Lorenzo Rosset
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Lorenzo Rosset has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Aurélien Decelle
Escuela Tecnica Superior de Ingenieros Industriales, Universidad Politecnica de Madrid
2 shared publications - Francesco Zamponi
Sapienza Università di Roma
2 shared publications - Martin Weigt
Sorbonne Universite
2 shared publications - Alessandra Carbone
Sorbonne University
1 shared publication - America Califano
University of Salerno
1 shared publication - Andrea Zanola
Department of Neuroscience, University of Padua, Padua, Italy
1 shared publication
Latest Publications
- Open Access
- 1 October 2026
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 7 (5) , 055050
From scarce functional labels to label-aware generation in homologous protein families
Accurately annotating and controlling protein function from sequence data remains a major challenge in protein engineering, especially when functional labels are scarce within large homologous families. Here, we study a two-stage light-supervision strategy for fine-grained functional annotation and label-aware sequence generation. First, we compare several sequence representations, including one-hot encodings, restricted Boltzmann machines (RBMs) and protein language model (pLM) embeddings, for predicting intra-family specificity labels from limited supervision. By using train/test splits that explicitly reduce phylogenetic leakage, we show that pLM-based representations do not systematically outperform family-specific RBM embeddings or even simple one-hot baselines in this regime. Second, we use the inferred annotations to train an annotation-aware RBM capable of generating artificial homologs conditioned on prescribed labels. For a specific protein family, we quantify how the number and quality of available labels determine the reliability of conditional generation. Our results show that scarce annotations can support label-aware protein design when they are accurately propagated, while also highlighting the importance of phylogeny-aware evaluation for assessing functional annotation methods within homologous families.
- 1 October 2026
- Published by IOP Publishing in Machine Learning: Science and Technology
- Vol. 7 (5) , 055050
Open AccessFrom scarce functional labels to label-aware generation in homologous protein families
Accurately annotating and controlling protein function from sequence data remains a major challenge in protein engineering, especially when functional labels are scarce within large homologous families. Here, we study a two-stage light-supervision strategy for fine-grained functional annotation and label-aware sequence generation. First, we compare several sequence representations, including one-hot encodings, restricted Boltzmann machines (RBMs) and protein language model (pLM) embeddings, for predicting intra-family specificity labels from limited supervision. By using train/test splits that explicitly reduce phylogenetic leakage, we show that pLM-based representations do not systematically outperform family-specific RBM embeddings or even simple one-hot baselines in this regime. Second, we use the inferred annotations to train an annotation-aware RBM capable of generating artificial homologs conditioned on prescribed labels. For a specific protein family, we quantify how the number and quality of available labels determine the reliability of conditional generation. Our results show that scarce annotations can support label-aware protein design when they are accurately propagated, while also highlighting the importance of phylogeny-aware evaluation for assessing functional annotation methods within homologous families.
- Open Access
From Scarce Functional Labels to Label-Aware Generation in Homologous Protein Families
Accurately annotating and controlling protein function from sequence data remains a major challenge in protein engineering, especially when functional labels are scarce within large homologous families. Here, we study a two-stage light-supervision strategy for fine-grained functional annotation and label-aware sequence generation. First, we compare several sequence representations, including one-hot encodings, Restricted Boltzmann Machines (RBMs), and ESM2-based protein language model embeddings, for predicting intra-family specificity labels from limited supervision. By using train/test splits that explicitly reduce phylogenetic leakage, we show that ESM2-based representations do not systematically outperform family-specific RBM embeddings or even simple one-hot baselines in this regime. Second, we use the inferred annotations to train an annotation-aware RBM capable of generating artificial homologs conditioned on prescribed labels. Across several protein families, we quantify how the number and quality of available labels determine the reliability of conditional generation. Our results show that scarce annotations can support label-aware protein design when they are accurately propagated, while also highlighting the importance of phylogeny-aware evaluation for assessing functional annotation methods within homologous families.
