Network Institute

Ancient Questions and Modern Answers:
Using Statistics to Describe Biblical Hebrew
Etienne van de Bijl, Master’s Student, Business Analytics,
Cody Kingham, Master’s Student, Biblical Studies and Digital Humanities

Our interdisciplinary research project, entitled “A Probabilistic Approach to Linguistic Variation and Change in Biblical Hebrew,” examines the differences in syntactic tendencies between various books in the ancient Hebrew Bible. Our method combines probabilistic methods from the business analytics world with the classical problem of textual origin in biblical studies. Doing this research has required us to bind and share knowledge between two very different disciplines. In week-to-week meetings, this often meant walking each other through new concepts, while seeking to use our expertise to address our research question. As a result, we each take new experiences back to our own disciplines. Here are some of the valuable insights we’ve gained.

From the Angle of Data Science

Roughly speaking, data science is all about gaining knowledge from data. Many applications of data science areas include marketing, manufacturing, telecommunications and many more! Therefore, applying data science to analyse biblical texts seems suitable. Two high-level goals in data science are predicting future events and/or describe objects of interest. Often, provided data sets are not consistent or contain a lot of missing values. As a data scientist, one has to deal with these kind of problems in an adequate manner. The data provided by the Eep Talstra Centre for Bible and Computer (Hebrew data) is well maintained, which makes it great to work with. The cooperation between a Biblical expert such as Cody and a data scientist as Etienne is fundamental for this kind of interdisciplinary research. On the one hand, linguistic knowledge and, as data scientist would refer, “Business knowledge” is required to do meaningful analysis on these texts, while on the other hand, methods and algorithms are required to capture and compare syntax.

From the Angle of Biblical Studies

The field of biblical studies is perhaps one of the oldest areas of academic interests in the western world. As such, every topic in this field is connected to many additional concerns that simultaneously touch on philosophy, theology, anthropology, history, and more. To bring in new methods, such as computational statistics, is to introduce something completely new to very old and relevant questions. And that is exciting! In our work, we are addressing the question of the Hebrew Bible’s textual origins, i.e. when was it written, or where did it come from. We do this by comparing the syntactic tendencies of texts, to see if there are differences that we can see with statistical models. So far we have found some very interesting things. And we are looking forward to presenting our final results.

About seven months ago we, Cérise Muller (Applied Linguistics) and Dana Hakman (Information Studies) embarked on our interdisciplinary journey as academy assistants for the Network Institute. Together we started researching the knowledge bases that are used within
the professional field of bilingualism. With our research we are aiming at providing the best knowledge source possible for speech therapists working with bilingual children in the Netherlands.

The starting point of our research has been an existing knowledge base, in the form of a wiki, created by the Vrije Universiteit Amsterdam and extensive research into the needs of the target group. From there, we have created a new knowledge architecture to enhance the experience of knowledge acquisition for professionals in the field. It has been very rewarding to speak to current users of the wiki and receive constructive feedback on how to make the work of speech therapists easier and therefore help more children who are struggling with languages impairments.

Our current focus is on building the new website which we hope to publish as soon as possible and finishing our research paper. It has already been extremely motivating that our research was accepted for the ICT Open conference earlier this year and we hope to be able to present our finished work at a conference soon. So far it has been an invaluable opportunity to work as an interdisciplinary team and we are looking forward to what lies ahead.

Cerise & Dana

Cerise is currently also writing her thesis on the use of Google Translate in secondary education and will pursue an internship at VSI Subtitling.

Dana is currently also writing her thesis in the form on an internship at the SAS Institute (Computer Software) where she is focusing on the gamification of Data Science.

Finally, Friday 5 PM! A week full of meetings, appointments and work has come to an end. Your holiday has started! You’ve been looking forward to the holiday during which you can finally get some rest and do the things you like and have been wanting to do for months: explore that lovely city, have dinner in that restaurant you’ve heard so much about, enjoy the sunshine, and most of all: have a good time. You’ve already made several representations of it in your head. The holiday flies by. The city was nice, but also a little too crowded; the restaurant was closed; you’ve barely seen the sun, and when you get back you think: “This holiday did actually not meet the expectations I had beforehand”. Sounds familiar?

