Previously known as Department of Intelligent Computer Systems https://www.um.edu.mt/library/oar/handle/123456789/8368 2026-09-10T19:38:03Z 2026-09-10T19:38:03Z Rašajski, Nemanja Makantasis, Konstantinos Liapis, Antonios Yannakakis, Georgios https://www.um.edu.mt/library/oar/handle/123456789/148974 2026-09-09T09:25:06Z 2026-01-01T00:00:00Z

Title: InvPatch : prefix-based conditional generation for inverse dynamics Authors: Rašajski, Nemanja; Makantasis, Konstantinos; Liapis, Antonios; Yannakakis, Georgios Abstract: Inverse Dynamics Models (IDMs) attempt to predict the actions that cause observable changes in a scene. Current methods for building accurate IDMs either rely on rule-based systems that exploit video metadata or require large-scale training over thousands of video hours. These approaches are inevitably limited to a single domain due to the difficulty of acquiring metadata or sufficient training data across different domains. In response to these challenges, this study draws inspiration from data-efficient video captioning methods, specifically prefix-based conditional generation. This approach maps visual features into prefix tokens that condition the action-prediction process. We introduce InvPatch, a framework that builds on prefix-based conditional generation and extends it by adding learned visual-representation compression. In InvPatch, attention-based patch selection and pooling are applied to features extracted from a ViT backbone, reducing the conditioning input from a set of frame-by-frame features to a single vector. We evaluate our framework across two diverse settings: 3D third-person real world (KIT Bimanual Actions) and 2D synthetic (MUGEN). Our method achieves 97.21% accuracy on MUGEN and surpasses state-of-the-art results on KIT Bimanual Actions. Our sensitivity analysis highlights the data efficiency of this approach, as InvPatch maintains comparable performance even when trained with 30% less data. Additionally, our runtime analysis demonstrates the computational efficiency of InvPatch, which requires fewer trainable parameters and performs fewer FLOPs compared to other methods.

2026-01-01T00:00:00Z Bartolo, Matthias Hili, Gabriel Seychell, Dylan Montebello, Matthew Debono, Carl Formosa, Saviour Makantasis, Konstantinos https://www.um.edu.mt/library/oar/handle/123456789/148973 2026-09-09T08:33:24Z 2026-01-01T00:00:00Z

Title: Litter detection from aerial imagery : a review of UAV-based approaches and deep learning techniques Authors: Bartolo, Matthias; Hili, Gabriel; Seychell, Dylan; Montebello, Matthew; Debono, Carl; Formosa, Saviour; Makantasis, Konstantinos Abstract: Litter pollution remains a pressing environmental issue, motivating the search for automated monitoring methods that can scale effectively. The problem is heightened by the difficulty of detecting small and diverse litter objects across wide areas, prompting interest in Unmanned Aerial Vehicles (UAVs) and deep learning as viable solutions. We conducted a systematic literature review by initially using Google Scholar, and then iteratively expanding our search through bibliographic references to identify relevant studies and datasets. In this review, we: synthesize the applicability of nine publicly available litter datasets; compile and analyze computer vision integrations with Litter Management Systems, with a particular emphasis towards UAV-based solutions; and review relevant literature addressing this issue; among others. Our analysis includes four UAV-based datasets (BDW, UAVVaste, HAIDA, SODA) and five non-UAV datasets (TrashNet, TACO, MJU-Waste, PlastOPol, ZeroWaste), examining dataset characteristics, preprocessing techniques, model architectures, and evaluation metrics across studies. Our synthesis of the literature highlights the varied approaches different studies undertake, reflecting the complexity of the task and the absence of standardised protocols. We conclude by discussing priorities for future work, notably the need for more publicly available in-the-wild UAV-acquired datasets and the potential of newer model architectures to address current limitations in automated litter detection.

2026-01-01T00:00:00Z Zhang, Ao Liu, Cong Makantasis, Konstantinos Chen, Xiaomin Ward, Tomas Cheng, Long https://www.um.edu.mt/library/oar/handle/123456789/148972 2026-09-09T08:07:47Z 2025-01-01T00:00:00Z

Title: DuMES : deep reinforcement learning‐based EV charging scheduling with dual‐layer safety modules Authors: Zhang, Ao; Liu, Cong; Makantasis, Konstantinos; Chen, Xiaomin; Ward, Tomas; Cheng, Long Abstract: Deep reinforcement learning (DRL) has become a promising approach for electric vehicle (EV) charging scheduling. However, its practical deployment poses potential risks to power infrastructure. DRL relies on trial-and-error interactions during training to approximate optimal policies, which may lead to unsafe decisions. To address this, a novel framework called dual-layer safety modules for EV charging scheduling (DuMES) is proposed. This framework introduces a decision-level safety layer into the conventional DRL architecture that adaptively detects and replaces unsafe actions. Furthermore, by integrating dual safety layers with reward shaping, the framework promotes convergence between raw and safe actions. This enhances training efficiency while ensuring power system stability during both training and deployment phases. The method was evaluated through simulation experiments on a charging station equipped with renewable energy and energy storage system (ESS). Comparative analyses with baseline methods demonstrate that DuMES effectively satisfies user charging demands, reduces operational costs and ensures compliance with safety constraints.

2025-01-01T00:00:00Z Pinitas, Kosmas Makantasis, Konstantinos Yannakakis, Georgios https://www.um.edu.mt/library/oar/handle/123456789/148971 2026-09-09T08:00:14Z 2025-01-01T00:00:00Z

Title: Privileged contrastive pretraining for multimodal affect modelling Authors: Pinitas, Kosmas; Makantasis, Konstantinos; Yannakakis, Georgios Abstract: Affective Computing (AC) has made significant progress with the advent of deep learning, yet a persistent challenge remains: the reliable transfer of affective models from controlled laboratory settings (in-vitro) to uncontrolled real-world environments (in-vivo). To address this challenge we introduce the Privileged Contrastive Pretraining (PriCon) framework according to which models are first pretrained via supervised contrastive learning (SCL) and then act as teacher models within a Learning Using Privileged Information (LUPI) framework. PriCon both leverages privileged information during training and enhances the robustness of derived affect models via SCL. Experiments conducted on two benchmark affective corpora, RECOLA and AGAIN, demonstrate that models trained using PriCon consistently outperform LUPI and end to end models. Remarkably, in many cases, PriCon models achieve performance comparable to models trained with access to all modalities during both training and testing. The findings underscore the potential of PriCon as a paradigm towards further bridging the gap between in-vitro and in-vivo affective modelling, offering a scalable and practical solution for real-world applications.

2025-01-01T00:00:00Z