Open AccessFrom Scarce Functional Labels to Label-Aware Generation in Homologous Protein Families
Accurately annotating and controlling protein function from sequence data remains a major challenge in protein engineering, especially when functional labels are scarce within large homologous families. Here, we study a two-stage light-supervision strategy for fine-grained functional annotation and label-aware sequence generation. First, we compare several sequence representations, including one-hot encodings, Restricted Boltzmann Machines (RBMs), and ESM2-based protein language model embeddings, for predicting intra-family specificity labels from limited supervision. By using train/test splits that explicitly reduce phylogenetic leakage, we show that ESM2-based representations do not systematically outperform family-specific RBM embeddings or even simple one-hot baselines in this regime. Second, we use the inferred annotations to train an annotation-aware RBM capable of generating artificial homologs conditioned on prescribed labels. Across several protein families, we quantify how the number and quality of available labels determine the reliability of conditional generation. Our results show that scarce annotations can support label-aware protein design when they are accurately propagated, while also highlighting the importance of phylogeny-aware evaluation for assessing functional annotation methods within homologous families.
- 11 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (2) , 1309-1316
Fast and Functional Structured Data Generators Rooted in Out-of-Equilibrium Physics
In this study, we address the challenge of using energy-based models to produce high-quality, label-specific data in complex structured datasets, such as population genetics, RNA or protein sequences data. Traditional training methods encounter difficulties due to inefficient Markov chain Monte Carlo mixing, which affects the diversity of synthetic data and increases generation times. To address these issues, we use a novel training algorithm that exploits non-equilibrium effects. This approach, applied to the Restricted Boltzmann Machine, improves the model's ability to correctly classify samples and generate high-quality synthetic data in only a few sampling steps. The effectiveness of this method is demonstrated by its successful application to five different types of data: handwritten digits, mutations of human genomes classified by continental origin, functionally characterized sequences of an enzyme protein family, homologous RNA sequences from specific taxonomies and real classical piano pieces classified by their composer.
- 11 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (2) , 1309-1316
Fast and Functional Structured Data Generators Rooted in Out-of-Equilibrium Physics
In this study, we address the challenge of using energy-based models to produce high-quality, label-specific data in complex structured datasets, such as population genetics, RNA or protein sequences data. Traditional training methods encounter difficulties due to inefficient Markov chain Monte Carlo mixing, which affects the diversity of synthetic data and increases generation times. To address these issues, we use a novel training algorithm that exploits non-equilibrium effects. This approach, applied to the Restricted Boltzmann Machine, improves the model's ability to correctly classify samples and generate high-quality synthetic data in only a few sampling steps. The effectiveness of this method is demonstrated by its successful application to five different types of data: handwritten digits, mutations of human genomes classified by continental origin, functionally characterized sequences of an enzyme protein family, homologous RNA sequences from specific taxonomies and real classical piano pieces classified by their composer.
- 7 July 2023
- Published by American Physical Society (APS) in Physical Review E
- Vol. 108 (1) , 014110
Unsupervised hierarchical clustering using the learning dynamics of restricted Boltzmann machines
Data sets in the real world are often complex and to some degree hierarchical, with groups and subgroups of data sharing common characteristics at different levels of abstraction. Understanding and uncovering the hidden structure of these data sets is an important task that has many practical applications. To address this challenge, we present a general method for building relational data trees by exploiting the learning dynamics of the restricted Boltzmann machine. Our method is based on the mean-field approach, derived from the Plefka expansion, and developed in the context of disordered systems. It is designed to be easily interpretable. We tested our method in an artificially created hierarchical data set and on three different real-world data sets (images of digits, mutations in the human genome, and a homologous family of proteins). The method is able to automatically identify the hierarchical structure of the data. This could be useful in the study of homologous protein sequences, where the relationships between proteins are critical for understanding their function and evolution.
