Marco Aldinucci

Matteo Pennisi, Federica Proietto Salanitri, Giovanni Bellitto et al.

In the medical field, multi-center collaborations are often sought to yield more generalizable findings by leveraging the heterogeneity of patient and clinical data. However, recent privacy regulations hinder the possibility to share data, and consequently, to come up with machine learning-based solutions that support diagnosis and prognosis. Federated learning (FL) aims at sidestepping this limitation by bringing AI-based solutions to data owners and only sharing local AI models, or parts thereof, that need then to be aggregated. However, most of the existing federated learning solutions are still at their infancy and show several shortcomings, from the lack of a reliable and effective aggregation scheme able to retain the knowledge learned locally to weak privacy preservation as real data may be reconstructed from model updates. Furthermore, the majority of these approaches, especially those dealing with medical data, relies on a centralized distributed learning strategy that poses robustness, scalability and trust issues. In this paper we present a federated and decentralized learning strategy, FedER, that, exploiting experience replay and generative adversarial concepts, effectively integrates features from local nodes, providing models able to generalize across multiple datasets while maintaining privacy. FedER is tested on two tasks -- tuberculosis and melanoma classification -- using multiple datasets in order to simulate realistic non-i.i.d. medical data scenarios. Results show that our approach achieves performance comparable to standard (non-federated) learning and significantly outperforms state-of-the-art federated methods in their centralized (thus, more favourable) formulation. Code is available at https://github.com/perceivelab/FedER

Gianluca Mittone, Nicolò Tonci, Robert Birke et al.

Decentralised Machine Learning (DML) enables collaborative machine learning without centralised input data. Federated Learning (FL) and Edge Inference are examples of DML. While tools for DML (especially FL) are starting to flourish, many are not flexible and portable enough to experiment with novel processors (e.g., RISC-V), non-fully connected network topologies, and asynchronous collaboration schemes. We overcome these limitations via a domain-specific language allowing us to map DML schemes to an underlying middleware, i.e. the FastFlow parallel programming library. We experiment with it by generating different working DML schemes on x86-64 and ARM platforms and an emerging RISC-V one. We characterise the performance and energy efficiency of the presented schemes and systems. As a byproduct, we introduce a RISC-V porting of the PyTorch framework, the first publicly available to our knowledge.

Federica Proietto Salanitri, Giovanni Bellitto, Simone Palazzo et al.

Early detection of precancerous cysts or neoplasms, i.e., Intraductal Papillary Mucosal Neoplasms (IPMN), in pancreas is a challenging and complex task, and it may lead to a more favourable outcome. Once detected, grading IPMNs accurately is also necessary, since low-risk IPMNs can be under surveillance program, while high-risk IPMNs have to be surgically resected before they turn into cancer. Current standards (Fukuoka and others) for IPMN classification show significant intra- and inter-operator variability, beside being error-prone, making a proper diagnosis unreliable. The established progress in artificial intelligence, through the deep learning paradigm, may provide a key tool for an effective support to medical decision for pancreatic cancer. In this work, we follow this trend, by proposing a novel AI-based IPMN classifier that leverages the recent success of transformer networks in generalizing across a wide variety of tasks, including vision ones. We specifically show that our transformer-based model exploits pre-training better than standard convolutional neural networks, thus supporting the sought architectural universalism of transformers in vision, including the medical image domain and it allows for a better interpretation of the obtained results.

Bruno Casella, Roberto Esposito, Antonio Sciarappa et al.

Training Deep Learning (DL) models require large, high-quality datasets, often assembled with data from different institutions. Federated Learning (FL) has been emerging as a method for privacy-preserving pooling of datasets employing collaborative training from different institutions by iteratively globally aggregating locally trained models. One critical performance challenge of FL is operating on datasets not independently and identically distributed (non-IID) among the federation participants. Even though this fragility cannot be eliminated, it can be debunked by a suitable optimization of two hyper-parameters: layer normalization methods and collaboration frequency selection. In this work, we benchmark five different normalization layers for training Neural Networks (NNs), two families of non-IID data skew, and two datasets. Results show that Batch Normalization, widely employed for centralized DL, is not the best choice for FL, whereas Group and Layer Normalization consistently outperform Batch Normalization. Similarly, frequent model aggregation decreases convergence speed and mode quality.

Bruno Casella, Roberto Esposito, Carlo Cavazzoni et al.

