Face the Facts! Evaluating RAG-based Fact-checking Pipelines in Realistic Settings
AI4TRUST partners FBK submitted a Paper with the goal to benchmark, under more realistic scenarios, RAG-based methods for the generation of verdicts i.e., short texts discussing the veracity of a claim – evaluating them on stylistically complex claims and heterogeneous, yet reliable, knowledge base in May 2025.
Improving the Perturbation-Based Explanation of Deepfake Detectors Through the Use of Adversarially-Generated Samples
AI4TRUST partners CERTH submitted a Paper introducing the idea of using adversarially-generated samples of the input images that were classified as deepfakes by a detector, to form perturbation masks for inferring the importance of different input features and produce visual explanations, at the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops (WACVW 2025) in Tucson, Arizona in February 2025.
VERGE in VBS 2025
AI4TRUST partners submitted a Paper which presents VERGE, an interactive system designed for searching and browsing video content. VERGE allows users to explore collections of images extracted from videos, offering various search options like free-text queries, concept-based searches, color matching, detection of faces and people, as well as searches based on visual and semantic similarities. This paper was presented at the 31st International Conference on MultiMedia Modeling (MMM2025) in Nara, Japan in January 2025.
ITI-CERTH participation in ActEV and AVS Tracks of TRECVID 2024
AI4TRUST partners submitted a Paper on the overview of the runs related to Ad-hoc Video Search (AVS) and Activities in Extended Video (ActEV) tasks at the TRECVID 2024 Workshop in November 2024.
Disturbing Image Detection Using LMM-Elicited Emotion Embeddings
AI4TRUST partners submitted a Paper dealing with the task of Disturbing Image Detection (DID), exploiting knowledge encoded in Large Multimodal Models (LMMs). at the IEEE Int. Conf. on Image Processing Workshops (ICIPW2024), in Abu Dhabi in October 2024.
Efficient training strategies for natural sounding speech synthesis and speaker adaptation based on FastPitch
AI4TRUST partners submitted a Paper at the 2024 IEEE 20th International Conference on Intelligent Computer Communication and Processing (ICCP 2024) , Cluj-Napoca, Romania, in October 2024.
Translating speech with just images
AI4TRUST partners presented a paper on visually grounded speech models link speech to images – linking images to text via an existing image captioning system, and as a result gain the ability to map speech audio directly to text, at Interspeech 2024, Kos, Greece in September 2024.
Hybrid-Diarization System with Overlap Post-Processing for the DISPLACE 2024 Challenge
AI4TRUST partners presented a paper at Interspeech 2024, Kos, Greece on the team’s collaborative efforts in participating in the Track 1 for Speaker Diarization of the Diarization of Speaker and Language in conversational Environments (DISPLACE) Challenge 2024, in September 2024.