Pure OAI Repository
Not a member yet
421387 research outputs found
Sort by
Investigating the effect of incorporated stimuli responsive groups on amphiphilic polymeric nanoparticles’ structure and its biological inte
No full textSingle-chain polymeric nanoparticles (SCPNs), formed via intramolecular folding of an amphiphilic polymer chain, offer potential in versatile applications like drug delivery and bio-orthogonal catalysis. In this project, we aim to obtain on-demand control on the particle conformation in aqueous media via incorporation of various stimuli-responsive moieties. The resulting variations in the polymer microstructure are envisaged to steer the particle interactions with biological surroundings such as proteins, receptors or cellular membranes. Hereto, we have developed nanometre-sized polymeric particles comprising photoswitches or H-bonding moieties and are currently investigating their structure-interaction relationship. Elucidating their size, shape and compactness is of great importance to understand the influence of the introduced stimuli responsive groups on the particle microstructure and in line their biological interactions
Investigating the effect of incorporated stimuli responsive groups on amphiphilic polymeric nanoparticles’ structure and its biological inte
No full textSingle-chain polymeric nanoparticles (SCPNs), formed via intramolecular folding of an amphiphilic polymer chain, offer potential in versatile applications like drug delivery and bio-orthogonal catalysis. In this project, we aim to obtain on-demand control on the particle conformation in aqueous media via incorporation of various stimuli-responsive moieties. The resulting variations in the polymer microstructure are envisaged to steer the particle interactions with biological surroundings such as proteins, receptors or cellular membranes. Hereto, we have developed nanometre-sized polymeric particles comprising photoswitches or H-bonding moieties and are currently investigating their structure-interaction relationship. Elucidating their size, shape and compactness is of great importance to understand the influence of the introduced stimuli responsive groups on the particle microstructure and in line their biological interactions
Panoptic-VSNet:Visual-semantic prior knowledge-driven multimodal 3D panoptic segmentation
No full textPrecise and robust perception is critical for ensuring the safe operation of autonomous vehicles. However, current methods are constrained by sparse image-LiDAR alignment, insufficient annotations, and ineffective structural discrepancy modeling, causing semantic degradation and generalization deficiency. Therefore, we propose Panoptic-VSNet, a visual-semantic prior knowledge-driven multimodal 3D panoptic segmentation network. Firstly, we design a progressive fusion semantic alignment module that effectively aggregates visual prior features obtained from the large Visual-Language model, establishing a point-semantic region association, thereby enhancing semantic awareness. Secondly, we propose an instance-aware superpixel cross-modal fusion module that incorporates instance prior knowledge, forming a unified representation with spatial precision and class consistency. Finally, we introduce a correlation-aware adaptive panoptic segmentation network that reduces parameter count while dynamically capturing contextual information and enhancing local details, thereby improving panoptic perception capabilities. Experimental evaluations on benchmark datasets show that Panoptic-VSNet outperforms state-of-the-art methods. Code is available at https://github.com/lixiao0125/panoptic-vsnet.git.</p
Panoptic-VSNet:Visual-semantic prior knowledge-driven multimodal 3D panoptic segmentation
No full textPrecise and robust perception is critical for ensuring the safe operation of autonomous vehicles. However, current methods are constrained by sparse image-LiDAR alignment, insufficient annotations, and ineffective structural discrepancy modeling, causing semantic degradation and generalization deficiency. Therefore, we propose Panoptic-VSNet, a visual-semantic prior knowledge-driven multimodal 3D panoptic segmentation network. Firstly, we design a progressive fusion semantic alignment module that effectively aggregates visual prior features obtained from the large Visual-Language model, establishing a point-semantic region association, thereby enhancing semantic awareness. Secondly, we propose an instance-aware superpixel cross-modal fusion module that incorporates instance prior knowledge, forming a unified representation with spatial precision and class consistency. Finally, we introduce a correlation-aware adaptive panoptic segmentation network that reduces parameter count while dynamically capturing contextual information and enhancing local details, thereby improving panoptic perception capabilities. Experimental evaluations on benchmark datasets show that Panoptic-VSNet outperforms state-of-the-art methods. Code is available at https://github.com/lixiao0125/panoptic-vsnet.git.</p
How does the layout of indoor communal spaces in low-income high-rise apartment buildings impact the social interactions between residents?
