Papers by Michalis Vazirgiannis
Abstractive Meeting Summarization: A Survey (2023.tacl-1)
Copied to clipboard
| Challenge: | Recent advances in deep learning have improved language generation systems, opening the door to improved forms of abstractive summarization. |
| Approach: | They propose to use neural encoder-decoder architectures to generate abstractive meeting summarizations that are particularly well-suited for multi-party conversation. |
| Outcome: | The proposed system could be used in a wide variety of real-world contexts, from business meetings to medical consultations to customer service calls. |
GreekMMLU: A Native-Sourced Multitask Benchmark for Evaluating Language Models in Greek (2026.findings-acl)
Copied to clipboard
Yang Zhang, Mersin Konomi, Christos Xypolopoulos, Konstantinos Divriotis, Konstantinos Skianis, Giannis Nikolentzos, Giorgos Stamou, Guokan Shang, Michalis Vazirgiannis
| Challenge: | Existing evaluation benchmarks for large language models are limited for Greek . Existing datasets are often machine-translated from English, failing to capture Greek linguistic and cultural characteristics. |
| Approach: | They propose a native-sourced benchmark for massive multitask language understanding in Greek . they publicize 16,857 samples and reserve 4,948 samples for a private leaderboard . |
| Outcome: | The proposed model is based on 21,805 multiple-choice questions across 45 subject areas . the model is publicly released and reserved for a private leaderboard . |
DATScore: Evaluating Translation with Data Augmented Translations (2023.findings-eacl)
Copied to clipboard
| Challenge: | Experimental results show that DATScore correlates better with human meta-evaluations than the other recent state-of-the-art metrics. |
| Approach: | They propose to use data augmented translations to improve the evaluation of machine translations by using two new scoring strategies. |
| Outcome: | The proposed metric improves on 3 NLG tasks other than translation. |
Unsupervised Word Polysemy Quantification with Multiresolution Grids of Contextual Embeddings (2021.eacl-main)
Copied to clipboard
| Challenge: | a new method to quantify polysemy is based on basic geometry in the contextual embedding space . word sense annotation has always been one of the tasks with the lowest interannotator agreement . |
| Approach: | They propose a method to estimate polysemy based on simple geometry in contextual embedding space. |
| Outcome: | The proposed method is fully unsupervised and data-driven . it can be used to sample sentences with different senses at no extra cost . |
The Curious Decline of Linguistic Diversity: Training Language Models on Synthetic Text (2024.findings-naacl)
Copied to clipboard
| Challenge: | a new study examines the effects of training language models on synthetic data generated by their predecessors. |
| Approach: | They propose to use recursive finetuning techniques to assess linguistic diversity of models. |
| Outcome: | The proposed metrics show a decrease in diversity of model outputs through successive iterations, especially for tasks demanding high levels of creativity. |
BARThez: a Skilled Pretrained French Sequence-to-Sequence Model (2021.emnlp-main)
Copied to clipboard
| Challenge: | Inductive transfer learning has taken the entire NLU field by storm, with models such as BERT and BART setting new state-of-the-art on countless tasks. |
| Approach: | They introduce a large-scale pretrained seq2seq model for French that is very competitive with state-of-the-art BERT-based French language models such as CamemBERT and FlauBERT. |
| Outcome: | The proposed model outperforms existing models on discriminative and generative tasks on a French summarization dataset. |
Beyond Random Sampling: Efficient Language Model Pretraining via Curriculum Learning (2026.eacl-long)
Copied to clipboard
| Challenge: | Curriculum learning has improved efficiency across machine learning domains, but remains underexplored for language model pretraining. |
| Approach: | They present a systematic investigation of curriculum learning in LLM pretraining . they use vanilla curriculum learning, pacing-based sampling, and interleaved curricula . |
| Outcome: | The proposed framework accelerates convergence in early and mid-training phases, reducing training steps by 18-45% to reach baseline performance. |
Lost in the Mix: Evaluating LLM Understanding of Code-Switched Text (2026.acl-long)
Copied to clipboard
| Challenge: | Code-switching (CSW) is widespread in multilingual communities and increasingly prevalent in online content. |
| Approach: | They propose a pipeline for producing linguistically grounded CSW variants of established benchmarks across five typologically diverse languages. |
| Outcome: | The proposed model sets show that inserting non-English tokens into English reduces accuracy on comprehension and reasoning benchmarks, whereas embedding English into non- English contexts often improves it. |
FrugalScore: Learning Cheaper, Lighter and Faster Evaluation Metrics for Automatic Text Generation (2022.acl-long)
Copied to clipboard
| Challenge: | Existing evaluation metrics are not reliable, but require significant computational resources. |
| Approach: | They propose a method to learn a fixed, low cost version of any expensive NLG metric while retaining most of its original performance. |
| Outcome: | The proposed approach retains most of the original performance while running faster and faster. |
GreekBART: The First Pretrained Greek Sequence-to-Sequence Model (2024.lrec-main)
Copied to clipboard
Iakovos Evdaimon, Hadi Abdine, Christos Xypolopoulos, Stamatis Outsios, Michalis Vazirgiannis, Giorgos Stamou
| Challenge: | Transfer learning has revolutionized the fields of Computer Vision and Natural Language Processing. |
| Approach: | They introduce a new language model, GreekBART, that is based on a BART-base architecture. |
| Outcome: | The proposed model outperforms BERT, GPT and other transformer-based models on discriminative tasks. |
FREDSum: A Dialogue Summarization Corpus for French Political Debates (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Recent advances in deep learning have improved the performance of abstractive summarization systems. |
