204 citations · 223 across the 11 of their papers we have counts for
14 papers
On the Token Collapse Phenomenon in Textual Graph Soft Prompting
Andrea Giuseppe Di Francesco, Rishabh Jain, Franco Maria Nardini +3
Recent methods for injecting graph structure into LLM prompts include graph soft prompting, in which a graph neural network (GNN) learns graph-specific tokens that capture relation…
A Picture of Agentic Search
Francesca Pezzuti, Ophir Frieder, Fabrizio Silvestri +2
With automated systems increasingly issuing search queries alongside humans, Information Retrieval (IR) faces a major shift. Yet IR remains human-centred, with systems, evaluation…
Statistical Foundations of DIME: Risk Estimation for Practical Index Selection
Giulio D'Erasmo, Cesare Campagnano, Antonio Mallia +3
High-dimensional dense embeddings have become central to modern Information Retrieval, but many dimensions are noisy or redundant. Recently proposed DIME (Dimension IMportance Esti…
Early-Exit Graph Neural Networks
Andrea Giuseppe Di Francesco, Maria Sofia Bucarelli, Franco Maria Nardini +3
Early-exit mechanisms allow deep neural networks to stop inference once prediction confidence is high, reducing latency and energy on easy inputs while retaining full-depth accurac…
AMAQA: A Metadata-based QA Dataset for RAG Systems
Davide Bruni, Marco Avvenuti, Nicola Tonellotto +1
Retrieval-augmented generation (RAG) systems are widely used in question-answering (QA) tasks, but current benchmarks lack metadata integration, limiting their evaluation in scenar…
Exploring the Effectiveness of Multi-stage Fine-tuning for Cross-encoder Re-rankers
Francesca Pezzuti, Sean MacAvaney, Nicola Tonellotto
State-of-the-art cross-encoders can be fine-tuned to be highly effective in passage re-ranking. The typical fine-tuning process of cross-encoders as re-rankers requires large amoun…
ECLIPSE: Contrastive Dimension Importance Estimation with Pseudo-Irrelevance Feedback for Dense Retrieval
Giulio D'Erasmo, Giovanni Trappolini, Nicola Tonellotto +1
Recent advances in Information Retrieval have leveraged high-dimensional embedding spaces to improve the retrieval of relevant documents. Moreover, the Manifold Clustering Hypothes…
Static Pruning in Dense Retrieval using Matrix Decomposition
Federico Siciliano, Francesca Pezzuti, Nicola Tonellotto +1
In the era of dense retrieval, document indexing and retrieval is largely based on encoding models that transform text documents into embeddings. The efficiency of retrieval is dir…
A Theoretical Analysis of Recommendation Loss Functions under Negative Sampling
Giulia Di Teodoro, Federico Siciliano, Nicola Tonellotto +1
Loss functions like Categorical Cross Entropy (CCE), Binary Cross Entropy (BCE), and Bayesian Personalized Ranking (BPR) are commonly used in training Recommender Systems (RSs) to…
A Reproducible Analysis of Sequential Recommender Systems
Filippo Betello, Antonio Purificato, Federico Siciliano +4
Sequential Recommender Systems (SRSs) have emerged as a highly efficient approach to recommendation systems. By leveraging sequential data, SRSs can identify temporal patterns in u…
A Tale of Trust and Accuracy: Base vs. Instruct LLMs in RAG Systems
Florin Cuconasu, Giovanni Trappolini, Nicola Tonellotto +1
Retrieval Augmented Generation (RAG) represents a significant advancement in artificial intelligence combining a retrieval phase with a generative phase, with the latter typically…
Cross-Document Neural Re-Ranking via Query-Induced Subgraphs
Andrea Giuseppe Di Francesco, Christian Giannetti, Nicola Tonellotto +1
Neural re-rankers typically score query-document pairs independently, neglecting cross-document context within the retrieved candidate set. We propose Graph Neural Re-Ranking (GNRR…
Generating Query Recommendations via LLMs
Andrea Bacciu, Enrico Palumbo, Andreas Damianou +2
Query recommendation systems are ubiquitous in modern search engines, assisting users in producing effective queries to meet their information needs. However, these systems require…
The Power of Noise: Redefining Retrieval for RAG Systems
Florin Cuconasu, Giovanni Trappolini, Federico Siciliano +5
Retrieval-Augmented Generation (RAG) has recently emerged as a method to extend beyond the pre-trained knowledge of Large Language Models by augmenting the original prompt with rel…