We, Benthe Spijkers and Suzana Bašić, are currently working on the "bright future" project as academy assistants for the Network Institute. Our study combines the fields of computational linguistics and social sciences to get insight into a possible optimistic bias in future thinking: are people in general more positive about their future than about their past? The project aims to get more insight into this optimistic bias by using sentiment analysis.

Several studies have found that people are more positive about the future than about the past when they are instructed to write about past or imagined future events. However, we wanted to analyse this social phenomenon in a more organic context, so we turned to the ultimate source of organically written human text – the World Wide Web. That is where the computational linguistics part of our project comes into play. The processing of large amounts of data requires the use of automated methods; it would hardly make sense to manually analyse millions and millions of social media messages. We believe that digital resources and methods can be invaluable tools for social science researchers, but we also believe that validating the method is of crucial importance.

Therefore, we are currently analysing the output of a commercial social media monitoring tool on the original and manipulated data in order to gain insight into the internal functioning of the tool. And don’t let the boring phrasing fool you – if you have ever taken something apart just to see how it works, you understand exactly how exciting it is! The only difference is that our tool is a black box, so we can’t actually take it apart, but that is why we are testing it on different datasets and analysing what happens every time, which makes it all the more exciting if you ask us!

We are still very busy working on the project, but we are very optimistic about our final results! 😉 What about you? Do you also often think the future has something positive in store for you?

As Academy Assistants, John Can Lokman (Artificial Intelligence) and Frederik Koenig (Political Science), we have been working on our project at the Network Institute together with Ali Khalili (AI) and Sascha Friesike (PS) for nearly six months now. From a conceptual perspective, our project focuses on interdisciplinarity in scientific research with the intention to explore the characteristics of interdisciplinarity and the effects it has on research. From a technical perspective, this translates into gathering bibliometric data from multiple sources (such as 'Web of Science') and creating a large database that houses a lot of metadata for each scientific publication. Using this data, we generate a citation network between articles, and a collaboration network between authors, institutions and disciplines. This allows us to explore patterns, such as scientific disciplines whose collaboration yields publications with more —or less— impact. We also plan to use this data to observe the evolution of such patterns over time. For example, to discover trends in scientific research, such as the early signs of the birth of a new discipline, which has begun its life as a collaboration between existing disciplines (an exciting possibility suggested by Prof. Peter van Besselaar [Social Sciences] during one of our meetings).

Although our responsibilities in the project are distinct, we meet on a regular basis, invite other experts to these meetings, and exchange our ideas. After six months, we carried the theoretical discussion forward and made significant technical progress. For both of us, this has been a challenging, rewarding, and academically enriching period, and we find it quite a nice addition to our studies to be Academy Assistants!

In our collaboration, Frederik and Sascha working on the theoretical part while John and Ali realizing the technical part. In the past months, Frederik completed a literature review and he is currently busy with finalizing his research on identifying the various dimensions of the impact of interdisciplinary research — a subject of debate in the literature, and a necessity for developing a way for accurate and comprehensive measurement. John, meanwhile, has been working on and recently completed a Python package that is capable of importing bibliographic records from multiple file formats (such as .bib, .csv, and .rdf files) into our central database, and then enriching each of these records with additional data from various places on the Internet using semantic web queries. The final component of our system is the Linked-Data Reactor (or 'LD-R', a framework being developed by Ali Khalili), which allows user-friendly exploration of our database on the web, and makes it accessible to users with no technical knowledge. This latter component alone, is an interesting technology and research topic in itself, as it lowers the threshold of collaboration between disciplines and makes data exploration a pleasure. We are now at the stage of integrating more data into this platform, and are about to begin analyses.

