Taking census of physics

TL;DR

This paper provides a comprehensive analysis of the evolution and dynamics of physics subfields by studying the careers and publication patterns of over 135,000 physicists, revealing how subfields emerge, grow, and interconnect over time.

Contribution

It introduces a large-scale, longitudinal census of physics careers, mapping subfield transitions, growth patterns, and collaboration impacts, which was not previously available.

Findings

Most physicists start in three main subfields.

Physicists tend to branch out to new areas later in their careers.

Large-scale collaborations have significantly affected productivity and impact.

Abstract

Over the past decades, the diversity of areas explored by physicists has exploded, encompassing new topics from biophysics and chemical physics to network science. However, it is unclear how these new subfields emerged from the traditional subject areas and how physicists explore them. To map out the evolution of physics subfields, here, we take an intellectual census of physics by studying physicists' careers. We use a large-scale publication data set, identify the subfields of 135,877 physicists and quantify their heterogeneous birth, growth and migration patterns among research areas. We find that the majority of physicists began their careers in only three subfields, branching out to other areas at later career stages, with different rates and transition times. Furthermore, we analyse the productivity, impact and team sizes across different subfields, finding drastic changes attributable to the recent rise in large-scale collaborations. This detailed, longitudinal census of physics can inform resource allocation policies and provide students, editors and scientists with a broader view of the field's internal dynamics.

Figures

Full text

\useunder

\ulined

Taking Census of Physics

Federico Battiston

Department of Network and Data Science, Central European University, Budapest, 1051, Hungary

Federico Musciotto

Department of Network and Data Science, Central European University, Budapest, 1051, Hungary

Dashun Wang

Kellogg School of Management, Northwestern University, Evanston, IL 60208, USA

Northwestern Institute on Complex Systems, Northwestern University, Evanston, IL 60208, USA

Albert-László Barabási

Department of Network and Data Science, Central European University, Budapest, 1051, Hungary

Network Science Institute, Northeastern University, Boston, MA 02115, USA

Center for Cancer Systems Biology, Dana-Farber Cancer Institute, Boston, MA 02115, USA

Michael Szell

Department of Network and Data Science, Central European University, Budapest, 1051, Hungary

Network Science Institute, Northeastern University, Boston, MA 02115, USA

Complexity Science Hub Vienna, Vienna, 1080, Austria

MTA KRTK Agglomeration and Social Networks Lendulet Research Group, Centre for Economic and Regional Studies, Hungarian Academy of Sciences, Budapest, 1094, Hungary

Roberta Sinatra

Department of Network and Data Science, Central European University, Budapest, 1051, Hungary

Department of Mathematics, Central European University, Budapest, 1051, Hungary

Network Science Institute, Northeastern University, Boston, MA 02115, USA

Complexity Science Hub Vienna, Vienna, 1080, Austria

There was a time when polymaths like Galileo knew all the physics that was there to be known. Over the centuries, however, the body of knowledge spanned by physics exploded, encompassing topics as diverse as gravitational waves, graphene, or network science. As physics expanded in breadth and depth, physicists were forced to specialise [1], segmenting researchers into their narrow, specialised communities. How many physicists work in each subfield of physics today and how does each subdiscipline evolve? In which subfield are physicists “born” into and where do they migrate, if at all? Here we take an intellectual census of physicists, their activities and career trajectories, helping us understand the evolution of the field and gaining quantitative insights about several fundamental scientific processes, from resource allocation to the exchange of knowledge. Advances in this direction were limited by the challenge in answering two fundamental questions: 1) Who can be counted as a physicist? 2) How do we survey their activities? The recent availability of large datasets of scientific publications finally offers opportunities to tackle these questions by exploring the production patterns of the scientific population [2, 3]. Indeed, the close to complete publication records of all physicists allow us to reconstruct their subfields of study and career changes, offering quantitative footprints not just for the field of physics, but its intimate relation with the broader scientific community [4, 5].

Combining large-scale data on physics publications and citations with recent data and network science techniques, here we ask: What are the impact and productivity differences between subfields? As a physics student choosing my future specialty, how do I know which subfields are growing? As a funding agency, how do I compare early-career physicists from different subfields? As a journal editor, how many papers should I expect from each subfield and how do I compare their impact?

