A5080524660- Video Analysis and Summarization
- Recommender Systems and Techniques
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Educational Games and Gamification
- Topic Modeling
- Multimodal Machine Learning Applications
- Multimedia Communication and Technology
- Semantic Web and Ontologies
- Digital Games and Media
- Web Data Mining and Analysis
- Context-Aware Activity Recognition Systems
- Music and Audio Processing
- Personal Information Management and User Behavior
- Augmented Reality Applications
- Innovative Human-Technology Interaction
- Data Stream Mining Techniques
- Emergency and Acute Care Studies
- Advanced Bandit Algorithms Research
- Data Visualization and Analytics
- Anomaly Detection Techniques and Applications
- Data Quality and Management
- Misinformation and Its Impacts
- Mobile Crowdsensing and Crowdsourcing
- Human Pose and Action Recognition
Koblenz University of Applied Sciences
2022-2024
Universität Koblenz
2022-2024
University of Sheffield
2018-2023
University of Koblenz and Landau
2022
Neusoft (China)
2018
University of Glasgow
2006-2017
Glasgow Life
2016
Technische Universität Berlin
2013-2015
Dublin City University
2004-2013
Clarity Centre for Sensor Web Technologies
2012-2013
Releasing datasets has fostered research in fields such as information retrieval and recommender systems. Datasets are typically tailored for specific scenarios. In this work, we present the plista dataset. The dataset contains a collection of news articles published on 13 portals. Additionally, comprises user interactions with those articles. We inctroduce dataset's main characteristics. Further, illustrate possible applications
Test collections have a long history of supporting repeatable and comparable evaluation in Information Retrieval (IR). However, thus far, no shared test collection exists for IR systems that are designed to index retrieve multimodal lifelog data. In this paper we introduce the first personal data, which has been employed NTCIR12-Lifelog task. paper, requirements motivated, process creating is described, along with an overview collection. Finally suggestions given possible applications
As user-generated contents thrive, so does the spread of toxic comment. Therefore, detecting comment becomes an active research area, and it is often handled as a text classification task. recent popular methods for tasks, pre-trained language model-based are at forefront natural processing, achieving state-of-the-art performance on various NLP tasks. However, there paucity in studies using such classification. In this work, we study how to best make use performances different models these...
Evaluation in empirical computer science is essential to show progress and assess technologies developed. Several research domains such as information retrieval have long relied on systematic evaluation measure progress: here, the Cranfield paradigm of creating shared test collections, defining search tasks, collecting ground truth for these tasks has persisted up until now. In recent years, however, several new challenges emerged that do not fit this very well: extremely large data sets,...
The CLEF NewsREEL challenge is a campaign-style evaluation lab allowing participants to evaluate and optimize news recommender algorithms. goal create an algorithm that able generate items users would click, respecting strict time constraint. challenges compete in either "living lab" (Task 1) or perform replays recorded streams 2). In this report, we discuss the objectives of lab, summarize last year's campaign outline main research can be addressed by participating 2016.
A portable, wireless system capable of in situ reagent-based colorimetric analysis is demonstrated. The based on a reconfigurable low cost optical detection method employing paired emitter detector diode device, which allows wide range centrifugal microfluidic layouts to be implemented. Due the communication, acquisition parameters can controlled remotely and results downloaded distant locations displayed real time. stand-alone capabilities system, combined with portability provide...
In the CLEF NEWSREEL lab, participants are invited to evaluate news recommendation techniques in real-time by providing recommendations actual users that visit commercial portals satisfy their information needs. A central role within this lab is communication between and users. This enabled The Open Recommendation Platform (ORP), a web-based platform which distributes users' impressions of articles returns readers. demo, we illustrate show how requests handled provide relevant real-time.
There have been multiple calls for integrating topics related to fairness, accountability, transparency, ethics (FATE) and social justice into Data Science curricula, but little exploration of how this might work in practice. This paper presents the findings a collaborative auto-ethnography (CAE) engaged by MSc teaching team based at University Sheffield (UK) Information School where FATE/Critical Studies (CDS) core part curriculum since 2015/16. In paper, we adopt CAE approach reflect on...
Evaluation in empirical computer science is essential to show progress and assess technologies developed. Several research domains such as information retrieval have long relied on systematic evaluation measure progress: here, the Cranfield paradigm of creating shared test collections, defining search tasks, collecting ground truth for these tasks has persisted up until now. In recent years, however, several new challenges emerged that do not fit this very well: extremely large data sets,...
COVID-19 infected millions of people and increased mortality worldwide. Patients with suspected utilised emergency medical services (EMS) attended departments, resulting in pressures waiting times. Rapid accurate decision-making is required to identify patients at high-risk clinical deterioration following infection, whilst also avoiding unnecessary hospital admissions. Our study aimed develop artificial intelligence models predict adverse outcomes by EMS clinicians.Linked ambulance service...
Detecting violent scenes in movies is an important video content understanding functionality e.g., for providing automated youth protection services. One key issue designing algorithms violence detection the choice of discriminative features. In this paper, we employ mid-level audio features and compare their power against low-level visual We fuse these cues with ones at decision level order to further improve performance detection. use Mel-Frequency Cepstral Coefficients (MFCC) as average...
The aim of this study is to survey reading habits users an online news portal. assumption motivating that insight into the can be helpful design better recommendation systems. We estimated transition probabilities who read article one category will move another (not necessarily distinct) category. For this, we analyzed users' click behavior within plista data set. Key findings are popularity local, loyalty readers same category, observing similar results when addressing enforced streams, and...