A closer look at our individual experiences

John Lokman
As a master student who dedicated this academic year to work on the Academy Assistant project, write his thesis, and find the PhD he's looking for —and not just any PhD—, I have been able to dedicate a good amount of time to the project and learn a lot of valuable skills meanwhile. The respite from the workload of simultaneous courses and the necessity to keep up with the course schedules, along with the financial help from the project budget, allowed me the luxury of a long and focused learning period. Furthermore, even though I ended up spending 30-40 hours on the project each week due to personal interest, the relatively lower pressure of a 10-hour per week agreement —and my supervisor's encouraging attitude & support— allowed me to be more explorative in my coding, and granted me precious time to bridge the skill/knowledge gaps I had due to being a programmer who came from a non-computer science background. I doubt that such an effective learning experience could have been possible if I did not have this rare-to-come-across balance of having a focus while having enough space to pursue it without having to rush things too much. For me, being able to give every action the time it deserved resulted in the feeling of working without working, work as play. This experience, repeated over the course of months, changed my view and practice of scholarship and by helping me to discover where my true career/research interests lie, it ultimately resulted in a certainty about my desired career path: research in computer science or a related discipline in a setting that allows me a deep and prolonged focus on a topic of my choice.

Frederik König
I had to dive deep in the extensive pool of literature on interdisciplinary research. Overall, I accessed more than 5000 papers and eventually selected 154 of them for my final review. That was truly a lot of reading on a topic that I had hardly known before starting the project. However, after some teething problems, I really started to enjoy my swim. Unexpectedly, many of the concepts I found were quite similar to what I had learned during my studies in Political Science. Moreover, the contact with researchers from different disciplines has truly broadened my horizon. Especially, the input of my team has encouraged me to further pursue my idea for an own tech-startup.

Five months have already whizzed by since we, Lesia Tkacz and Robin Kumar Sharma, began working as Network Institute Academy Assistants, but we are more excited than ever to be on the Enhancing Quality Assessment Using Perspective Detection project as it progresses!

Our project aims to alleviate the burden of information overload that web users can face while researching a topic online, by helping the users to be more informed about the information they consume. Our work builds on a preexisting tool that evaluates web documents on 5 quality dimensions (completeness, precision, neutrality, accuracy, readability, overall quality), which we contextualize by using Natural Language Processing technology to indicate what perspective a document’s content may express with regards to a controversial debate. We will convey this information through a visualization that we are developing, which will represent how groups of documents can cluster according to both the perspective that they may express and the quality of their information.

Designing this visualization is the most challenging part of the project, but we've formed a good habit for approaching it: every work morning, we grab coffee at the WN science building on the VU campus, bust out some pens and paper, and start discussing and sketching our latest ideas for the visualization. We explain our domain-specific approaches to each other, and work together to realize how best to visualize information quality and the perspective. We power this iterative process with both constructive feedback from our supervisors and from our own inspiration from research. In particular, it has been interesting to negotiate and combine our visions and knowledge of mathematics & computer science, and of language & art, towards a shared goal. While this visualization design phase is now drawing to a close, our project is progressing into a demo implementation phase, and we are looking forward to submitting a poster abstract with our preliminary results to the WebSci Amsterdam Conference.

It's very rewarding to see our plans take a more sophisticated shape over time.

While sometimes it may feel like we are making slow progress, it must be remembered that working once a week for 10 months is the equivalent of about 2 months of full time work. In this equivalent time, we will have created a multiple document quality and perspective visualization demo, an interesting approach to a language processing problem, given a few conference presentations, and most importantly, have had an invaluable opportunity to garner knowledge and experience working on a serious project with a supportive interdisciplinary team. It certainly feels like an achievement to be proud of.

Lesia & Robin

Lesia Tkacz is extending her areas of expertise in English Linguistics & Literature, and Fine Art, with a Research MA in Linguistics: Human Language Technology.

Robin Sharma is completing his Research MA in Computer Science, and is also working on the successful launch of a start-up project.

Both students plan to pursue a PhD.

Life after being an Academy Assistant

One of the goals of the Network Institute is to interest master students in conducting scientific research and pursuing an academic career. But how successful is the Network Institute in reaching this goal? To answer this question, we spoke to three former Academy Assistants to see how their life is like today and to find out if and how their Academy Assistant project have helped them in their careers.

1. What are you doing today?

Ellen Droog: I'm currently working as a junior researcher and teacher at the Communication Science department of the VU. I'm researching the metaphorical framing of cyberinfrastructure, which is a new and complex technology.

Enrico Rotundo: After my master in Artificial Intelligence at VU, I joined HAL24K a company focused on developing solutions based on machine learning to the real world. I'm Data Scientist and my role consists of analysing large datasets and designing data-driven products.

