it is good to have such an activity introducing MICEMS and that we identified mathematical physics as an area for us to pursue further within the centre, since this is what I personally can do and that it also covers the existing research collaboration in chaotic systems.

I have been telling my colleagues that we should continue supporting MICEMS as an initiative that was given support by governments of both countries. In fact, in my personal view, Malaysia has much more to gain, given that Italy has a longer scientific tradition than us. Italy was right there when Western civilization began its scientific revolution. I think all of us have at least heard the story of Galileo Galilei who helped revolutionize science in the west. Within mathematical sciences, we have from Italy, the following historical figures: Leonardo Fibonacci, Gerolamo Cardano and Giuseppe Luigi Lagrangia. In fact, DISMA has Lagrange's name attached to the department's name. Besides this, it is worthwhile to note that ISI has also launched the Lagrange Project to venture into complex systems and data science. Other notable Italian historical scientific figures are listed here. Today, Italy continues to 'export' its intellectuals worlwide in many different sciences. Apart from a few exceptional cases, our country has yet to come to this level. We will probably need to progress a lot more before attaining such height. In particular, with respect to mathematical sciences, an area often forgotten locally, we have to do much more. Perhaps MICEMS can help do this. For now, I hope to pursue MICEMS at the personal level, irrespective of the outcome of the institute's restructuring. It is sometimes difficult to get researchers to commit to some idea unless it is proven successful and thus the responsibility I personally feel. In any case, I would like to rally colleagues and friends to help support MICEMS initiatives so that we can achieve at least in part what we would like to see happen for mathematical sciences in Malaysia.

Returning to My Roots

The past few weeks, my mind was occupied regarding matters of the institute's restructuring to the point I had little time for other things. This is probably reflected in my last few posts. I have to convey my apologies to my guests who are here for our International Seminar on Recent Trends in Mathematical Physics (coorganized by INSPEM and MICEMS) but I did not quite have the time for them. Also, the next few days, I will be spending time with my family for my first son's convocation.

Our proposal for the restructuring has also taken a different turn. The proposal paper has been written and hopefully submitted for consideration. It took into consideration both the wishes of the researchers on the ground and the university administration. Hopefully it will be well-received for otherwise, we will be criticised by both sides.

Given that the proposal paper is done, my wish is that the matter will not be prolonged any further (noting that there are risks). Much time has been spent on this and it has only caused anxieties. In fact, the whole affair made me wonder whether I should stay on as deputy director. Particularly when my guests were around, I found better excitement when discussing with them, exploring new research ideas. Besides quantization, ideas followed by Syed (my visitor), we also stumbled onto the possibility of using some ideas of quantization that can be used for chaotic systems which is the area of research of our research fellow, Santo. That excitement reminded me of my old days and made me realise something. While matters of the institute is important and to ensure the tradition for mathematical sciences established firmly is what I believe in, perhaps I should be best left to research in mathematical physics and should have not gone into administration. Given my term ends on May 31, if I would be given any choice, I prefer to end my administrative days. I prefer to contribute only as a simple researcher and focus in graduating my students.

In my mind, I had different ideas of a mathematical institute than the one we have; one that explores sophisticated mathematics, very much like the famous institutes abroad. One of the things that we found out while writing up the proposal is the fact all these well-known institutes had relatively few publications (much less than us) but the work they produced are highly cited because the depth of ideas they explore. Also another important character of such institutes is that they are based on very strong visitor programs. Such an institute will probably not survived the number-chasing research environment and an administration that is overly concerned about expenses over long-term benefits. The composition of research areas in such institutes is often not dominated by only applied mathematics but at least a good balance of pure mathematics. This will probably helped us better in international recognition (see mathematics rankings of UM and USM). It is however unfortunate that our local mathematicians tend to shy away from such fields. If ever there is any who started off as pure mathematics, it is almost always, they get pressured by circumstances to change their fields. It is for this reason, I get furious when my junior colleagues (ex-students of mine doing theoretical physics) are asked to change fields to a more 'productive' one. One of our speakers at our recent event, Prof. Teo has assembled a team of mathematical scientists at Xiamen University, which I thought is very interesting. I am pretty sure if they ever wanted to be ranked, they will probably be doing well. Such would be my dream of what the mathematical institute would be. Another aspect would be the interdisciplinary character, which again I stress that this includes cross-overs within the mathematics sub-disciplines. Of course, I can only dream.