A census of physics subfields

To offer a data-driven answer to these questions [2, 3], we identify the relevant physics papers and citations within Web of Science (WoS). We start by selecting $\sim$3.2 million physics papers, published in 294 physics journals indexed by WoS. This core represents, however, only a fraction of all physics papers  [5, 6], missing for example those published in interdisciplinary journals like Nature or Science, or papers published in journals of other disciplines but that are of direct relevance for the physics community. To map out the complete physics literature we then set to detect physics papers by virtue of their patterns of citations among the other $\sim$47 million papers in WoS. A paper is a potential physics publication if its references and citations to the core physics literature are significantly higher than in a null model in which each paper’s citations are assigned randomly, regardless of a paper’s journal or research area. We identified $\sim$4.5 million papers whose patterns of citations and references are indistinguishable from papers in physics journals (SI Section S1), obtaining overall a dataset of $\sim$7.7 million publications of interest to the physics community.

We characterise further this physics corpus by classifying each paper into nine major subfields, according to the Physics and Astronomy Classification Scheme (PACS) used by the American Physical Society [7] between 1985 and 2015. We use this dataset to reconstruct the publication profile of 135,877 physicists with a persistent productivity between 1985 and 2015. See Box 1 and SI Section S3 for more details on the dataset curation and validation.

The first step in developing a census is to count the number of physicists working in each subfield. Such counting is, however, not straightforward, as physicists may contribute to publications in different subfields. We therefore associate each physicist with a primary subfield if the number of her publications in the subfield is higher, in a statistically significant manner, than expected for a typical physicist (Box 1 and SI Section S4). The obtained subfield demographics offer us a first summary statistic (Fig. 1a): we find that the largest subfield is CondMat (condensed matter physics) with more than $62,000$ physicists, capturing 46% of the entire physicist population. It is followed by General ($34,000$), HEP (high energy physics, $33,000$), Interdisc (Interdisciplinary physics, $32,000$), Classical ($28,000$), Nuclear ($24,000$), AMO (Atomic and molecular physics, $20,000$) and Astro physics ($19,000$). Plasma is the smallest subfield of physics, with less than $11,000$ researchers.

Given the highly specialised nature of the physics subfields, one might suspect that most physicists work in a single subfield. Yet, we find that highly specialised physicists are the exception rather than the rule: The majority of physicists (63%) are active in two or more subfields (Fig. 1b). This prompts us to ask: Which subfields have particularly low or high rates of specialisation? The differences between subfields are striking, defining two different groups (Fig. 1c): six subfields have less than $10\%$ specialised physicists. Among these subfields, Interdisc has less than 1% of specialised physicists, in line with the expectation that interdisciplinary physicists bridge multiple subfields. In contrast, the percentage of specialised physicists in CondMat, HEP and Nuclear is 42%, 34% and 25% respectively, at least an order of magnitude larger than in the other group of subfields. What drives the different levels of specialisation between subfields?

A physicist working on two or more subfields combines the collective know-how of these fields, a process deemed essential for novel discoveries in science [8, 9, 10]. To understand which of the physics subfields cross-pollinate most significantly, we calculate the co-activities of individual physicists between each pair of subfields. Co-activities are defined by weighted links between subfields, where the weights measure the observed versus expected co-activities based on a randomised null model (SI Section S6). Starting with the highest weighted links, we plot the minimum number of links needed to have a connected network of subfields (Fig. 1d). The network reveals a non-trivial co-activity structure, clustering all physics subfields into three broader areas, 1) Interdisc and CondMat, 2) Classical, AMO, and Plasma, 3) HEP, Astro, and Nuclear, all held together by General. This research space captures the intellectual affinities between subfields, facilitating movements between close subfields, while limiting cross-pollination between distant ones like Interdisc and Nuclear[11]. For example, the diversity of topics within CondMat and Classical and their adaptable approaches, like statistical mechanics applied to multiple systems composed of large numbers of entities, makes it easier for those working in these subfields to take their tools to different disciplines. In contrast, more specialised subfields like HEP or Nuclear require their members to acquire familiarity with large-scale, long-term projects. While scientists working in such fields may have deep knowledge and expertise on the subject they specialise in, they face a greater burden that limits their ability to explore other areas. The observed network is similar to the citation network[5] between subfields, showing that the flow of knowledge is captured through multiple metrics, both by paper citations and by the activities of individual physicists[11].

