arXiv subjects

Filippo Menolascina

Publications and source records attributed to Filippo Menolascina.

6 recordsLinked to original sources

arXiv2026

Learning the Koopman Operator using Attention Free Transformers

Learning Koopman operators with autoencoders enables linear prediction in a latent space, but long-horizon rollouts often drift off the learned manifold, leading to phase and amplitude errors on systems with switching, continuous spectra, or strong transients. We introduce two complementary components that make Koopman predictors more robust. First, we add an attention-free latent memory (AFT) block that aggregates a short window of past latents to produce a corrected latent before each Koopman update. Unlike multi-head attention, AFT operates in linear time and adds only $\approx$30k parameters ($3d^2 + T^2$, fewer than matched multi-head attention), yet captures the local temporal context needed to suppress error divergence. Second, we propose dynamic re-encoding: lightweight, online change-point triggers (EWMA, CUSUM, and sequential two-sample tests) that detect latent drift and project predictions back onto the autoencoder manifold. Across three benchmark systems -- Duffing oscillator, Repressilator, IRMA -- our model consistently reduces error accumulation compared to a Koopman autoencoder and matched-capacity multi-head attention. We also compare against GRU and Transformer autoencoders, evaluated both from initial conditions and with a 50-step context, and find that Koopman+AFT (with optional re-encoding) attains markedly lower long-horizon error while maintaining lower inference latency. We report improvements over horizons up to 1000 steps, together with ablations over trigger policies. The result is a fast, compact predictor that stays on the learned manifold over long horizons.

cs.LG↗

arXiv2025

ModernBERT + ColBERT: Enhancing biomedical RAG through an advanced re-ranking retriever

Retrieval-Augmented Generation (RAG) is a powerful technique for enriching Large Language Models (LLMs) with external knowledge, allowing for factually grounded responses, a critical requirement in high-stakes domains such as healthcare. However, the efficacy of RAG systems is fundamentally restricted by the performance of their retrieval module, since irrelevant or semantically misaligned documents directly compromise the accuracy of the final generated response. General-purpose dense retrievers can struggle with the nuanced language of specialised domains, while the high accuracy of in-domain models is often achieved at prohibitive computational costs. In this work, we aim to address this trade-off by developing and evaluating a two-stage retrieval architecture that combines a lightweight ModernBERT bidirectional encoder for efficient initial candidate retrieval with a ColBERTv2 late-interaction model for fine-grained re-ranking. We conduct comprehensive evaluations of our retriever module performance and RAG system performance in the biomedical context, fine-tuning the IR module using 10k question-passage pairs from PubMedQA. Our analysis of the retriever module confirmed the positive impact of the ColBERT re-ranker, which improved Recall@3 by up to 4.2 percentage points compared to its retrieve-only counterpart. When integrated into the biomedical RAG, our IR module leads to a state-of-the-art average accuracy of 0.4448 on the five tasks of the MIRAGE question-answering benchmark, outperforming strong baselines such as MedCPT (0.4436). Our ablation studies reveal that this performance is critically dependent on a joint fine-tuning process that aligns the retriever and re-ranker; otherwise, the re-ranker might degrade the performance.

cs.CL↗

arXiv2025

Bridging Clinical Narratives and ACR Appropriateness Guidelines: A Multi-Agent RAG System for Medical Imaging Decisions

The selection of appropriate medical imaging procedures is a critical and complex clinical decision, guided by extensive evidence-based standards such as the ACR Appropriateness Criteria (ACR-AC). However, the underutilization of these guidelines, stemming from the difficulty of mapping unstructured patient narratives to structured criteria, contributes to suboptimal patient outcomes and increased healthcare costs. To bridge this gap, we introduce a multi-agent cognitive architecture that automates the translation of free-text clinical scenarios into specific, guideline-adherent imaging recommendations. Our system leverages a novel, domain-adapted dense retrieval model, ColBERT, fine-tuned on a synthetically generated dataset of 8,840 clinical scenario-recommendation pairs to achieve highly accurate information retrieval from the ACR-AC knowledge base. This retriever identifies candidate guidelines with a 93.9% top-10 recall, which are then processed by a sequence of LLM-based agents for selection and evidence-based synthesis. We evaluate our architecture using GPT-4.1 and MedGemma agents, demonstrating a state-of-the-art exact match accuracy of 81%, meaning that in 81% of test cases the predicted procedure set was identical to the guideline's reference set, and an F1-score of 0.879. This represents a 67-percentage-point absolute improvement in accuracy over a strong standalone GPT-4.1 baseline, underscoring the contribution that our architecture makes to a frontier model. These results were obtained on a challenging test set with substantial lexical divergence from the source guidelines. Our code is available at https://anonymous.4open.science/r/demo-iclr-B567/

q-bio.QM↗

arXiv2025

Reinforcement Learning for Clinical Reasoning: Aligning LLMs with ACR Imaging Appropriateness Criteria

