Chamezopoulos Savvas

University of Amsterdam

Master's of Science on Information Studies: Data Science

Aristotle University of Thessaloniki

Bachelor of Engineering with integrated Masters of Engineering in Electrical and Computer Engineering

Certificate of Proficiency in English

Issued by Univeristy of Michigan

Diplome d'etudes en Langue Française DELF B2

Issued by La Commision National du DELF et du DALF

Greek

Mother Tongue

Overview of the DagPap24 Shared Task on Detecting Automatically Generated Scientific Paper

Association for Computational Linguistics · Aug 16, 2024

Abstract: This paper provides an overview of the 2024 ACL Scholarly Document Processing workshop shared task on the detection of automatically generated scientific papers. Unlike our previous task, which focused on the binary classification of whether scientific passages were machine-generated or not, one likely use case for text generation technology in scientific writing is to intersperse human-written text with passages of machine-generated text. We frame the detection problem as a multiclass span classification task: given an expert of text, label token spans in the text as human-written or machine-generated. We shared a dataset containing excerpts from human-written papers as well as artificially generated content collected by Elsevier publishing and editorial teams. As a test set, the participants were provided with a corpus of openly accessible human-written as well as generated papers from the same scientific domains of documents. The shared task saw 457 submissions across 28 participating teams and resulted in three published technical reports. We discuss our findings from the shared task in this overview paper.

Evaluating approaches to identifying research supporting the United Nations Sustainable Development Goals

Quantitative Science Studies · May 15, 2024

Abstract: The United Nations (UN) Sustainable Development Goals (SDGs) challenge the global community to build a world where no one is left behind. Recognizing that research plays a fundamental part in supporting these goals, attempts have been made to classify research publications according to their relevance in supporting each of the UN’s SDGs. In this paper, we outline the methodology that we followed when mapping research articles to SDGs and which is adopted by Times Higher Education in its Social Impact rankings. We compare our solution with other existing queries and models mapping research papers to SDGs. We also discuss various aspects in which the methodology can be improved and generalized to other types of content apart from research articles. The results presented in this paper are the outcome of the SDG Research Mapping Initiative, which was established as a partnership between the University of Southern Denmark, the Aurora European Universities Alliance (represented by Vrije Universiteit Amsterdam), the University of Auckland, and Elsevier to bring together broad expertise and share best practices on identifying research contributions to UN’s Sustainable Development Goals.

Article Classification with Graph Neural Networks and Multigraphs

ELRA and ICCL · May 30, 2023

Abstract: Classifying research output into context-specific label taxonomies is a challenging and relevant downstream task, given the volume of existing and newly published articles. We propose a method to enhance the performance of article classification by enriching simple Graph Neural Networks (GNN) pipelines with edge-heterogeneous graph representations. SciBERT is used for node feature generation to capture higher-order semantics within the articles' textual metadata. Fully supervised transductive node classification experiments are conducted on the Open Graph Benchmark (OGB) ogbn-arxiv dataset and the PubMed diabetes dataset, augmented with additional metadata from Microsoft Academic Graph (MAG) and PubMed Central, respectively. The results demonstrate that edge-heterogeneous graphs consistently improve the performance of all GNN models compared to the edge-homogeneous graphs. The transformed data enable simple and shallow GNN pipelines to achieve results on par with more complex architectures. On ogbn-arxiv, we achieve a top-15 result in the OGB competition with a 2-layer GCN (accuracy 74.61%), being the highest-scoring solution with sub-1 million parameters. On PubMed, we closely trail SOTA GNN architectures using a 2-layer GraphSAGE by including additional co-authorship edges in the graph (accuracy 89.88%)

The CLIN33 Shared Task on the Detection of Text Generated by Large Language Models

Computational Linguistics in the Netherlands Journal · Mar 21, 2023

Abstract: The Shared Task for CLIN33 focuses on a relatively novel yet societally relevant task: the detection of text generated by Large Language Models (LLMs). We frame this detection task as a binary classification problem (LLM-generated or not), using test data from up to 6 different domains and text genres for both Dutch and English. Part of this test data was held out entirely from the contestants, including a ”mystery genre” that belonged to an unknown domain (later revealed to be columns). Four teams submitted 11 runs with substantially different models and features. This paper gives an overview of our task setup and contains the evaluation and detailed descriptions of the participating systems. Notably, included in the winning systems are both deep learning models as well as more traditional machine learning models leveraging task-specific feature engineering.

Link Prediction in Signed Social Networks: The Case of Bitcoin Users

XIV Balkan Conference on Operational Research · Dec 3, 2020

Abstract: During the last decade, social networks appear in many aspects of modern life. By their nature, these networks are dynamic objects and, thus, questions have been emerged regarding their evolvement during time. The availability of large datasets encoding network information along with the novel machine learning algorithms/solutions have made possible the extensive study of social network properties and structural features. In our work, we study the well-studied link prediction problem which seeks to accurately predict future possible links on the network or missing links due to incomplete data. The most common modelling approach is to represent these networks as graphs, where the nodes represent entities while the edges/links represent the association between entities. We focus on weighted signed social networks and try to predict new edges in a real-world dataset. Specifically, a Bitcoin network is being employed where different users rate the level of trust (on a scale ranging from - 10 to 10, excluding 0) they have in other users. Three different frameworks for representation learning on large graphs have been used; namely, Node2Vec, CDTNE and GraphSage. Following, standard steps involved in supervised learning, the performance of the selected learning functions have been measured using well-known metrics (e.g. accuracy, precision, AUC-score) for each implementation strategy employed in our analysis. All three employed techniques are compared under the aforementioned Bitcoin-related network and the results provide distinct useful insights on the network’s future formation. Additionally, the same methodologies are applied to a well-known dataset citation network of scientific publications (known as CORA dataset) in order to validate further the conclusions of the preceding analysis. Finally, we discuss how the different methodologies regarding network embeddings and link prediction frameworks can be combined effectively to achieve better results regarding the link prediction problem.