Birth, growth, and migration

Why are there so considerable differences in specialised physicists between similarly sized subfields, like Nuclear and Interdisc (Fig. 1a,b)? To understand this heterogeneity, we first assess the relative growth rate of each subfield over time, measuring the fraction of physicists entering a subfield every year (Fig. 2a). We find that the growth rates of Interdisc and Astro increased from a few percent in 1985 to over 20% and 27% respectively after 2010, substantially reshaping the physics landscape in recent years. An opposite trend characterises CondMat: while it had the largest share of new physicists in 1985, its share dramatically decreased over time, falling below 5% after 2010. HEP also displayed a receding trend just before 2010, but the spur of new research connected to the activity of the Large Hadron Collider in Geneva injected new forces into the field. In particular, HEP’s sharp peak in 2010 can be attributed to the first ATLAS and CMS publications [12] (SI Section S7).

Figure 2a mixes together physicists who start their careers in a particular subfield with those who make career transitions to other subfields. There are remarkable examples of physicists who never changed their subfield, like Klaus von Klitzing, whose first publication was in CondMat, and contributed over 500 papers to the subfield, earning him the Nobel Prize in 1985 for the discovery of the quantised Hall effect. In contrast, Rainer Weiss, best known for inventing the laser interferometric technique at the heart of LIGO, which earned him the Nobel Prize in 2017, published his first paper on an unrelated topic in AMO, “Magnetic Moments and Hyperfine-Structure Anomalies of $Cs_{133}$, $Cs_{135}$ and $Cs_{137}$”. To distinguish such different careers, we next systematically explore career transitions within physics [13], asking: Where are physicists “born”, and how do they “migrate” between subfields? When do these transitions typically occur?

Figure 2b shows how many physicists began their careers in each subfield (top rectangles). Remarkably, $64\%$ of the physicists began their careers by publishing in either CondMat, HEP, or Nuclear ($37\%$ of all physicists start out in CondMat). These three subfields capture “curricular” physics topics, the natural ending points of many undergraduate courses, hence the typical starting point of research careers. General, covering topics of interests to a wide set of physicists, accounts for $14\%$ of first publications. In contrast, only $4\%$ of physicists started publishing in Interdisc, and as low as $3\%$ began in Astro. As Interdisc integrates other disciplines, it might be difficult to start out as an Interdisc physicist; the low percentage of Astro starts may be rooted in the fact that traditionally it has not been a “curricular” subfield.

Box 1

**Identifying subfields

We classify papers into 9 subfields, based on the 1-digit Physics and Astronomy Classification Scheme (PACS) by the American Physical Society (APS):**

•

General: Mathematical Methods, Quantum Mechanics, Relativity, Nonlinear Dynamics, Metrology

•

HEP: The Physics of Elementary Particles and Fields

•

Nuclear: Nuclear Structure and Reactions

•

AMO: Atomic and Molecular Physics

•

Classical: Electromagnetism, Optics, Acoustics, Heat Transfer, Classical Mechanics, and Fluid Dynamics

•

Plasma: Physics of Gases, Plasmas, and Electric Discharges

•

CondMat: Structural, Mechanical, and Thermal Properties; Electronic Structure, Electrical, Magnetic, and Optical Properties

•

Interdisc: Interdisciplinary Physics and Related Areas of Science and Technology

•

Astro: Astrophysics, Astronomy, and Geophysics

PACS were consistently used in papers published in APS journals between 1985 and 2015 (SI Section S2). Using an algorithm that evaluates the patterns of citations and references between papers, we propagate subfield labels from APS papers to other papers: if the fraction of references and citations between a given paper and papers in a particular subfield is larger than expected by the null model, the paper is assigned to that subfield. A paper may be assigned to multiple subfields, in line with APS papers reporting multiple PACS. In panel a) we show an example of an unclassified paper which references in CondMat, Plasma and Astro, and which is cited by CondMat, Astro and another publication still lacking a PACS. The publication is first assigned to CondMat and then to Astro, but not to Plasma, as it lacks statistical significant links to the subfield. The algorithm is run iteratively until convergence for each subfield, helping us associate at least one subfield to 1,137,670 papers (SI Section S3).