No full textThe scarcity of urban land restricts outdoor communal spaces, especially in low-income high-rise complexes. As a result, indoor communal spaces become essential in stimulating social interaction among high-rise residents. This study, therefore, investigates how the layout of indoor communal space in low-income high-rise apartment buildings affects social interactions between residents using a combination of comparative floorplan-analysis and semi-structured interviews. The investigation draws on data collected from four low-income high-rise building complexes in Hanoi, Vietnam. Space syntax analyses, including the visual integration and connectivity drawn from Visibility Graph Analysis (VGA), were used for the comparative floorplan-analysis. Thematic analysis was used to analyze semi-structured interviews to understand residents’ perceptions and use of communal spaces for social interaction. Findings indicate that residents are more likely to use functional communal spaces when they are accessible, visible, and flexible. Small-scale corridors and sequences of spaces can provide a platform for interaction and, at the same time, guarantee privacy. The seating options, advertisement screen display, (digital) community noticeboards in the common area, and good lighting/ventilation can invite people to stay longer and interact with others. The results of this study offer design implications, suggesting that indoor-communal spaces are significant for social interaction
An Assembly-Line Mechanism for In-Vitro Encapsulation of Fragmented Cargo in Virus-Like Particles
No full textThe ability of virus shells to encapsulate a wide range of functional cargoes, especially multiple cargoes─siRNAs, enzymes, and chromophores─has made them an essential tool in biotechnology for advancing drug delivery applications and developing innovative new materials. Since therapeutic cargo may be formulated in different physical states (size, surface charge, etc.) we have investigated the spontaneous encapsulation of multiple, charged, small nanoparticles which repel when free in solution, inside a spherical cage formed of brome mosaic virus (BMV) coat proteins. Unlike the cases of coassembly of virus-like particles (VLPs) from multiple oligonucleotides and single nanoparticle cargo, the structure of virus-like particles thus obtained is consistent with that of the native icosahedral BMV capsid. Working with small metal nanoparticles as cargo allowed the pathway of assembly to be followed by electron and liquid atomic force microscopy and cryoelectron tomography. Based on the structural identification of nanoparticle─BMV protein intermediates, we have found that multiple cargo encapsulation occurs in stages through a specific "assembly line" pathway that is different from the previously described in vitro assembly mechanisms of virus-like particles (VLP). We propose a model that explains the experimental findings, some of which will be important for delivery applications, for instance, the pronounced nanoparticle size selectivity in competition experiments where different nanoparticle sizes are present.</p
SPartAn Visualization:Design of a Tool for Interpreting the SPartAn Planning Optimization Model Outcomes
No full textExploring the Socio-Epistemic Dynamics of Scientific Inquiry through Agent-Based Models
No full textThe success of scientific inquiry is fundamentally a collective achievement. Although scientific research can be driven by individual researchers, it is only through communication, disagreement, and shared evaluation that knowledge accumulates and matures. Yet, the relationship between individual-level practices and differences and collective epistemic outcomes is far from straightforward. Practices that appear rational or beneficial to individual agents can, in some contexts, hinder the epistemic goals of the community. Conversely, certain seemingly flawed or irrational behaviours may unintentionally promote collective progress. This tension lies at the heart of my dissertation. To assess what truly counts as good scientific practice, we must go beyond traditional case studies or thought experiments. We need tools capable of capturing the complexity of social interaction, the heterogeneity of agents, and the indirect effects of individual behaviour. Agent-based models (ABMs) provide precisely this capacity. By simulating scientific communities composed of interacting, boundedly rational agents, ABMs allow us to explore how micro-level reasoning gives rise to macro-level epistemic patterns. The thesis is structured around six chapters, grouped into two thematic parts. Part I investigates the impact of diverse