Marit Blank: I'm currently working as a junior teacher at the Department of Organization Sciences. I'm teaching several courses in the bachelor Administrative & Organization Sciences and the master Policy, Communication & Organization. Besides teaching I supervise bachelor and master theses.

2. Did the Academy assistant project help you in any way to get where you are today?

Ellen Droog: Yeah, it definitely did! Luckily we had two wonderful and very helpful supervisors in our project and one of them also supervised my Master Thesis. He also knew that I wanted to stay in academia and then one day he came to me and said: "I have a job for you"! So that is kind of the story of how I got this job, and if it wasn't for the Academy Assistant project, I wouldn't be here today.

Enrico Rotundo: My role requires a rational thinking and advanced research skills. I'm confident in saying the Academy Assistant experience had a significant role in building up such know-how. Indeed, some of my daily taks are actually very similar to what I did for the AA project like studying academic literature and implementing state-of-the-art alghorithms.

Marit Blank: Yes, it absolutely did. My supervisor asked me if I was interested in becoming a teacher at the department of Organization Sciences, so without the research project I wouldn't be having this job. I never really thought about staying in academia after graduating, but while I was doing this research project, I realized how interesting it was to work in academia and to be surrounded by seniors who can learn you a lot.

3. What did you like about being an Academy Assistant the most?

Ellen Droog: During and after writing my Bachelor Thesis I started to think about the possibility of staying in academia. And a few months later when I was starting my Master I saw a vacancy for an Academy Assistant project. And this project really provided me the opportunity to find out what conducting research is actually like and if an academic career would suit me. I also enjoyed working in a project team and it also resulted in a good friendship with my fellow Academy Assistant!

Enrico Rotundo: I was able to complete my master while working in a research setting with a part-time schedule, isn't that great?!

Marit Blank: What I liked the most was to present our research project at a conference in Germany. Together with my supervisor I went to Germany for 4 days and we attended many conference speeches. I met a lot of interesting people and collected new research ideas. Besides that, it was just a lot of fun with nice hotel, food, drinks and parties. Furthermore, it was very nice to work with Enrico from Computer Science because it's such a different discipline that I learned a lot from working with him.

4. What are your plans for the future?

Ellen Droog: I really want to become PhD student, but unfortunately PhD spots are really hard to find. But the research experience I gained due to the Academy Assistant project will probably help me when I apply for a PhD position!

Enrico Rotundo: I'm planning on improving my Dutch and visiting different areas of the country; I spent most of my time in the capital… and now I'm looking forward the upcoming spring season to explore the national parks of this wonderful country.

Marit Blank: I never really make any plans for the future. At this moment I like the teaching, but I also experience that education - in contrast to research - is not really a priority at the VU. As a junior teacher (who's not interested in pursuing a PhD) that really breaks my heart. So, maybe my future is outside academia, who knows!

Food waste happens everywhere: during production, storage and transport, but also in stores, restaurants and even at home. Even though consumers are said to waste the most food, supermarkets and producers play a key role in the food waste chain. Signalling this problem to the media is often taken up by non-governmental organisations and social movements, such as and . For their project, Academy Assistants Sarah Kamphuis and Melanie Jong Tjien Fa are studying how different stakeholders in the food (waste) chain are framed in the public debate. To do this, they will leverage automated content analysis techniques to produce longitudinal social and semantic network maps. These maps will show how the framing of food waste in the public debate changed over time, tracing the relative positioning of the different stakeholders.

In 60 seconds: Wasted food in numbers - BBC News Subscribe to BBC News www.youtube.com/bbcnews 1/3 of all food produced for human consumption is lost or wasted before it is eaten, the UN's Food and Agricult...

At the kick-off meeting of the academic year 2017-2018 all the new Academy Assistants were able to meet each other and hear all about the interdisciplinary projects of this year. With projects about language impairment, generic optimistic bias, biblical Hebrew, knowledge flows, food waste and quality assessment, it's definitely going to be an exciting year! You can read all about this year's projects on the website of the Network Institute. http://www.networkinstitute.org/academy-assistants/academy-projects-17/

Today Romy van den Heerik, Ellen Droog and Melanie Jong Tjien Fa presented their Academy Assistant research project about creative and constructive language production at the Anela Juniorendag!