**Assigning physicists to subfields

We analyse all careers with at least $5$ labeled papers between 1985 and 2015, capturing the careers of 135,877 physicists. We consider a physicist working in a subfield if her share of publications in the subfield is higher than that of the average physicist. The statistical criterion  [14] we used, guarantees that each scientist is assigned to at least one subfield, and takes into account the different sizes of subfields. As an example, we show the result of the criterion applied to the career of Stephen Hawking in panel b). In the physics dataset Hawking has 124 papers associated to different subfields. Of these subfields, only General (95 papers) and Astro (77 papers) are assigned to the physicist through the statistical criterion, whereas HEP (23 papers) and Classical (1 paper) are not statistically significant, which is consistent with Hawking being known as a theoretical physicist and cosmologist.

**

For validation and further methods see SI Sections S3, S4, S5.

The links of Fig. 2b capture the significant flows between subfields, linking the subfield where a physicist published her first paper, to the subfields that best characterised her later careers (SI Section S6). This diagram indicates that CondMat is the starting point for many physicists who later specialised in Interdisc, Classical, and General. HEP and Nuclear tend to swap researchers while feeding talents into Astro, a pattern that may be rooted in the fact that all three subfields study radiation or nuclear and subnuclear processes. We find that most Interdisc physicists did not start their career there, but migrated from CondMat and General, consistent with the hypothesis that one needs to acquire expertise in at least two fields before being able to bring them together. Finally, Plasma and Astro welcome physicists with many different backgrounds, but rarely feed into other subfields. The diversity of the incoming flows to Plasma and Astro suggests their accessibility to physicists with many different backgrounds.

We also measure the average time it takes to transition to a different subfield, captured by the vertical axis of Fig. 2b. Once again, HEP, Nuclear and CondMat top the list: physicists who did not start their career in these subfields tend to transition towards them the earliest, typically by the third or fourth year of their research career. The opposite trend was observed for Interdisc and Astro, which not only have the highest transition rates among subfields, but are also characterised by the longest time to transition. Indeed, on average a physicist publishes her first paper on these two topics $6$ to $7$ years into her career, roughly double the transition time towards HEP, Nuclear and CondMat. Interdisc displays a late switch, consistent with the hypothesis that it takes time to gather expertise in multiple fields. Similarly, physicists tend to switch to Astro typically after a relatively long experience in HEP.

The flow diagram of Fig. 2b helps us better understand the research space captured by Fig. 1d. For instance, in the bottom right triple, HEP plays the leading role in producing physicists who transition to its tightly connected subfields, Nuclear and Astro. In the top two nodes of the network, CondMat is the main force feeding Interdisc. The observed widespread career transitions may reflect potential benefits to the whole field, cross-pollinating one physics community with ideas and methods developed by a different subfield [8, 9].

The role of chaperones

The future prosperity of young scholars has often been linked to access to valuable mentorship at the early stages of a scientific career [15, 16, 17]. For example, a surprising fraction of Nobel laureates had a mentor-mentee or a co-authorship relation with another Nobel laureate [18, 19], and scientists who co-author early with an established scientist are more likely to have higher impact and higher chances to publish as lead author than other scientists [20]. Taken together, a senior scientist who acts as “chaperone” during a scientist’s early career might foster the acquisition of skills, passing on experience and knowledge necessary for high achievements later in a career.

To quantify the chaperone effect, we measure how many physicists co-author their first paper in a subfield with a physicist who has published in that subfield before [20]. We find that the chaperone effect is particularly strong for HEP, Nuclear and CondMat, where over $90\%$ of physicists wrote their first paper with someone who published before in the same subfield (Fig. 2c and SI Section S8). This large share of chaperoned physicists could have several reasons, like the documented high number of physicists starting their career in these three subfields, or the need to access large facilities, which require early-career physicists to collaborate with established scientists. Note that the typical large co-authorships patterns of HEP can not explain the magnitude of the chaperone effect characterising this subfield (SI Section S8).

Other subfields have a lower fraction of chaperoned physicists, especially Interdisc and Astro. These subfields are often explored by more senior physicists who received mentorship at a previous stage of their careers in a different subfield and often decide to explore the new area without close supervision (26% of physicists are not chaperoned in Interdisc and Astro, Fig. 2c). On top of this, applications of computational physics, like computational biophysics or complex systems, classified as Interdisc, require lower financial resources compared to experimental research and could also play a significant role in explaining the low chaperone effect [21]. Taken together, the chaperone effect is strong in physics, with an average rate of 82% chaperoned physicists across subfields. The effect signals a research culture where physicists often get introduced to their future research area by senior colleagues in a collaborative setting, in contrast with disciplines like mathematics, where the majority of scientists start their career with publishing solo-author papers[20].

