About
I am Full Professor at the department of Computer Science and Technology of the University of Cambridge and I am a member of the Artificial Intelligence group. I am a member of the Cambridge Centre for AI in Medicine.
Background: PhD in Complex Systems and Non Linear Dynamics (School of Informatics, dept of Engineering of the University of Firenze, Italy) and PhD in (Theoretical) Genetics (University of Pavia, Italy). More information is on my personal homepage
Other Affliations: Fellow and member of the Council of Clare Hall College, member of Ellis, the European Lab for Learning & Intelligent Systems, I am member of the Academia Europaea; I am listed in www.topitalianscientists.org/Top_italian_scientists_VIA-Academy.aspx
I am happy to receive enquiries for PhD applications. I have successfully completed the equality and diversity essentials.
My research interest focuses on developing Artificial Intelligence and Computational Biology models to understand diseases complexity and address personalised and precision medicine. Current focus is on Graph Neural Network modeling to build predictive models based on the integration of multi scale, multi omics and multi physics data; integrate deep learning and mechanistic approaches; explainability and interpretability in medicine; exploiting short and long range communications in the human body, between cells and tissues and predict emerging mechanistic properties at systemic medicine level. Develop an AI-based medical digital twin to increase self-awareness; Develop an AI personal decision support system to increase social awareness.
Skills & Expertise
Machine Learning
Genomics
Neural Networks
Dynamics
Bioinformatics and biostatics
Computational Intelligence
Research Interests
Artificial Intelligence
Machine Learning
Deep Learning
Medical Imaging Technology
nonlinear dynamics
Computer Science
Biology
Computer Vision
Bioinformatics
Medicine
deep learning
Complex Systems
AI methodology
Geometric deep learning
Graph Neural Networks
Hypergraphs
Neural ODE
PDE
Latent force models
Neural execution of algorithms
multimodality
Neuroimaging
Precision Medicine
Computational Biology
Personalised Medicine
Disease Complexity Modeling
Explainable AI
Mechanistic Modeling
Systems Medicine
Multi-Omics Data Integration
Medical Digital Twin
Decision Support Systems
Human Body Communication
Theoretical Genetics
Nonlinear Dynamics
AI in Medicine
Predictive Modeling
Digital Health
Health Informatics
Stroke MRI Analysis
Real-Time Cancer Imaging
Feature Visualization
Clinical Decision Support Systems
Chronic Disease Prediction
Translational Data Science
Data Science Algorithms
EEG and ECG Analysis
Secondary Data Utilization
Automated Disease Diagnosis
Interpretable Machine Learning
Connect With Me
Experience
Professor
Visiting Professor
Research Associate
Research Associate in Statistical
Epidemiology, Princess Anne Hospital, University of Southampton
Lecturer
Education
University of Florence (UF)
University of Pavia (UP)
Projects
MICA: Mental Health Data Pathfinder
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details. Technical Summary With strong NHS partnerships and recent contributions to national mental health (MH) informatics, we shall add novel methods, epidemiology and phenotyping to the MH Platform. We envisage a modular pipeline that de-identifies MH data; supports flexible consent for sharing/contact; and links MH, cognitive, physical, psychosocial and biomarker data. Project (P) 1. Our open-source tools de-identify clinical records to create CPFT’s Research Database, supporting research and participation. We shall extend them to generate anonymised subsets and link data from consenting patients across MH/community services, acute care, and research organizations, including from existing deeply phenotyped longitudinal cohorts. We emphasize rigorous interface standards and NHS governance over identifiable/pseudonymised data. We shall collaborate on a national natural language processing framework, allowing NHS/research organizations to generate structured data from free text. P2. We have created novel open-source neuropsychiatric assessment software. We shall extend it for broad and integrated NHS and research use. This will take automated cognitive testing into routine clinical practice. As a bold but tractable exemplar with research and clinical applications, we shall use it to apply electronic diagnostic algorithms and neuropsychiatric phenotyping, and link these detailed data to clinical records and biomarkers that include immunophenotyping. P3. We shall apply P1 tools to a public health crisis: the premature death of those with serious mental illness. We shall link MH, national and acute Trust data and use machine learning to develop early predictors of mortality. P4. We shall democratize MH research though broad consultation on generic tiered consent models for data-sharing and participation, by giving the research database direct clinical interfaces, and by enhancing data visualization to help clinicians and service users develop research and the NHS improve local services.
Chemometric Histopathology via Coherent Raman Imaging for Precision Medicine
Ultrafast Holographic FTIR Microscopy
Trophy (G112412 NRAG/752), Horizon Europe UKRI Underwrite Innovate, Title:
Ultrafast Holographic FTIR Microscopy (2022-2026)
Conferences & Seminars (2)
Denoising Probabilistic Diffusion Models for Synthetic Healthcare Image Generation
Healthcare data are an essential resource in Machine Learning (ML) and Artificial Intelligence (AI) to improve clinical practice, empower patients and enhance drug development with the aim to discover new medical knowledge. In particular, the biomedical imaging analysis plays a important role in the health-care context producing a huge amount of data that can be used to study complex diseases and their evolution in a deeper way or to predict their onsets. In this work we consider an approach based on Denoising Diffusion Probabilistic Models (DDPM) which is a type of generative model that uses a parameterized Markov chain and variational inference to generate synthetic samples that match real data. In particular, we execute a study by training on Malaria images and generating high-quality synthetic samples in order (i) to test the performance of the DDPMs, (ii) to estimate the association between original and synthetic data and (iii) to understand how the natural and human-made environmental factors impact Malaria disease. Finally, we use a well-defined convolutional neural network for classification tasks to assess the DDPM's goodness in generating the synthetic images.
7th International Conference on Computational and Mathematical Biomedical Engineering
EVENTS: 20/10/20 talk at TEDxCambridgeUniversity
23/11/20 talk on AI and Medicine at FuturoRemoto, Naples; 28-30 June 2021: Plenary Lecture at CMBE21 (7th International Conference on Computational and Mathematical Biomedical
Engineering); Politecnico di Milano.
Certificates & Licenses (1)
BITS Bioinformatics Italian Society
Awards & Achievements (2)
🏆 Best student paper
🏆 Listed in the top Italian Scientists
Thesis Guided (1)
Deep concept reasoning: beyond the accuracy-interpretability trade-off
Institution: University of Cambridge
Professional Memberships (2)
University of Cambridge
Country: United Kingdom
University of Cambridge
Country: United Kingdom
Invited Position (3)
DEGAS at GSP Workshop 2023: AI and Medicine: Graph and Hypergraph Representation Learning
OxML 2023
The 8th Cambridge-UTokyo Joint Symposium: Parallel Session 3 AI Trends, Opportunities and Threats
Publications (220)
Android malware seriously affects the use of Android applications, and a growing number of Android malware developers are using adversarial attacks to evade detection by deep learning models. This wor...
A heterogeneous information network (heterogeneous graph) federated learning plays a crucial role in enabling multi-party collaboration in the IoT system. However, due to differences in business and d...
Neural Algorithmic Reasoning (NAR) aims to optimize classical algorithms. However, canonical implementations of NAR train neural networks to return only a single solution, even when there are multiple...
Heterogeneous graphs, with nodes and edges of different types, are commonly used to model relational structures in many real-world applications. Standard Graph Neural Networks (GNNs) struggle to proce...
Graph neural networks (GNNs) have gained significant attention for their ability to learn representations from graph-structured data, in which message passing and feature fusion strategies play an ess...
dd