| Approach: | They present a dataset of french political debates to enhance resources for multi-lingual dialogue summarization. |
| Outcome: | The proposed dataset will be made publicly available for use by the research community. |
Bias in the Mirror : Are LLMs opinions robust to their own adversarial attacks (2025.acl-long)
Copied to clipboard
| Challenge: | Existing work on large language models lacks robustness, highlighting the limitations of such models. |
| Approach: | They propose a novel approach where two LLMs engage in self-debate to persuade a neutral version of the model. |
| Outcome: | The proposed approach examines whether large language models are robust during interactions and whether they are susceptible to reinforcing misinformation or shifting to harmful viewpoints. |
Speaker-change Aware CRF for Dialogue Act Classification (2020.coling-main)
Copied to clipboard
| Challenge: | Recent work in Dialogue Act (DA) classification approaches the task as a sequence labeling problem, using neural network models coupled with a Conditional Random Field (CRF) as the last layer. |
| Approach: | They propose to modify the CRF layer to take speaker-change into account and learn meaningful transition patterns conditioned on speaker-changing DA labels. |
| Outcome: | The proposed model outperforms the original model with wide margins for some DA labels. |
Unsupervised Abstractive Meeting Summarization with Multi-Sentence Compression and Budgeted Submodular Maximization (P18-1)
Copied to clipboard
Guokan Shang, Wensi Ding, Zekun Zhang, Antoine Tixier, Polykarpos Meladianos, Michalis Vazirgiannis, Jean-Pierre Lorré
| Challenge: | a novel graph-based framework for abstractive meeting speech summarization is developed . instead of grammatical, well-segmented sentences, the input is made of often ill-formed and ungrammatically ungrammatized text fragments called utterances. |
| Approach: | They propose a graph-based framework for abstractive meeting speech summarization that is fully unsupervised and does not rely on annotations. |
| Outcome: | The proposed framework improves on the state-of-the-art on the AMI and ICSI corpus. |
Evaluation of Greek Word Embeddings (2020.lrec-1)
Copied to clipboard
| Challenge: | Word embeddings are the most popular input for many NLP tasks. |
| Approach: | They propose to use Greek word embeddings as an unsupervised learning tool . they use a Greek word analogy test set and a morphological test collection to evaluate word similarities . |
| Outcome: | The proposed model is able to create meaningful representations of Greek words . the proposed model can be adapted to Greek language and polysemy . |
LLM as a Broken Telephone: Iterative Generation Distorts Information (2025.acl-long)
Copied to clipboard
| Challenge: | Large language models are increasingly responsible for online content, but they can be distorted by repeated transmission. |
| Approach: | They investigate whether large language models distort information through iterative generation. |
| Outcome: | The findings raise important questions about the reliability of LLM-generated content in iterative workflows. |
Questioning the Validity of Summarization Datasets and Improving Their Factual Consistency (2022.emnlp-main)
Copied to clipboard
| Challenge: | Abstractive summarization systems have a lack of a defined definition for the task . factual consistency is a key factor in summarizing, but there are still deficiencies . a new study shows that summarized summarisation models achieve improved performance . |
| Approach: | They propose a filtered summarization dataset with improved factual consistency to address this problem . they argue that the dataset should become a valid benchmark for developing and evaluating summarizing systems . |
| Outcome: | The proposed model improves on a popular summarization dataset with improved factual consistency. |
Fast-Decoding Diffusion Language Models via Progress-Aware Confidence Schedules (2026.findings-acl)
Copied to clipboard
| Challenge: | *SchED* is a training-free, model-agnostic early-exit algorithm that terminates diffusion decoding using a progress-aware confidence threshold. |
| Approach: | They propose a training-free, model-agnostic early-exit algorithm that terminates diffusion decoding using a progress-aware confidence threshold. |
| Outcome: | The proposed algorithm achieves 4 speedups on instruction-tuned models while maintaining baseline performance on average. |
Energy-based Self-attentive Learning of Abstractive Communities for Spoken Language Understanding (2020.aacl-main)
Copied to clipboard
| Challenge: | Abstractive community detection is an important spoken language understanding task, whose goal is to group utterances according to whether they can be jointly summarized by a common abstractive sentence. |
| Approach: | They propose a neural contextual utterance encoder with three types of self-attention mechanisms and train it using the siamese and triplet energy-based meta-architectures. |
| Outcome: | The proposed system outperforms multiple energy-based and non-energy based baselines on the AMI corpus. |
Automatic Analysis of Substantiation in Scientific Peer Reviews (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing systems to analyze peer reviews' quality are inadequate due to the increasing workload of reviewers and the lack of domain experts . |
| Approach: | They propose to use a claim-evidence pair extraction problem to analyze substantiation in peer reviews and train an argument mining system to do the same. |
| Outcome: | The proposed system could be used by conference managers and reviewers to analyze the quality of peer reviews. |
Scalable graph-based method for individual named entity identification (D19-53)
Copied to clipboard
| Challenge: | Named entity recognition (NED) is a method for identifying named entities within a knowledge base. |
| Approach: | They propose a method for individual identification requiring few annotated data samples. |
| Outcome: | The proposed method is well-motivated for integration in real systems. |