Hateful and abusive speech presents a major challenge for all online social media platforms. Recent advances in Natural Language Processing and Natural Language Understanding allow for more accurate detection of hate speech in textual streams. This study presents a new multimodal approach to hate speech detection by combining Computer Vision and Natural Language processing models for abusive context detection. Our study focuses on Twitter messages and, more specifically, on hateful, xenophobic, and racist speech in Greek aimed at refugees and migrants. In our approach, we combine transfer learning and fine-tuning of Bidirectional Encoder Representations from Transformers (BERT) and Residual Neural Networks (Resnet). Our contribution includes the development of a new dataset for hate speech classification, consisting of tweet IDs, along with the code to obtain their visual appearance, as they would have been rendered in a web browser. We have also released a pre-trained Language Model trained on Greek tweets, which has been used in our experiments. We report a consistently high level of accuracy (accuracy score = 0.970, f1-score = 0.947 in our best model) in racist and xenophobic speech detection.

Floods are lethal and destructive natural hazards. The Mediterranean, including Greece, has recently experienced many flood events (e.g., Medicanes Zorbas and Ianos), while climate change results in more frequent and intense flood events. Accurate flood mapping in river areas is crucial for flood risk assessment, planning mitigation measures, protecting existing infrastructure, and sustainable planning. The accuracy of results is affected by all simplifying assumptions concerning the conceptual and numerical model implemented and the quality of geospatial data used (Digital Terrain Models—DTMs). The current research investigates flood modelling sensitivity against geospatial data accuracy using the following DTM resolutions in a mountainous river sub-basin of Thessaly’s Water District (Greece): (a) open 5 m and (b) 2 m data from Hellenic Cadastre (HC) and (c) 0.05 m data from an Unmanned Aerial Vehicle (UAV) topographical mission. RAS-Mapper and HEC-RAS are used for 1D (steady state) hydraulic simulation regarding a 1000-year return period. Results include flood maps and cross section-specific flow characteristics. They are analysed in a graphical flood map-based empirical fashion, whereas a statistical analysis based on the correlation matrix and a more sophisticated Machine Learning analysis based on the interpretation of nonlinear relationships between input–output variables support and particularise the conclusions in a quantifiable manner.

In this paper we describe our work towards building a generic framework for both multi-modal embedding and multi-label binary classification tasks, while participating in task 5 (Multimedia Automatic Misogyny Identification) of SemEval 2022 competition. Since pretraining deep models from scratch is a resource and data hungry task, our approach is based on three main strategies. We combine different state-of-the-art architectures to capture a wide spectrum of semantic signals from the multi-modal input. We employ a multi-task learning scheme to be able to use multiple datasets from the same knowledge domain to help increase the model's performance. We also use multiple objectives to regularize and fine tune different system components.