Having said I prefer not to continue in the new institute's administration, here are my wishes for the future leadership of the institute. I hope the person will be very open to developments of new mathematical fields and be knowledgeable in the general ideas of such fields. Another wish is for him/her to be more open to international outlook. Particularly whenever international guests are present in the institute, the person can strike interesting intellectual discussion with these guests. Such gestures often give a good impression of the institute, being aware of the development of mathematics as a whole.

So for now, I wish to return to my roots in trying my best to contribute in a challenging area of mathematical science.

The Fight for Mathematical Sciences Part 2

As I watched Avengers' Endgame new trailer today and the suspense it builds around it, I can't help thinking that I'm experiencing the suspense build around another endgame for the institute itself. Two days ago, when I chaired the institute meeting, there were voices of dissatisfaction, voices of the wants of mathematical community, some of which sounded patronising. Then I had the spur of realisation that this will not gel well with the university. The university management too has their wants that ought to be considered by all, to which I remarked in the meeting, if we can't frame our wants in the context of what benefits the university, we will forever be battling.

In the meeting, I had prepared a presentation to assess the progress that we had made as a community of mathematical scientists in UPM, particularly with reference to INSPEM. I was trying to move away from the position of only responding to the pressure of external factors of restructuring and high KPIs, which are not in our control, to a position of where we as a community of mathematical scientists would want to go amidst constraints. I had spent long hours on the presentation, thinking of what to say. But when it was presented, I didn't get to say as much as I wanted to say. So perhaps this post is to make up what I have missed.

I began with our recent QS subject rankings. Now I know there is this love-hate relationship between academics and rankings and some even spoke of ignoring rankings altogether. I do not subscribe to this view. Some go on to mock by taking a single datum, make singular comparison and confusing different datasets (between individual and organizational achievements) altogether. I prefer to approach the matter as a scientist by viewing the rankings are simply datasets and there is valuable information that comes with them. A single datum does not mean much but better information can be gotten when this is seen in a bigger context within one dataset or even correlating them with another dataset. I do agree however, rankings are open to abuse (as much as anything else) and administrators begin to become obsessive about them. But we should separate this form of poor judgement from seeking valuable information from the rankings (and yes, it takes effort to uncover useful information).

I have mentioned in the last post that UPM was ranked 201-250 for the subject of mathematics in 2019 along with UKM and UTM. I am pretty sure some of those who attended the meeting do not know about this. But what perhaps even more not known that we were also ranked 201-250 in the year before (at least I do not know it myself until I start to check this recently).

Perhaps what I wanted to point out is that we did not quite care about this before because of the different context we are experiencing and because of our own apathy. We are made aware of the ranking this year because of two things: the imminent threat of restructuring and the fact that there was a social media announcement that took the length of even reporting the top 300 and top 400 subjects being ranked in the university but mathematics was not mentioned. In other words, if we do not look for this piece of information, we will remain oblivious about the matter. Now our own 201-250 ranking information does not mean much but the fact that we are in the same bracket as UKM and UTM while UM and USM ranked above us, does mean something. It will be more meaningful if we look deeper into the ranking scores.

What we could see that our ranking relied mostly through our academic reputation, which we have build over the years. The things we need to further improve on are our citations per paper and the h-index scores. When I say improve, I don't mean manipulatively but more from the intrinsic value of each of our publication. As I said, one can make use of this ranking information to help us progress better in mathematics rather than being apathetic or even emotional about it. I also took the trouble of looking at the ranking of materials science in UPM (251-300) due to many of our university researchers are in this area. The scores reveal the opposite from that of mathematics; here the researchers need to work on the academic reputation as the citations per paper score is the highest among the Malaysian universities.

QS subject ranking relied more on citations. With growing concern of our increased KPI, the publication volume dataset will be another relevant source of information that we can look to. Note however, that larger number of scientific publications does not necessarily entail better scientific progress but may only increase its probability of being so. Let me show the UPM total math publication again from the last post.

One can see clearly the steep increase of mathematics publications is after the declaration of research university status. It must be noted however, that not all of these publications come from mathematics department or the institute. Other entities also contribute to this volume of publications. The breakdown can be seen from the following piechart.

