Papers by Gianluca Moro
Can Large Language Models Win the International Mathematical Games? (2025.emnlp-main)
Copied to clipboard
Alessio Cocchieri, Luca Ragazzi, Giuseppe Tagliavini, Lorenzo Tordi, Antonella Carbonaro, Gianluca Moro
| Challenge: | Recent advances in large language models (LLMs) have demonstrated strong mathematical reasoning abilities, even in visual contexts. |
| Approach: | They propose a benchmark of 2,183 high-quality mathematical problems in an open-ended format that enables a structured evaluation of LLMs’ mathematical and logical reasoning abilities. |
| Outcome: | The new benchmark spans seven age groups and a skill-based taxonomy and enables a structured evaluation of LLMs’ mathematical and logical reasoning abilities. |
PORTS: Preference-Optimized Retrievers for Tool Selection with Large Language Models (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing retrieval-based methods to pre-select tools are often misaligned with tool-calling LLMs due to separate training processes. |
| Approach: | They propose a method to fine-tune retrievers to find useful tools by using a frozen LLM. |
| Outcome: | The proposed method fine-tunes retrievers to find useful tools using a frozen LLM . it improves tool selection accuracy and can be generalized to new queries and tools . |
Text-to-Text Extraction and Verbalization of Biomedical Event Graphs (2022.coling-1)
Copied to clipboard
| Challenge: | Biomedical events represent complex, graphical, and semantically rich interactions expressed in the scientific literature. |
| Approach: | They propose a framework to solve event extraction and event verbalization with a unified text-to-text approach. |
| Outcome: | The proposed framework achieves greater state-of-the-art performance than single-task competitors and can generate coherent natural language utterances from structured data. |
NLG-Metricverse: An End-to-End Library for Evaluating Natural Language Generation (2022.coling-1)
Copied to clipboard
| Challenge: | Natural language generation models are a key component of deep learning, says aaron eliott . he says it is crucial to develop and apply better metrics for NLG evaluation . |
| Approach: | a new open-source library for NLG evaluation is created to facilitate researchers to judge the effectiveness of their models. the framework provides a living collection of NLG metrics in a unified and easy-to-use environment. |
| Outcome: | a new open-source library for NLG evaluation aims to improve performance of models . the framework provides tools to apply, analyze, compare, and visualize the metrics . |
BioReader: a Retrieval-Enhanced Text-to-Text Transformer for Biomedical Literature (2022.emnlp-main)
Copied to clipboard
| Challenge: | Recent research has equipped language models with the ability to attend over relevant and factual information from non-parametric external sources, drawing a complementary path to architectural scaling. |
| Approach: | They propose a retrieval-enhanced text-to-text model that augments the input prompt by fetching and assembling relevant scientific literature chunks from a neural database centered on PubMed. |
| Outcome: | The proposed model outperforms state-of-the-art models on a broad array of downstream tasks while using up to 3x fewer parameters. |
OpenBioNER: Lightweight Open-Domain Biomedical Named Entity Recognition Through Entity Type Description (2025.findings-naacl)
Copied to clipboard
Alessio Cocchieri, Giacomo Frisoni, Marcos Martínez Galindo, Gianluca Moro, Giuseppe Tagliavini, Francesco Candoli
| Challenge: | Biomedical Named Entity Recognition (BioNER) is a computationally expensive and limited tool . specialized 7B NER LLMs and GPT-4o can't match textual spans with entity types . |
| Approach: | They propose a lightweight BERT-based cross-encoder architecture that can identify any biomedical entity using only its description. |
| Outcome: | The proposed system outperforms existing models that match textual spans with entity types rather than descriptions on biomedical benchmarks. |
Discriminative Marginalized Probabilistic Neural Method for Multi-Document Summarization of Medical Literature (2022.acl-long)
Copied to clipboard
| Challenge: | Existing solutions for multi-document summarization ignore potential summary-relevant contents, causing problems in the medical domain. |
| Approach: | They propose a discriminative marginalized probabilistic method to generate a multi-document summary from a cluster of topic-related medical documents using token probability marginalization. |
| Outcome: | The proposed method outperforms the current state-of-the-art on a biomedical dataset for multi-document summarization. |
To Generate or to Retrieve? On the Effectiveness of Artificial Contexts for Medical Open-Domain Question Answering (2024.acl-long)
Copied to clipboard
| Challenge: | Medical open-domain question answering requires substantial access to specialized knowledge. |
| Approach: | They propose a framework that generates multiple-choice questions from a set of open-book parameters and a small-scale reader that can outcompete closed-book questions by 706x using fewer parameters. |
| Outcome: | The proposed framework outperforms closed-book models on MedQA-USMLE, MedMCQA, and MMLU while using up to 706x fewer parameters. |
