MADec 18, 2025
Don't Guess, Escalate: Towards Explainable Uncertainty-Calibrated AI Forensic AgentsGiulia Boato, Andrea Montibeller, Edward Delp et al.
AI is reshaping the landscape of multimedia forensics. We propose AI forensic agents: reliable orchestrators that select and combine forensic detectors, identify provenance and context, and provide uncertainty-aware assessments. We highlight pitfalls in current solutions and introduce a unified framework to improve the authenticity verification process.
DCJan 15, 2019
Blockchain enabled fog structure to provide data security in IoT applicationsMozhdeh Farhadi, Daniele Miorandi, Guillaume Pierre
IoT provides services by connecting smart devices to the Internet, and exploiting data generated by said devices to enable value-added services to individuals and businesses. In such cases, if data is exposed, tampered or lost, the service would not behave correctly. In this article, we discuss data security in IoT applications across five dimensions: confidentiality, integrity, authenticity, non-repudiation and availability. We discuss how distributed ledger technology could be used to overcome these issues and propose to use a fog computing architecture as decentralized computational support to deploy the ledger.
IRApr 6, 2018
A Wikipedia-based approach to profiling activities on social mediaChristian Torrero, Carlo Caprini, Daniele Miorandi
Online user profiling is a very active research field, catalyzing great interest by both scientists and practitioners. In this paper, in particular, we look at approaches able to mine social media activities of users to create a rich user profile. We look at the case in which the profiling is meant to characterize the user's interests along a set of predefined dimensions (that we refer to as categories). A conventional way to do so is to use semantic analysis techniques to (i) extract relevant entities from the online conversations of users (ii) mapping said entities to the predefined categories of interest. While entity extraction is a well-understood topic, the mapping part lacks a reference standardized approach. In this paper we propose using graph navigation techniques on the Wikipedia tree to achieve such a mapping. A prototypical implementation is presented and some preliminary results are reported.