- 7 July 2023
- Published by American Physical Society (APS) in Physical Review E
- Vol. 108 (1) , 014110
Unsupervised hierarchical clustering using the learning dynamics of restricted Boltzmann machines
Data sets in the real world are often complex and to some degree hierarchical, with groups and subgroups of data sharing common characteristics at different levels of abstraction. Understanding and uncovering the hidden structure of these data sets is an important task that has many practical applications. To address this challenge, we present a general method for building relational data trees by exploiting the learning dynamics of the restricted Boltzmann machine. Our method is based on the mean-field approach, derived from the Plefka expansion, and developed in the context of disordered systems. It is designed to be easily interpretable. We tested our method in an artificially created hierarchical data set and on three different real-world data sets (images of digits, mutations in the human genome, and a homologous family of proteins). The method is able to automatically identify the hierarchical structure of the data. This could be useful in the study of homologous protein sequences, where the relationships between proteins are critical for understanding their function and evolution.
- 20 May 2022
- Published by Emerald Publishing in International Journal of Building Pathology and Adaptation
- Vol. 42 (1) , 18-34
Natural climate reconstruction in the Norwegian stave churches through time series processing with variational autoencoders
In the field of heritage science, especially applied to buildings and artefacts made by organic hygroscopic materials, analyzing the microclimate has always been of extreme importance. In particular, in many cases, the knowledge of the outdoor/indoor microclimate may support the decision process in conservation and preservation matters of historic buildings. This knowledge is often gained by implementing long and time-consuming monitoring campaigns that allow collecting atmospheric and climatic data. Sometimes the collected time series may be corrupted, incomplete and/or subjected to the sensors' errors because of the remoteness of the historic building location, the natural aging of the sensor or the lack of a continuous check of the data downloading process. For this reason, in this work, an innovative approach about reconstructing the indoor microclimate into heritage buildings, just knowing the outdoor one, is proposed. This methodology is based on using machine learning tools known as variational auto encoders (VAEs), that are able to reconstruct time series and/or to fill data gaps. The proposed approach is implemented using data collected in Ringebu Stave Church, a Norwegian medieval wooden heritage building. Reconstructing a realistic time series, for the vast majority of the year period, of the natural internal climate of the Church has been successfully implemented. The novelty of this work is discussed in the framework of the existing literature. The work explores the potentials of machine learning tools compared to traditional ones, providing a method that is able to reliably fill missing data in time series.
- 20 May 2022
- Published by Emerald Publishing in International Journal of Building Pathology and Adaptation
- Vol. 42 (1) , 18-34
Natural climate reconstruction in the Norwegian stave churches through time series processing with variational autoencoders
In the field of heritage science, especially applied to buildings and artefacts made by organic hygroscopic materials, analyzing the microclimate has always been of extreme importance. In particular, in many cases, the knowledge of the outdoor/indoor microclimate may support the decision process in conservation and preservation matters of historic buildings. This knowledge is often gained by implementing long and time-consuming monitoring campaigns that allow collecting atmospheric and climatic data. Sometimes the collected time series may be corrupted, incomplete and/or subjected to the sensors' errors because of the remoteness of the historic building location, the natural aging of the sensor or the lack of a continuous check of the data downloading process. For this reason, in this work, an innovative approach about reconstructing the indoor microclimate into heritage buildings, just knowing the outdoor one, is proposed. This methodology is based on using machine learning tools known as variational auto encoders (VAEs), that are able to reconstruct time series and/or to fill data gaps. The proposed approach is implemented using data collected in Ringebu Stave Church, a Norwegian medieval wooden heritage building. Reconstructing a realistic time series, for the vast majority of the year period, of the natural internal climate of the Church has been successfully implemented. The novelty of this work is discussed in the framework of the existing literature. The work explores the potentials of machine learning tools compared to traditional ones, providing a method that is able to reliably fill missing data in time series.