Scientific innovations and societal developments happen in rapid succession: quantum computers, AI, the pandemic… When communicating about such developments with non-expert audiences, journalists often use metaphors. Metaphors describe abstract, unfamiliar concepts in terms of more concrete, more accessible concepts (e.g. AI as a friend). The metaphors we choose influence perceptions and decision-making, making it crucial to study their use in media texts.
However, manual metaphor identification is very labor-intensive. On average, one in six words in news texts is used metaphorically. However, most metaphors will not be used to describe the topic of interest (e.g. AI). To speed up this process, various methods relying on supervised machine learning have been proposed. While these can achieve acceptable performance on data from the same source as the training data, performance tends to drop on data covering new topics. It is unlikely that a supervised system could identify metaphors it has never seen in its training data. Even recent models (e.g., ChatGPT), perform poorly.
To resolve this issue, we develop a hybrid computationally-assisted approach to metaphor identification, combining simple and transparent NLP methods with targeted manual annotation. The goal is to limit manual labor by identifying passages about the target topic containing potential metaphors automatically. First, we identify passages about our target topics semi-automatically. Then, we identify metaphors in these passages (initially based on manual analysis). Text passages and metaphors identified as such will subsequently be fed back into the system, creating a loop to find more relevant passages and metaphors. This approach combines machine learning with human feedback and verification procedures to combat blind spots. We aim to develop a standardized procedure applicable to various domains. As use-cases, we focus on news about covid, climate technologies and AI.
Researchers:
-
-
Gudrun Reijnierse
-
-
Pia Sommerauer
Mental health illnesses have long been among the most serious and prevalent public health issues. In particular, depression is a leading cause of disability and significantly increases the risk of suicidal ideation and suicide attempts. According to Statistics Netherlands’ review of 2023, young people’s mental health has not fully recovered since the pandemic. Numerous surveys have shown a rise in mental health issues compared to two years earlier, especially among young women. Specifically, 18 percent of young people in the Netherlands (aged 12 to 24) were considered mentally unhealthy in 2023, up from 11 percent in 2019 and 2020.
With the increasing availability of data on individuals’ mental health, artificial intelligence (AI) and machine learning (ML) techniques are being used to deepen our understanding of mental health conditions and enhance patient care. These technologies are also aiding clinicians and doctors in making more informed clinical decisions.
In recent years, Graph Neural Networks (GNNs) that focus on functional magnetic resonance imaging (fMRI) data have become popular for diagnosing depression in both AI and medical communities. However, existing GNN-based approaches face challenges related to explainability and generalization. Specifically, GNNs are often viewed as black boxes with poor explainability, meaning clinicians do not understand the rationale behind their decisions. Additionally, these trained machine learning models typically exhibit poor generalization, performing well in local hospitals but experiencing significant performance drops when tested on data from other hospitals.
This project aims to advance existing AI approaches from two perspectives. First, we aim to develop a new, explainable GNN to enhance both explainability and generalization. Second, we aim to further improve the detection accuracy by incorporating other modalities of neuroimaging data, such as structural magnetic resonance imaging (sMRI). Thus, creating a multi-modal depression detection system is another key objective.
Researchers:
-
-
Shujian Yu
-
-
Guido van Wingen
Against the backdrop of several societal crises currently taking place in the Netherlands – the climate crisis, the nitrogen crises, several health-related crises, and the housing crisis – a potential new crisis is already in the making: congestion. Since COVID-19, commuting has picked up like never before, traffic jams are at record length, and public transportation is creeping toward full capacity. A crisis of congestion would not merely be one of daily annoyance and discomfort, it would also be one that strains our infrastructure, environment, climate, health, and labor productivity. Calls to increase capacity (i.e. widen roads, lengthen or heighten train) are ill-informed. Such interventions only attract more traffic and also are increasingly inefficient given that most infrastructure is underused during slow hours. The only real – and evidencebased – solution to congestion problems is peak spreading.
Peak spreading – like other transitions – requires behavioral transformations in citizens; transformations that can best be instigated by combining economic and psychological insights into motivations for, and resistance toward, change. This is where the interdisciplinary team behind this project adds its value: Erik Verhoef is a leading researcher in transport economics, with a proven track record in developing solutions for urban transport. Ivar Vermeulen is a persuasion researcher, studying not only how attitudes and behavior can be changed, but also how resistance to change can be effectively addressed.
The current project proposes to combine these insights to design a large-scale peak spreading experiment in one of the key hubs of Dutch congestion (the Amsterdam Zuidas), and in its key traveler hub (the VU). The project team will be supported by two public transport companies (GvB; NS) to obtain traveler volume data. The team has will also request from the VU directorship, in line with its determination of being responsible in containing the university’s negative societal impacts.