Sycophants in the Courtroom: Are LLMs Fragile to Juridical Authority and Evolving Legal Standards? (2026.acl-long)
Copied to clipboard
| Challenge: | Recent advances have seen large language models (LLMs) achieve remarkable performance across high-stakes specialized domains. |
| Approach: | They propose a diagnostic framework that evaluates legal reasoning against medical baselines along four axes (knowledge recall, grounding, confidence, and robustness) they uncover a sharp domain asymmetry when applied to a benchmark that encodes temporal validity and normative relationships. |
| Outcome: | The proposed framework evaluates legal reasoning against medical baselines along four axes (knowledge recall, grounding, confidence, and robustness) it shows that legal LLMs struggle to assess when retrieved citations are useful or misleading, exhibiting overconfidence in perturbed contexts and sensitivity to superficial formatting cues. |
What Are You Token About? Differentiable Perturbed Top-k Token Selection for Scientific Document Summarization (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing datasets suffer from a deficiency in source heterogeneity, hindering effective model training and generalizability. |
| Approach: | They propose a dataset that includes technical and lay summaries from multiple journals . they propose 'prinepert' transformer-based model that prunes irrelevant tokens in end-to-end learning. |
| Outcome: | The proposed model achieves a 2x speed-up compared to a state-of-the-art linear transformer, remaining comparable in effectiveness. |
ZeroNER: Fueling Zero-Shot Named Entity Recognition via Entity Type Descriptions (2025.findings-acl)
Copied to clipboard
Alessio Cocchieri, Marcos Martínez Galindo, Giacomo Frisoni, Gianluca Moro, Claudio Sartori, Giuseppe Tagliavini
| Challenge: | Existing zero-shot learning methods rely on entity type names for generalization . current solutions require large datasets and prioritize a handful of commonly occurring types . |
| Approach: | They propose a description-driven framework that enhances hard zero-shot NER in low-resource settings. |
| Outcome: | The proposed framework outperforms existing models by up to 16% in the F1 score . it also surpasses baseline models that use type names alone . |
LLMs (Almost) Never Abstain Under Medical Uncertainty (2026.acl-long)
Copied to clipboard
| Challenge: | Medical multiple-choice question answering (MCQA) benchmarks assume large language models should always commit to an answer. |
| Approach: | They propose a benchmark to evaluate medical abstention under uncertainty . they remove the gold answer and introduce an explicit "I abstain" option . results highlight abstinence as a critical but overlooked dimension of medical decision-making evaluation . |
| Outcome: | The new benchmark evaluates medical abstention under uncertainty. |
ReMedQA: Are We Done With Medical Multiple-Choice Benchmarks? (2026.eacl-long)
Copied to clipboard
| Challenge: | Multiple-choice question answering (MCQA) benchmarks show near-human accuracy . but a single accuracy score is a poor proxy for competence . |
| Approach: | They propose a medical multiple-choice question answering (MCQA) benchmark that augments three standard medical MCQA datasets with open-ended answers and systematically perturbed options. |
| Outcome: | The proposed benchmarks show that high MCQA accuracy masks low reliability . MCQ is the dominant paradigm for assessing medical knowledge in large language models . |
“What do you call a dog that is incontrovertibly true? Dogma”: Testing LLM Generalization through Humor (2025.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) have shown strong performance in NLP tasks like text summarization and question answering. |
| Approach: | They propose a new humor-based question-answering benchmark to assess LLMs’ reasoning through carefully crafted puns. |
| Outcome: | Experiments on pun comprehension, resolution, and generation reveal that most LLMs struggle with generalization, even on simple tasks, consistently underperforming the human baseline. |
MemeWeaver: Inter-Meme Graph Reasoning for Sexism and Misogyny Detection (2026.findings-eacl)
Copied to clipboard
| Challenge: | Existing methods to detect hate speech on social media are limited by heuristic graph construction, shallow modality fusion, and instance-level reasoning. |
| Approach: | They propose a multimodal framework for detecting sexism and misogyny using a graph reasoning mechanism that can be used to train multiple visual-textual fusion strategies. |
| Outcome: | The proposed framework outperforms state-of-the-art methods on MAMI and EXIST benchmarks while achieving faster training convergence. |
Unknown Claims: Generation of Fact-Checking Training Examples from Unstructured and Structured Data (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for fact-checking are labor-intensive and time-consuming. |
| Approach: | They propose a framework that generates training instances for FC systems automatically using textual and tabular content. |
| Outcome: | The proposed framework generates training instances for FC systems using textual and tabular content. |