Thomson Reuter's website Sciencewatch.com has a special section out on citation and paper data for the last ten years of quantum computing. More below the fold.
The first interesting thing that everyone wants to know about such list is, of course, "who's number 1?!?" Well the most cited paper over the last 10 years according to this survey was "A scheme for efficient quantum computation with linear optics" by Knill, Laflamme, and Milburn (Nature, 409, 46-52 (2001)). Not too surprising given that this paper established a completely new method for building a quantum computer and, equally important, was clearly the first word in this subject area, not the last. Indeed if you look down the list of top 20 papers over there is a common theme: novel methods for building a quantum computer (linear optics, one-way QC, a proposal for QC with molecular magnets, topological QC, teleportation of gates are the top five papers). This is probably because such methods are accessible across a wide swath of physics and as such can be cited by entire fields. The cynic in me would add the words "who want to grab some quantum computing funding" to the last sentence, but I grumpily digress.
The data on paper and citation data for countries is kind of interesting, though in line with what I'd expect. Here one sees that over the last ten years, the US has gotten only (at most) 31 percent of citations and produced only (at most) 24 percent of all papers. Poor China gets rammed for have a ton of papers, but only a small number of citations, though I bet the database used in the survey wouldn't actually catch the Chinese literature well.
Another interest fact is that Nature has published twice as many quantum articles as Science...something you might consider next time you submit a paper to one of the glossies. Indeed Science has one article in the top 20 over the last ten years, while Nature does exceptionally well.
And yes this is all a lot of navel gazing. But it is Thursday afternoon.
- Log in to post comments
Hey, I think I see some lint in there...
I would say that the KLM paper very much deserves the number one spot, but it was certainly not the first word in the subject area. Quite a lot of people at the time (including myself) were trying to figure out how to do efficient quantum computing with photons and linear optics.
Thanks Dave.
These surveys always need to be taken with a large pinch of salt, but they are important, not least since governments are increasingly turning to Thompson citation data when allocating grants and funding.
I was surprised to see Canada so low on the ranking - is this just because PI / IQC didn't get going till mid-decade?
The main question that strikes me, though, is "Where are all the computer scientists?".
Is it the different citation / publishing / conference culture? Could it be that Thompson is missing data from the CS conferences? Or is it simply that CSists are just not using the right keywords in their abstracts?
Here's how the data was collated:
Dan I thought about the lack of CS as well. Shor's paper, for example is too old even though Google Scholar shows it having well north of 1500 citations.
But I suspect the main reason is that the CS community in quantum computing is small and is nearly completely and totally isolated from the larger CS community. If I write a paper on quantum algorithms, then at best it will pick up a few citations from the STOC and FOCS crowd. If I write a paper on universal quantum computing, every physicist who is building a quantum computer might be inclined to site it.
So it seems to me what this shows is how marginalized quantum CS is. Whether this is deserved or not, is a question I'll leave up to others :)
Hi Peter,
Sorry for the wording, but I can't figure out how to word it better. Certainly people were thinking about linear optics and photons, but was there any progress that was in the line of using prep and measurement for the needed "nonlinearity"? My understanding was that all prior work tried to use some sort of nonlinear medium to work and KLM's big achievement was that these could be the preparation and measurement.
I think there are two main reasons that the CS-ish papers are so poorly represented. First, as I read their methodology, they only count citations to papers in journals, not conference proceedings, and presumably also don't count citations *from* conference proceedings. So CS papers of all kinds take a big hit right off.
Second, I am fairly sure that there are just many more people doing experiments related to quantum computing or theory related to experiments related to quantum computing than there are people doing theoretical CS-ish work (of all kinds, not just quantum information related). Even taking conference citations into account correctly, the numbers are going to be much lower.
Hi Dave,
I would say that KLM managed to pull off three major breakthroughs in their paper: 1) they showed that you can construct a probabilistic two-qubit gate with linear optics and feed-forward measurements; 2) they showed how to do gate teleportation with optical Bell measurements (highly nontrivial); and 3) they constructed an error correction scheme that mitigates the unavoidable errors in their procedure.
The first part was looked into by many people at the time (Tim Ralph, Bill Munro, Chris Adami, Nicholas Cerf, Sam Braunstein and me, to name a few), and the second part fit into a tradition of trying to figure out what kind of transformations were possible with LOQC, starting with Reck, Zeilinger et al in the mid nineties, again Cerf and Adami, and just before the KLM paper by Lutkenhaus, Vaidman, etc. And these are just the theorists.
I believe that one of the reasons the paper made such a splash is precisely that there were so many people already working on this, and it showed both theorists and experimentalists the way forward. Ironically, few people understood the technical details of the KLM paper in the early days; the conclusion that LOQC was possible was enough to turbo-charge the field.