Researchers:
-
-
Erik Verhoef
-
-
Ivar Vermeulen
The Dutch Republic (1588-1795) was a leading empire with settlements in America, Africa and Asia, and remained a colonial power as a Kingdom until its decolonization after World War II (1815-1962/1975). State and church closely worked together in sending professional ministers and lay pastors through the East Indian and West Indian Companies to the overseas territories. After the revolutionary era, missionary organizations did the same for more than a century. Thousands of Dutch men from middle and lower class – together with wives and children! – went to one or more world regions and a part of them returned to patria, building a long tradition of cultural exchange of ideas, practices and objects. In spite of a strong Dutch colonial historiography, the religious part of the story is not paid much attention. This VU-project will contribute to a structural development of an urgent research agenda.
As spin-off from a former NWO-Vernieuwingsimpulsproject, the first applicant composed a biographical database of all Dutch Reformed ministers between 1555 and 2004. This database was recently curated, enriched and linked to other data in the public domain with the Netherlands e-Science Center, projecting a Digital Dutch Religion Portal 1500-2000 (NLESC.SSIH.2022a.SSIH009). The main applicant has also made up a database of all religious workers in the Dutch colonies, initially for the early modern period (1602-1795/1811), later extended to the mission era (1797/1815-1961). The ‘colonial’ database deserves to be added to the digital research infrastructure as well, especially given the public interest in colonialism, racism and slavery in Dutch history. Making available the data on lives and careers of circa 7000 ‘missionaries’ in the Dutch colonial world empire will open up many research challenges in the fields of general and religious history in connection to digital analysis of biographical and geographical data, archival and printed texts and sources.
Researchers:
-
-
Fred van Lieburg
-
-
Victor de Boer
Immersive technologies are rapidly reshaping our communication methods. Immersive experiences within virtual environments can elicit profound emotional responses and capture the viewer’s attention, enabling the more effective transmission of messages and ideas. VR technologies have been suggested as potentially effective tools to enhance awareness about environmental problems (Thoma et al., 2023), for example, by helping individuals visualize and comprehend future climate scenarios, otherwise perceived as distant and abstract. However, it is unclear whether VR is more effective in sustaining behavior change (Horen et al., 2024; Ferreira & Banerjee, 2024).
This project aims to assess whether VR technology can positively change multiple human behaviours in an environmental context. One way to study this is by evaluating spillover effects of behavioral interventions (Galizzi & Whitmarsh, 2019). Spillover effects assess whether individuals who have adopted more sustainable behaviour in one domain (eating less meat) will also become more susceptible to changing their behaviour in another domain (reduced energy consumption).
We propose developing an environment where participants interact with various scenarios to increase awareness and support behavior change regarding the environment through digital intervention. This will be tested using a randomized control trial with four conditions: (1) a 2D environment on a computer monitor, (2) a 3D environment in virtual reality (VR), (3) a placebo scenario unrelated to the environment in VR, and (4) a control condition where participants read scenario text. Participants in each condition will make two environmental choices, allowing us to determine the direct impact of VR versus 2D technology and any indirect spillover effects.
The study consists of multiple methods: we administer a digital intervention experiment embedded in a longitudinal study to measure real behaviors over time, using surveys to measure participants’ baseline characteristics and pre- and post-experimental attitudes. We will use state-of-the-art econometric methods to provide causal estimates of our treatments.
Researchers:
-
-
Charlotte Gerritsen
-
-
Sanchayan Banerjee
In an age of concerns about misinformation, audience fragmentation, and polarization, it has become more urgent than ever to understand how news and media impact the way people perceive the world. While the proliferation of media devices, platforms, and information sources has made it challenging to capture people’s news and information use, paradoxically, the digitalization and platformization of society have simultaneously enabled unprecedented access into people’s practices and experiences. This project harnesses these developments via state-of-the-arts methods in communication science and media/journalism studies. Specifically, it combines data donations with WhatsApp-diaries and in-depth interviews to answer the following question: How and under which circumstances do societal issues ‘break through’ the continuous stream of information and resonate with citizens?
To answer this question, first, participants (N=30) will fill in a longitudinal (three-month) diary, weekly reflecting on the societal issues they notice. They will also submit contextual information: Where did they encounter this issue? Did they discuss it with anyone? What do they think about it? Afterward, these participants will donate their digital trace data – spanning the same period of three months – from platforms such as Google, YouTube, Instagram and X. For this, we will use Digital Data Donation Infrastructure (D3I), developed by a consortium led by Theo Araujo (UvA), which enables individuals to download their data from any organization that tracks them and to donate these data to academic research. Comparing the digital trace data and the diary entries will allow us to better understand if, when and how issues people encounter online actually enter their lifeworld (i.e. the world as they subjectively experience it). Creative semi-structured interviews (including data walkthroughs) at the end will help us further make sense of participants’ practices and experiences. Mixed-methods empirical inquiry into everyday meaning-making is crucial for better understanding how media function in today’s society.
Researchers:
-
-
Tim Groot Kormelink
-
-
Kasper Welbers