Productivity, impact, and team size across subfields

Productivity and impact, capturing the number of papers published and citations received by a physicist, are frequently used metrics in the assessment of scientific careers [22, 23]. These quantities have implications for decisions and policies involving predicting, nurturing, and funding early career scientists. Yet, the proper interpretation of these metrics must account for the highly heterogeneous productivity and citation patterns characterising different subfields  [24] and for different team sizes [25], both of which vary in time.

Team size, i.e. the number of coauthors per paper, has been increasing steadily over the past decades in all fields, capturing an increasing collaboration in science [26]. Are there particular differences in collaborative patterns in the different physics subfields, and what are their implications on productivity and impact? To answer this question, we assess the diversity and evolution of collaboration, productivity, and citation standards in the different subfields of physics. First, the tendency of scientists to work in increasingly large teams has been particularly pronounced in HEP (especially after 2005), Nuclear (especially after 2010) and Astro (especially after 2000) (Fig. 3a). The observed explosive growth in these three subfield is partly rooted in large-scale projects like ATLAS (SI Section S7). They also result in an increased productivity: as physicists were involved in more and larger teams, the average number of papers they published each year increased by a factor of 10 for HEP and by a factor of 2 for Nuclear and Astro from 1985 to 2015 (Fig. 3b). However, for the other six subfields productivity has stayed constant over 30 years, and for all subfields productivity has increased at a slower rate than team sizes. These different rates of increase explain why fractional productivity, i.e. the ratio between the number of papers and the average team size, decreased across all subfields (Fig. 3c). The effect is the strongest in HEP, Nuclear, and Astro, where team size grew disproportionately. It is worth noting that in these subfields authors are usually ordered alphabetically due to the large average team size, making the assessment of credits for single authors more problematic [27]. Taken together, we find that the amount of knowledge produced per capita decreases in all subfields despite the increase in the total number of physicists and physics papers.

Given the explosive increase in both team size and the number of papers per physicists in HEP, do HEP physicists today have more or less impact than they had decades earlier? To answer this question we measured the average impact in number of citations after 5 years (Fig. 3d) and the fractional impact (ratio between number of citations and average team size, Fig. 3e) per physicist per subfield. Interestingly, the average impact of HEP shows a growth of comparable magnitude as the growth in average productivity, leading to an unchanged fractional impact. In other words, large-scale projects like ATLAS produce papers that generate a large number of citations, compensating for the massive numbers of co-authors (hundreds or more).

Given some of the large productivity differences between different subfields, we also expect differences in impact [28, 29, 30, 31], measured in terms of cumulative citations over a career. For instance, how much impact does it take to be a scientific leader in HEP and how is that different in CondMat? In Fig. 3f and Fig. 3g we show the total number of papers and citations acquired over an average career by the top 5% of physicists in each subfield (in terms of productivity). In both terms, HEP is by far the most rewarding subfield, whose top scientists coauthor 169 papers and accumulate over 7,000 citations. In contrast, top Interdisc physicists coauthor only 18 papers with less than 1,000 citations. The large discrepancy is not explained by paper citation rates [32, 33], which are roughly constant across subfields (SI Section S9), but by the high or low number of papers per author in the respective subfield (Fig. 3b). As a consequence, when physicists with different specialties compete for positions or grants, caution is needed in comparing their profiles using metrics based on citations or productivity, as subfield-dependent differences appear from the very beginning of a career.

What about the rate of top papers in the different subfields? We selected the top 1% of all physics papers (in terms of citations) and assessed into which subfield they fall (Fig. 3h). The majority falls into CondMat, General and HEP, however, this result is trivial as these fields produce the most papers. To unveil the significant effects we measured the surplus between this top 1% distribution and the distribution of subfields of all physics papers. As Fig. 3i shows, Interdisc papers are 40% more likely to be in the top 1% than expected, while Nuclear and Plasma papers are 40% less likely to be found in the top 1%. The high rate of Interdisc among the top cited papers might be partially explained by the finding that papers which are 15% novel and 85% conventional often have high impact [9]. Interdisc is more likely to achieve this balance, since interdisciplinary research must be novel and, at the same time, must adhere to established principles. Another explanation is that Interdisc is more likely to initiate new topics or emerging subfields. Papers that do open such new avenues are known to acquire a high number of citations as they become milestones, cited by subsequent papers once the field is established [34, 35].

