Yen-Chia Hsu

Scarcity of pixel-level labels is a significant challenge in practical scenarios. In specific domains like industrial smoke, acquiring such detailed annotations is particularly difficult and often requires expert knowledge. To alleviate this, weakly supervised semantic segmentation (WSSS) has emerged as a promising approach. However, due to the supervision gap and inherent bias in models trained with only image level labels, existing WSSS methods suffer from limitations such as incomplete foreground coverage, inaccurate object boundaries, and spurious correlations, especially in our domain, where emissions are always spatially coupled with chimneys. Previous solutions typically rely on additional priors or external knowledge to mitigate these issues, but they often lack scalability and fail to address the model's inherent bias toward co-occurring context. To address this, we propose a novel WSSS framework that directly targets the co-occurrence problem without relying on external supervision. Unlike prior methods that adopt a single network, we employ a teacher-student framework that combines CNNs and ViTs. We introduce a knowledge transfer loss that enforces cross-architecture consistency by aligning internal representations. Additionally, we incorporate post-processing techniques to address partial coverage and further improve pseudo mask quality.

Digitally-supported participatory methods are often used in policy-making to develop inclusive policies by collecting and integrating citizen's opinions. However, these methods fail to capture the complexity and nuances in citizen's needs, i.e., citizens are generally unaware of other's needs, perspectives, and experiences. Consequently, policies developed with this underlying gap tend to overlook the alignment of multistakeholder perspectives, and design policies based on the optimization of high-level demographic features. In our contribution, we propose a method to enable citizens understand other's perspectives and calibrate their positions. First, we collected requirements and design principles to develop our approach by involving stakeholders and experts in policymaking in a series of workshops. Then, we conducted a crowdsourcing study with 420 participants to compare the effect of different text and images, on people's initial and final motivations and their willingness to change opinions. We observed that both influence participant's opinion change, however, the effect is more pronounced for textual modality. Finally, we discuss overarching implications of designing with empathy to mediate alignment of citizen's perspectives.

By deploying machine learning (ML) and deep learning (DL) algorithms, we address the problem of smell event modelling in the Pittsburgh metropolitan area. We use the Smell Pittsburgh dataset to develop a model that can reflect the relation between bad smell events and industrial pollutants in a specific urban territory. The initial dataset resulted from crowd-sourcing citizen reports using a mobile phone application, which we categorised in a binary matter (existence or absence of smell events). We investigate the mapping of smell data with air pollution levels that were recorded by a reference station located in the southeastern area of the city. The initial dataset is processed and evaluated to produce an updated dataset, which is used as an input to assess various ML and DL models for modelling smell events. The models utilise a set of air quality and climate data to associate them with a smell event to investigate to what extent these data correlate with unpleasant odours in the Pittsburgh metropolitan area. The model results are satisfactory, reaching an accuracy of 69.6, with ML models mostly outperforming DL models. This work also demonstrates the feasibility of combining environmental modelling with crowd-sourced information, which may be adopted in other cities when relevant data are available.

The use of Computer Vision, through a Perceptive Visual Urban Analytics (VUA) paradigm, has been proposed as a way for municipalities to more easily monitor their cities. However, prior studies fall short of actually investigating whether Perceptive VUA is ready for municipal use. In this paper we take a critical look at this paradigm by comparing key methods and evaluating them on usability and trustworthiness with municipal experts as well as Responsible AI and Computer Vision researchers. Based on on this evaluation we find that Perceptive VUA is not (yet) ready for municipal use as they do not incorporate domain knowledge and overly rely on spurious correlations. We conclude by providing recommendations for how to progress Perceptive VUA such that it may actually contribute to improving the liveability and quality of urban environments.

The EmpathiCH Workshop aims to blend a diverse set of expertise to expand upon the nascent field of Empathy-Centric Design. Building on the discussions in previous editions of the workshop, this iteration invites contributions which scrutinize the use of empathy as a design principle in digital interfaces. We encourage inquiry in a number of research dimensions: examining the multifaceted nature of empathy; establishing both the requirements and shortcomings of empathy in design research; discussing key post-human stakeholders in digital interfaces (social groups, causes, digital avatars, artificial agents etc.); and expanding the scope of empathy research beyond preliminary perspective-taking. The workshop, structured as a combination of author panels, expert discussion, and interactive activities, provides the ideal venue to foster a critical discussion on the nature of the suitability of empathy in digital design, especially in the rapidly approaching context of its role in post-humanist HCI.

Artificial Intelligence (AI) is increasingly used to analyze large amounts of data in various practices, such as object recognition. We are specifically interested in using AI-powered systems to engage local communities in developing plans or solutions for pressing societal and environmental concerns. Such local contexts often involve multiple stakeholders with different and even contradictory agendas, resulting in mismatched expectations of these systems' behaviors and desired outcomes. There is a need to investigate if AI models and pipelines can work as expected in different contexts through co-creation and field deployment. Based on case studies in co-creating AI-powered systems with local people, we explain challenges that require more attention and provide viable paths to bridge AI research with citizen needs. We advocate for developing new collaboration approaches and mindsets that are needed to co-create AI-powered systems in multi-stakeholder contexts to address local concerns.

Empowering communities through citizen science.