A5039642040- Misinformation and Its Impacts
- Translation Studies and Practices
- Ethics and Social Impacts of AI
- Data-Driven Disease Surveillance
- Semantic Web and Ontologies
- Cancer-related Molecular Pathways
- Protein Degradation and Inhibitors
- Opinion Dynamics and Social Influence
- Multimedia Communication and Technology
- Visual Attention and Saliency Detection
- Theatre and Performance Studies
- Neutrophil, Myeloperoxidase and Oxidative Mechanisms
- Cinema and Media Studies
- Sentiment Analysis and Opinion Mining
- Aesthetic Perception and Analysis
- Image Retrieval and Classification Techniques
- Fashion and Cultural Textiles
- COVID-19 Digital Contact Tracing
- Color perception and design
- Human Mobility and Location-Based Analysis
- Muscle Physiology and Disorders
- Face recognition and analysis
- Categorization, perception, and language
- Advanced Database Systems and Queries
- Evolutionary Psychology and Human Behavior
Cyprus Institute of Neurology and Genetics
1996-2022
Citard Services (Cyprus)
2019-2021
Active Technologies (Italy)
2019-2020
Wageningen University & Research
2016
Hellenic Cooperative Oncology Group
2004
University of Cyprus
2002
There are increasing expectations that algorithms should behave in a manner is socially just. We consider the case of image tagging APIs and their interpretations people images. Image taggers have become indispensable our information ecosystem, facilitating new modes visual communication sharing. Recently, they widely available as Cognitive Services. But while offer developers an inexpensive convenient means to add functionality creations, most opaque proprietary. Through cross-platform...
Machine-learned computer vision algorithms for tagging images are increasingly used by developers and researchers, having become popularized as easy-to-use "cognitive services." Yet these tools struggle with gender recognition, particularly when processing of women, people color non-binary individuals. Socio-technical researchers have cited data bias a key problem; training datasets often over-represent contexts that convey social stereotypes. The psychology literature explains learn...
Myotonic dystrophy type 1 (DM1) is the most common form of adult-onset muscular dystrophy, which characterised by progressive muscle wasting and discovery reliable blood-based biomarkers could be useful for disease progress monitoring. There have been some reports showing that presence specific miRNAs in blood correlates with DM1. In one these, our group identified four muscle-specific miRNAs, miR-1, miR-133a, miR-133b miR-206, correlated progression observed DM1 patients. The levels were...
Abstract In the healthcare sector, phytocompounds are known to be beneficial by contributing or alleviating a variety of diseases. Studies have demonstrated progressive effects on immune-related diseases and exhibit anticancer effects. Graviola tree is an evergreen with its extracts (leafs seeds) been reported having properties, but precise target action not clear. Using in silico approach, we predicted that annonacin, Acetogenin , active agent found leaf extract (GLE) potentially act as...
Crowdsourcing plays a key role in developing algorithms for image recognition or captioning. Major datasets, such as MS COCO Flickr30K, have been built by eliciting natural language descriptions of images from workers. Yet elicitation tasks are susceptible to human biases, including stereotyping people depicted images. Given the growing concerns surrounding discrimination algorithms, well data used train them, it is necessary take critical look at this practice. We conduct experiments Figure...
Image analysis algorithms have become an indispensable tool in our information ecosystem, facilitating new forms of visual communication and sharing. At the same time, they enable large-scale socio-technical research which would otherwise be difficult to carry out. However, their outputs may exhibit social bias, especially when analyzing people images. Since most are proprietary opaque, we propose a method auditing for biases. To able compare how interpret controlled set images, collected...
Vision-based cognitive services (CogS) have become crucial in a wide range of applications, from real-time security and social networks to smartphone applications. Many focus on analyzing people images. When it comes facial analysis, these can be misleading or even inaccurate, raising ethical concerns such as the amplification stereotypes. We analyzed popular Image Tagging CogS that infer emotion person's face, considering whether they perpetuate racial gender stereotypes concerning emotion....
Abstract Algorithms have greatly advanced and become integrated into our everyday lives. Although they support humans in daily functions, often exhibit unwanted behaviors perpetuating social stereotypes, discrimination, other forms of biases. Regardless their accuracy on task, many algorithms do not get scrutinized for unintended a systematic way. This phenomenon can propagate amplify existing societal issues or even create new ones. Many called human supervision (human oversight)...
We report two cases of compound heterozygote patients for the --(MED I) and Hb Agrinio [alpha29(B10)Le-->uPro (alpha2)] anomalies in unrelated Greek Cypriot families. The first patient had a serious form H disease died at age 21 due to complications arising during an operation. second showed severe hematological picture has been regularly transfused since early age. This exhibits bone abnormalities as well hepatosplenomegaly. severity these incidences emphasizes need inclusion screening test...
Image analysis algorithms have been a boon to personalization in digital systems and are now widely available via easy-to-use APIs. However, it is important ensure that they behave fairly applications involve processing images of people, such as dating apps. We conduct an experiment shed light on the factors influencing perception "fairness." Participants shown photo along with two descriptions (human- algorithm-generated). They then asked indicate which "more fair" context site, explain...
Modern information access systems extensively use personalization, automatically filtering and/or ranking content based on the user profile, to guide users most relevant material. However, this can also lead unwanted effects such as "filter bubble." We present an interactive demonstration system, designed educational and research tool, which imitates a search engine, personalizing results returned for query user's characteristics. The system be tailored suit any type of audience context,...
Following the literature on dehumanization via technology, we audit six proprietary image tagging algorithms (ITAs) for their potential to perpetuate dehumanization. We examine ITAs' outputs a controlled dataset of images depicting diverse group people tags that indicate presence human in image. Through an analysis (mis)use these tags, find there are some individuals whose 'humanness' is not recognized by ITA, and often from marginalized social groups. Finally, compare findings with use...
Abstract Evaluating the algorithmic behavior of interactive systems is complex and time-consuming. Developers increasingly recognize importance accountability for their creations’ unanticipated resulting implications. To mitigate this phenomenon, developers not only need to concentrate on observable inaccuracies that can be measured quantitatively but also more subjective outcomes perpetuate social bias, which are challenging identify. We require a new approach involves humans in...
The aim of this work was to develop a system based on modular neural networks and multi-feature texture analysis that facilitates the automated interpretation cloud images. This speeds up process provides continuity in application satellite imagery for weather forecasting. A series infrared images from geostationary METEOSAT7 were employed. Nine different feature sets (a total 55 features) extracted segmented using following algorithms: first order statistics, spatial gray level dependence...
The present randomized phase III trial was designed to detect a 15% benefit in relapse-free survival (RFS) or overall (OS) from the incorporation of adjuvant tamoxifen combination CNF [cyclophosphamide, 500 mg/m2; mitoxantrone (Novantrone), 10 fluorouracil, mg/m2 chemotherapy and ovarian ablation premenopausal patients with node-positive breast cancer conversely postmenopausal patients. From April 1992 until March 1998, 456 operable one nine infiltrated axillary nodes entered study....
We segment the structural units of electron microscope muscle images using a novel AM-FM image representation. This approach is shown to be effective in describing sarcomeres and mitochondrial regions images.
"Filter bubbles," a phenomenon in which users become caught an information space with low diversity, can have various negative effects. Several tools been created to monitor the users' actions make them aware of their own filter bubbles, but these disadvantages (e.g., infringement on privacy). We propose standalone demo that does not require any personal data. It emulates Facebook, well-known and popular social network. demonstrate how each user interaction may affect selection subsequent...