Recognition of physics subfields

Do impact differences affect the way in which the overall scientific community perceives the different subfields of physics? As a rough proxy of this recognition we take the Nobel Prizes awarded from 1985 to the present, highlighting each awarded subfield (Fig. 3j, SI Section S10). Although the Nobel Prize often recognises research undertaken much before the selection year, the timing of Nobel prize selections could affect the way in which the relative importance of different physics communities are perceived by the committee. As a comparison between Fig. 2a and Fig. 3j shows, Nobel Prizes are not related to the number of physicists flocking into specific physics communities, nor do they show significant temporal clusters. However, the general distribution of awarded subfields reveals interesting tendencies: a large fraction of Nobel Prizes have been awarded to the “curricular” topics, like CondMat, the subfield with the largest number of active researchers, and HEP. Surprisingly, Astro, despite the relatively moderate size of its community, comes in third, with five Nobel Prizes. This success might be linked to the perception of astrophysics as a field that studies the universe on a grand scale, as well as to its strong ties to HEP, a regular recipient of Nobels. Other well established areas with a long history, such as AMO and Classical have also been recognised. In contrast, since 1985 Plasma and Interdisc have not been awarded a Nobel Prize. The omission of Interdisc likely comes from the charter of the Nobel Prize to award clear-cut categories (e.g. physics, chemistry, medicine/physiology) rooted in 19th century discriminating against interdisciplinary discoveries [36, 37].

Conclusions

As one of the oldest scientific disciplines, physics plays a fundamental role in the development of science. As the aperture of physics widens, the focus of individual physicists narrows, leading progressively to the formation of specialised communities and subfields. Here we offered an intellectual census of these subfields, exploring how physicists migrate between them, how they specialise and collaborate to create impactful research.

We observed that subfields rarely live in isolation but rather tend to overlap, with individual scientists working in multiple subfields and transitioning between fields during their career. Mapping these overlaps reveals a highly non-trivial research space, displaying deep intellectual links between some subfields and large gaps between others.

Physicists who are confronted with heated arguments on the allocation of resources to different subfields and departments, often use metrics of productivity or impact to seek priority. However, our research suggests that such arguments should be taken with scepticism. Indeed, there are considerable field-specific differences in the patterns of productivity and impact. Publication rates have exploded in recent years in HEP, Nuclear and Astro, whereas fractional productivity is declining. In some subfields, such as HEP, researchers co-authors an exceptionally large number of papers, partly rooted in their unique culture of collaboration. By contrast, interdisciplinary physicists produce papers at a much lower rate but their papers tend to garner a disproportionally higher impact, once we factor in the relative size of the subfield. Understanding these field differences within physics represents the first step towards a deeper understanding of our discipline. As tomorrow’s physicists working on different topics compete for the same position and resources, these insights may prove pertinent for the sustainable vitality of physics as a discipline.

Our study is based on Web of Science data, lacking the literature that has been exclusively published in preprint servers like arXiv[38], leading to unavoidable (but small) differences in subfield representation due to diverse publication cultures in different communities. For example, the proportion of HEP and Astro papers in arXiv is higher compared to our dataset and WoS, reflecting the common practice of these communities to communicate findings in preprints rather than journal papers. However, there is a high overlap in the coverage of the physics literature between different databases [39] and a high correlation of the representation of physics subfields (SI Section S3), indicating that our findings should agree if repeated on a different database.

In this study we focused on careers of physicists within physics. However, these days, many scientists with a background in the physical sciences contribute to fields outside of physics, from biology to finance, both in academia and the private sectors [40]. For this reason, the investigation of the connection between physics and other scientific disciplines, and the career transitions away from physics, remains as fruitful future work. Indeed, such an investigation, possibly aided with data sources that go beyond scientific publications, could shed light on the role of physics and its subfields in the entire ecosystem of science and beyond.

Acknowledgments

This work was supported by the John Templeton Foundation Grant #61066 (A.-L.B., F.B., R.S., and M.S.), the ITI project “Just Data” funded by Central European University (F.M. and R.S.), the National Science Foundation grant SBE 1829344 (D.W.), and the Air Force Office of Scientific Research grants FA9550-15-1-0077 (A.-L.B., R.S., and M.S.), FA9550-15-1-0364 (A.-L.B. and R.S.), FA9550-15-1-0162 (D.W.) and FA9550-17-1-0089 (D.W.).