It can be seen that computer science and engineering contributed substantially to the total mathematics publications. This is followed by materials science and physics. This might seems to be a negative point to mathematicians but one should make use of this knowledge positively. We can use this to plan for a better future version of the institute and hence meet the interdisciplinary need of the institute.

It should be interesting also to compare the mathematics publication output of the other research universities with that of UPM. To do this, we limit the period of publications to only the last five years (2014-2018), so that each dataset is comparable to the other. This is shown below.

Apart from UKM, all the research universities (including UPM) follow the same trend and hence can be attributed to some conditions happening at the national level. This is particular the case for the dip in mathematics publications for last year (2018) and hence UPM's decrease in mathematics 2018 publications must not be seen in isolation. The UKM case seems to be anomalous but the dip in 2016 seems to be present also in UM, USM and UTM.

Let us go back to the mathematics publications of UPM. Do INSPEM contributions follow the same trend as UPM? Below we show the mathematics publications for the period of 2003-2019 (to coincide with INSPEMs) followed by INSPEM's own mathematics publications. (Please ignore the last data point since this only represents the publication volume for the first three months of 2019.)

INSPEM publications (the second one) do not follow closely the curve for UPM (the top one) but there is a similarity in trend. The dip in 2014 for UPM is shown in INSPEM as a decreasing trend and this decreasing trend on average continues. One can also see that the decrease in publications in 2018 is seen in both UPM and INSPEM. Hence decrease in 2018 INSPEM publications can be attributed to something that happens at the university level or even at the national level (referring to the earlier diagrams for all RUs). Let us do a more quantitative comparison by listing the fraction of UPM mathematics publications that come from INSPEM (see below).

Note that from 2010-2014, INSPEM' s contribution is always above 50% (apart from 2011). From 2015 onwards, it fell below 50% which is worrying. The year 2015 was the year before our second director took over from the first and I can only remember this was the period when we had much reduced budget and even lost some contract staff. Nevertheless, such decreasing trend should send us warning signals that we need to rectify the situation. Perhaps the university's insistence for higher KPIs will help, though my personal humble opinion that this should be done with care. As I have once said that the usual response to a threat is fight or flight. It would be much better that if the administrators can convince all researchers of the need to address the problem and by them owning up this problem, there will be a better working and research environment and hence help the university reach its desired targets. Again numbers are just numbers and they are only meaningful if there is indeed a genuine scientific progress made in the research ecosystem.

I ended the presentation with the same subject bubble plot that I have shown in my last post. The basic idea is to further refine the Malaysian mathematics publication analysis right down to the subfields within mathematics. The plot consists of a vertical axis of cites per document which shows the intrinsic (average) citation that a subfield has, while the horizontal axis is its (extrinsic) h-index whose value are determined by citations across a history of documents. Bubbles of each subfield are plotted in this parameter space with the size of the bubble is determined by the number of publications in the subfield for the considered period. In a way, these gives us an idea of how each subfield is trending in Malaysia. I gave both plots of the 2013-2014 period (five years ago) and the 2017-2018 period (last year). This is shown respectively below as the first one and the second one.

As clearly shown above, mathematics research is very dynamical. Let me take the example of mathematical physics, the field that I am in. In 2013-2014 diagram (the first one), it is represented by the bubble next to analysis. It has about 6 cites per document and the h-index of about 22 (taken from 33 documents). In 2017-2018, the cites per document is reduced to about 0.9 but h-index remains about 22 for mathematical physics. As stated earlier, the absolute values themselves may not mean much but the relative position in the parameter space reveals perhaps much more meaning. For mathematical physics, the intrinsic cites per document became the highest in mathematics but its h-index remains modest relative to other fields. The subject with the highest h-index is Modeling and Simulation (the bubble besides Applied Mathematics) with the value of 55. Its cites per document (about 0.4) is relatively high than most other subjects but is still much lower than Mathematical Physics.

Perhaps another important point that can be drawn by the diagram is that even the subfields of mathematics shows variations of citation culture, even within mathematics itself. It will probably be of no surprise if one is to plot the bubble diagram for the whole of science, one can see much more variation in the citation culture. I did not show such diagram in my presentation but below I give it for Malaysian publications in the period of 2017-2018.