Medical imaging has revolutionized diagnosis, yet unnecessary procedures are rising, exposing patients to radiation and stress, limiting equitable access, and straining healthcare systems. The American College of Radiology Appropriateness Criteria, developed through extensive multidisciplinary review, provide evidence-based guidance but remain underutilized. Leveraging advances in LLM reasoning, we introduce a Reasoning Agent trained with Reinforcement Learning (RL), specifically Group Relative Policy Optimization (GRPO), to replicate expert clinical reasoning from the ACR Criteria. We present a novel RL approach for structured medical reasoning, systematically comparing reasoning-focused reward functions and evidence integration strategies. Our lightweight 8B model, MedReason-Embed, improves macro F1 by 18% over baseline, shows stronger reasoning alignment, and outperforms both larger and alternatively trained models, showing that reasoning-based supervision enables efficient, trustworthy clinical AI. Building on this, we design a modular end-to-end agentic architecture that automates imaging referrals: mapping diagnoses to ICD codes, retrieving PubMed evidence, and recommending optimal procedures. Crucially, the ability to generalize beyond static ACR guidelines not only enables clinicians to handle out-of-distribution cases, but also supports scaling the guideline development process itself, potentially reducing the significant effort required to create and update them. This work shows the potential of reasoning-focused RL within agentic architectures to deliver transparent, scalable, and reliable clinical decision support. Our code is available at: https://anonymous.4open.science/r/agentic-imaging-recommender-iclr-877D

q-bio.QM↗

arXiv2024

Medical Graph RAG: Towards Safe Medical Large Language Model via Graph Retrieval-Augmented Generation

We introduce a novel graph-based Retrieval-Augmented Generation (RAG) framework specifically designed for the medical domain, called \textbf{MedGraphRAG}, aimed at enhancing Large Language Model (LLM) capabilities for generating evidence-based medical responses, thereby improving safety and reliability when handling private medical data. Graph-based RAG (GraphRAG) leverages LLMs to organize RAG data into graphs, showing strong potential for gaining holistic insights from long-form documents. However, its standard implementation is overly complex for general use and lacks the ability to generate evidence-based responses, limiting its effectiveness in the medical field. To extend the capabilities of GraphRAG to the medical domain, we propose unique Triple Graph Construction and U-Retrieval techniques over it. In our graph construction, we create a triple-linked structure that connects user documents to credible medical sources and controlled vocabularies. In the retrieval process, we propose U-Retrieval which combines Top-down Precise Retrieval with Bottom-up Response Refinement to balance global context awareness with precise indexing. These effort enable both source information retrieval and comprehensive response generation. Our approach is validated on 9 medical Q\&A benchmarks, 2 health fact-checking benchmarks, and one collected dataset testing long-form generation. The results show that MedGraphRAG consistently outperforms state-of-the-art models across all benchmarks, while also ensuring that responses include credible source documentation and definitions. Our code is released at: https://github.com/MedicineToken/Medical-Graph-RAG.

cs.CV↗

arXiv2008

Developing a Theoretical Framework for Optofluidic Device Designing for System Identification in Systems Biology: the EGFR Study Case

Identification of dynamics underlying biochemical pathways of interest in oncology is a primary goal in current systems biology. Understanding structures and interactions that govern the evolution of such systems is believed to be a cornerstone in this research. Systems theory and systems identification theory are primary resources for this task since they both provide a self consistent framework for modelling and manipulating models of dynamical systems that are best suited for the problem under investigation. We address herein the issue of obtaining an informative dataset ZN to be used as starting point for identification of EGFR pathway dynamics. In order to match experimental identifiability criteria we propose a theoretical framework for input stimulus design based on dynamical properties of the system under investigation. A feasible optofluidic design has been designed on the basis of the spectral properties of the driving inputs that maximize information content after the theoretical studies.

q-bio.OT↗