Nicola Tonellotto

Large language models exhibit a critical vulnerability to distractor interference in retrieval-augmented contexts: they fail to prioritize relevant, factually correct documents over topically similar but misleading content. We introduce Lat-Defuse, a mechanistic framework that corrects this failu...

Dense retrieval uses embedded dense representations for documents to be obtained from language models ingested in vector stores. To reduce large space requirements and/or efficiency concerns, approximate nearest neighbours algorithms can be deployed, typically at the cost of effectiveness. Among...

Modern sequential recommendation systems often rely on negative sampling to efficiently train models over vast item corpora. However, common strategies such as uniform, popularity-based sampling, or hard-negative mining, often yield uninformative negatives, introduce popularity bias, or suffer fr...

We introduce a load-aware selective embedding-pruning method to be used in dense retrieval systems. Our method adapts the embedding dimensionality at inference time based on per-query deadlines and the current queue occupancy. Under light load, full-dimensional embeddings maximise effectiveness;...