One can see that mathematics citation culture is among the lowest of all the sciences. The other lower subjects (the bubbles to the left under mathematics) are Economics, Econometrics & Finance and Arts & Humanities. It is perhaps no wonder that researchers among the social sciences often felt underappreciated when they are being compared to the other sciences. But such feeling can be extended to the subject of mathematics itself. It is best that university administrators understand this cultural difference in citations among the different sciences or even more generally the cultural difference in research among the different sciences.

On the next day after the institute meeting, I met the former director who actually founded INSPEM to relay to him about our current worries. I can see his disappointment about the imminent changes to the institute. He mentioned his wishes that we (me and the management) will be able to carry INSPEM through in the future. I left his room with the feeling of a heavy burden on my shoulders. After coming back home from work, I had another spike in my blood pressure and hence rested thereafter. My nap was shortened as my mind went through many scenarios and having thoughts of this particular post. So how should we do the plot of the endgame?

The Fight for Mathematical Sciences

These days of information glut, one needs a conscientiouss effort to thrust forward valuable information for our use. Otherwise valuable gems of information may simply get squashed by the volume of other information. Recently QS came out with their 2019 subject rankings and our university quickly highlighted those subjects that were ranked in the social media (see picture below).

Not seeing Mathematics in the picture, my gut feeling told me that it must have got ranked given the progress we have made so far. So I checked the QS Rankings for the subject of mathematics and yes, we were ranked at 201-250 (see below).

I informed my management colleagues and they felt that it should be highlighted by the university as well in the social media. Hence I mentioned it in one of the private university forums and they said that they will change this (so far I have not seen it). I do not think the absence of mathematics was purposely done but more probably is that less interest is taken in mathematics (which is a shame). I have no intent to make this as a big issue but such lack of interest in the progress made by our mathematical scientists can be misused by others, giving the impression that we have not progressed as much as in other sciences. On the other hand, while being ranked as such, we know ourselves that we still need better progress to move forward.

As part of the reason on why I would be surprised that we would be unranked is that there have been other studies showing that our country is progressing well in mathematics research. For instance, the paper "Mathematics Research in Association of Southeast Asian Nations Countries: A Scientometric Analysis of Patterns and Impacts" shows Malaysia is one of the four countries (the rest are Singapore, Vietnam and Thailand) that contribute most in the mathematics publications (95%) in ASEAN. In a way I was slightly surprised that we are comparable to Vietnam since Vietnam has a longer mathematical tradition associated to both French and Russian traditions (see particularly the math highly cited publications in the table below.

Some of these, as expected, are having international collaborators, which according to the paper makes our papers to be better cited. Malaysia, so it seems, are mainly having collaborators from Iran, India and Saudi Arabia (I was surprised by the third) - see the graph plot below. Uzbekistan, whose mathematical community in Malaysia is significant, seems to come fifth after USA. To digress, it was through Uzbek mathematician, Abdumalik Rakhimov (once our research fellow in the institute), that I got my Erdos number 4 (which of course does not say anything significant) - see below. Another observation is the graph density (not calculated in the paper); ours are as dense as Vietnam's but not as dense as Singapore's, showing better international diversity.

Now the question is how much of this is contribution from our university, UPM. First note that the dataset used in the above is from the period of 2006-2015. While I can't find (at the time of this writing) how mathematics contribution are distributed among say, the research universities, the distribution among broad science categories can be found here. Let us just focus on mathematics in UPM where I have done some information filtration through Scopus before this. Here is the mathematics publications affiliated with UPM from 1980-2019 (see below).

Prior to 1992, there seems to be negligible Scopus-cited mathematics publication from UPM. An appreciable increase in publications can be seen from 2000 onwards. A steep rise can be seen after 2006. It is plausible to make connections with some known events. Some notable events are (i) the formation of the institute INSPEM in April 2002; and (ii) the status of UPM as research university in October 2006. Of course, there is a multitude of factors that can be considered. We will next consider only those publications affiliated with INSPEM.