background assumptions. Scientists often operate with divergent aims, values, and theoretical frameworks. In some cases, scientists misinterpret evidence due to incorrect assumptions. In others, they apply equally legitimate but conflicting standards. The first two chapters model evidence misinterpretation. Chapter 2 extends the Hegselmann-Krause model to show that excess data can intensify polarisation when scientists interpret evidence differently. Chapter 3 uses a bandit model framework to show that evidence misinterpretation can lead to epistemic benefits. Finally, Chapter 4 employs an NK model to investigate value and standard pluralism, a scenario of scientific inquiry in which scientists hold different but equally legitimate standards. It finds that interaction with agents with different standards can enhance epistemic performance, highlighting the epistemic benefits of evaluative diversity. Part II addresses the epistemic consequences of reasoning biases. Building on the work of Mercier and Sperber, who identified flaws such as myside bias and selective laziness in human reasoning, I examine how these biases play out in collective scientific deliberation. Chapter 5 surveys the landscape of argumentative agent-based models, ABMs that represent and simulate argument exchange. Chapter 6 explores the effect of myside bias and its mitigation through shared epistemic norms, finding that bias can be offset by background diversity and agreement on argumentative standards. Chapter 7 shows how the combination of laziness and bias can enhance collective epistemic performance under constraints of time or information, echoing ideas from the literature on fast and frugal heuristics. Together, these results offer a nuanced picture of the relationship between individual reasoning and community-wide epistemic success. They challenge conventional wisdom about what counts as virtuous scientific behaviour and show how certain individual vices can be epistemically productive at scale
OfficeSense:Fostering Situated Data Sensemaking to Enhance Office Well-being
No full textOffice sensor systems often fail to present environmental data related to office well-being in ways that are accessible and comprehensible for non-expert users. To address this, we introduce Situated Data Sensemaking, a concept that enhances individuals’ ability to understand and use data by embedding it within their physical environment. Implemented through OfficeSense, a system that physicalizes environmental parameters—light levels, sound, air quality (CO2), and temperature—within office spaces, this concept was evaluated in a 4-week study (N=11) including a baseline week. OfficeSense influenced perceived data literacy and understanding of office well-being. Findings indicate that Situated Data Sensemaking supported participants in aligning their perceptions of environmental conditions with real-time data, influenced by spatial and temporal contexts, personal preferences, and collaborative interactions. This study demonstrates that embedding data in shared physical spaces can enhance users’ confidence in interpreting environmental data and foster collective awareness of workplace conditions
3D-technologie in de periferie:Ontwerpen van een systeem waarbinnen de kwaliteit en toekomstbestendigheid van het gebruik van 3D technologie binnen Gelre gewaarborgd is
No full textDe toepassing van 3D-technologie in de zorg groeit snel en biedt mogelijkheden voor gepersonaliseerde zorg, zoals preoperatieve planningen en anatomische modellen. De implementatie ervan vraagt echter om investeringen in expertise, processen en naleving van wet- en regelgeving, waaronder de MDR. Gelre ziekenhuizen gebruikt 3D-technologie nog beperkt en wil inzicht krijgen in kwalitatieve en toekomstbestendige organisatievorm.Het doel van dit project is het ontwerpen van een systeem waarmee de kwaliteit en toekomstbestendigheid van het gebruik van 3D‑technologie binnen Gelre ziekenhuizen kan worden gewaarborgd. Dit wordt gerealiseerd door een theoretisch haalbaarheidsadvies waarin vier scenario’s worden vergeleken: volledig in-huis, volledig extern of twee hybride vormen. Het advies ondersteunt Gelre ziekenhuizen bij het bepalen welk scenario het beste aansluit op strategie, middelen en regelgeving.In het haalbaarheidsadvies zijn processen, risico’s, eisen, oplossingen en kosten per scenario uitgewerkt. Op basis hiervan is een kostendashboard ontwikkeld waarmee verschillende 3D-toepassingen vergeleken kunnen worden. Het project biedt tevens adviezen voor een passend kwaliteitssysteem, benodigde software en hardware en de benodigde competenties van de medewerkers voor een professioneel 3D-proces