It is interesting to note that the first publication begins in 2003. It is very much plausible that there may be researchers of the institute that have published before this but without INSPEM affiliation since the institute is a new phenomenon. One can also note that from the side numbers, the number of publications for each year is lower than that for the whole of UPM. Until 2016, the percentage of INSPEM publication relative to UPM total math publication hovers between 42-60%. It is to be noted that researchers have the freedom to put down the INSPEM affiliation or not (even if the research grant is parked at the institute) in their papers. There is also the policy that the publications from the institute will still be counted at the various faculties. While this can be said as a disadvantage, the institute has always given the freedom to the researchers though we recommended to them to include the institute affiliation in their publications. The percentage of INSPEM publications dropped to around 30% beginning 2017. We are not quite sure why this is so but I have argued before that certainly this coincides with the period of reduced fundings and shrinkage of the number of staff.

Having observed the above, it is only natural for the institute to study whatever weaknesses and threats there are to the institute. This is the more reason why we at the institute should consolidate our strength and resources to ensure the survival of the institute. We have already heard the news of possible restructuring of the institute, which undoubtedly has brought the morale down and makes it even more difficult for us to progress. Recent setting of the KPIs has also worsened the situation for the institute mainly because the KPI setting is done without differentiating the research culture of each discipline. It is widely known that mathematics has modest publication rates, which is at best is two or lower per year - see the issued statement of American Mathematical Society over here. Another study of American mathematicians in fact revealed that it is even lower than one per year (see figure below).

Similarly for citations, the citation culture is different in mathematics, often lower than other sciences (and hence also for other indicators that depend on citations like h-index) - see the following study by International Mathematical Union and the following discussion (pg 32-34) in European Mathematical Society Newsletter. It would be terribly unjust to ignore these differences. It will be no surprise then by setting the KPI of publication rate at five papers per year (as main author), for instance, scares everyone away from being a member of the institute. Had it been KPIs being merely targets and unrelated to salary movement or promotion assessments, it would have been more acceptable. Unfortunately, this is usually not the case and most people see KPIs in the context of threats. The usual response would be fight or flight.

Setting a blanket KPI would be an easy route for university administrators to monitor progress, but one that will come at a price. Unrealistic KPIs would invite unethical approaches by researchers that often may embarrass the universities themselves (see Retraction Watch posts on Malaysia). At best, researchers will show indifference towards to the vision and goals of the university. Some may ask why should the university allow different treatments for a particular discipline. Why should we be considered 'special', as someone asked. Then one can respond back with the same reasoning of why should any ranking agencies be interested in ranking of universities by subjects. It is just a simple acknowledgement of the fact that there are research cultural differences one can't just ignore. I understand the fact the university wishes to progress and not be contented with present achievements. I believe everyone in their right mind, wants to be so. That is why I suggest that any progress be done realistically and incrementally based on past achievements and present resources. We could look to the subject ranking for mathematics as a guide. For instance, for the publication KPI that one needs to set is perhaps the average number of publications of a mathematician from the bracket of top 151-200 universities, since we have already being ranked at top 201-250. Now the criteria for the QS ranking was based mainly on citations and overall reputation and not publication rate, but with some efforts, we can get the average number of publications from Scopus dataset. Below are some breakdowns of the ranking indicators used.

The first column is the overall reputation, the second is citations per paper and the third is h-index citations. We can immediately see that we really need to work on the citations (and I don't mean manipulatively). One should now work on publishing our research materials in journals where the relevant community flocks to (and this does not necessarily means Q1 and Q2 journals as some would like to think). Each journal tends to have a particular following and we can look for those respectable ones where our research output would be most read by, particularly those which are more specialised. There are other 'strategies' that one could also use; for instance uploading preprints to some repository like the arXiv, making our work more visible (and hence also open to more critics). We can even go to more details; within the mathematics subfields, which are the ones more identifiable with our institute, build reputation around it. Here is a snapshot of the mathematical subfields found visible for Malaysia in 2017-2018 given by Scimago.

Let me just end this post with a saying that I have read elsewhere: as the university gets more critical with our achievements, one should also be equally critical with the critieria that we have been measured against. One can be easily deluded by 'numbers' that appears objective, but it is what do these numbers actually mean, which is the subjective part. Instead of being preoccupied with such numbers, we should be more concerned with the science that we grow. Deep within us, we know ourselves whether we are doing alright or not.