Classic Machine Learning techniques require training on data available in a single data lake. However, aggregating data from different owners is not always convenient for different reasons, including security, privacy and secrecy. Data carry a value that might vanish when shared with others; the ability to avoid sharing the data enables industrial applications where security and privacy are of paramount importance, making it possible to train global models by implementing only local policies which can be run independently and even on air-gapped data centres. Federated Learning (FL) is a distributed machine learning approach which has emerged as an effective way to address privacy concerns by only sharing local AI models while keeping the data decentralized. Two critical challenges of Federated Learning are managing the heterogeneous systems in the same federated network and dealing with real data, which are often not independently and identically distributed (non-IID) among the clients. In this paper, we focus on the second problem, i.e., the problem of statistical heterogeneity of the data in the same federated network. In this setting, local models might be strayed far from the local optimum of the complete dataset, thus possibly hindering the convergence of the federated model. Several Federated Learning algorithms, such as FedAvg, FedProx and Federated Curvature (FedCurv), aiming at tackling the non-IID setting, have already been proposed. This work provides an empirical assessment of the behaviour of FedAvg and FedCurv in common non-IID scenarios. Results show that the number of epochs per round is an important hyper-parameter that, when tuned appropriately, can lead to significant performance gains while reducing the communication cost. As a side product of this work, we release the non-IID version of the datasets we used so to facilitate further comparisons from the FL community.

Gianluca Mittone, Filip Svoboda, Marco Aldinucci et al.

Aggregating pharmaceutical data in the drug-target interaction (DTI) domain has the potential to deliver life-saving breakthroughs. It is, however, notoriously difficult due to regulatory constraints and commercial interests. This work proposes the application of federated learning, which we argue to be reconcilable with the industry's constraints, as it does not require sharing of any information that would reveal the entities' data or any other high-level summary of it. When used on a representative GraphDTA model and the KIBA dataset it achieves up to 15% improved performance relative to the best available non-privacy preserving alternative. Our extensive battery of experiments shows that, unlike in other domains, the non-IID data distribution in the DTI datasets does not deteriorate FL performance. Additionally, we identify a material trade-off between the benefits of adding new data, and the cost of adding more clients.

Gianluca Mittone, Walter Riviera, Iacopo Colonnelli et al.

Since its debut in 2016, Federated Learning (FL) has been tied to the inner workings of Deep Neural Networks (DNNs). On the one hand, this allowed its development and widespread use as DNNs proliferated. On the other hand, it neglected all those scenarios in which using DNNs is not possible or advantageous. The fact that most current FL frameworks only allow training DNNs reinforces this problem. To address the lack of FL solutions for non-DNN-based use cases, we propose MAFL (Model-Agnostic Federated Learning). MAFL marries a model-agnostic FL algorithm, AdaBoost.F, with an open industry-grade FL framework: Intel OpenFL. MAFL is the first FL system not tied to any specific type of machine learning model, allowing exploration of FL scenarios beyond DNNs and trees. We test MAFL from multiple points of view, assessing its correctness, flexibility and scaling properties up to 64 nodes. We optimised the base software achieving a 5.5x speedup on a standard FL scenario. MAFL is compatible with x86-64, ARM-v8, Power and RISC-V.

Bruno Casella, Matthias Jakobs, Marco Aldinucci et al.

Time series classification (TSC) is a critical task with applications in various domains, including healthcare, finance, and industrial monitoring. Due to privacy concerns and data regulations, Federated Learning has emerged as a promising approach for learning from distributed time series data without centralizing raw information. However, most FL solutions rely on a client-server architecture, which introduces robustness and confidentiality risks related to the distinguished role of the server, which is a single point of failure and can observe knowledge extracted from clients. To address these challenges, we propose DROCKS, a fully decentralized FL framework for TSC that leverages ROCKET (RandOm Convolutional KErnel Transform) features. In DROCKS, the global model is trained by sequentially traversing a structured path across federation nodes, where each node refines the model and selects the most effective local kernels before passing them to the successor. Extensive experiments on the UCR archive demonstrate that DROCKS outperforms state-of-the-art client-server FL approaches while being more resilient to node failures and malicious attacks. Our code is available at https://anonymous.4open.science/r/DROCKS-7FF3/README.md.

Miruna Beţianu, Abele Mălan, Marco Aldinucci et al.

Large language models (LLMs) increasingly serve as the backbone for classifying text associated with distinct domains and simultaneously several labels (classes). When encountering domain shifts, e.g., classifier of movie reviews from IMDb to Rotten Tomatoes, adapting such an LLM-based multi-label classifier is challenging due to incomplete label sets at the target domain and daunting training overhead. The existing domain adaptation methods address either image multi-label classifiers or text binary classifiers. In this paper, we design DALLMi, Domain Adaptation Large Language Model interpolator, a first-of-its-kind semi-supervised domain adaptation method for text data models based on LLMs, specifically BERT. The core of DALLMi is the novel variation loss and MixUp regularization, which jointly leverage the limited positively labeled and large quantity of unlabeled text and, importantly, their interpolation from the BERT word embeddings. DALLMi also introduces a label-balanced sampling strategy to overcome the imbalance between labeled and unlabeled data. We evaluate DALLMi against the partial-supervised and unsupervised approach on three datasets under different scenarios of label availability for the target domain. Our results show that DALLMi achieves higher mAP than unsupervised and partially-supervised approaches by 19.9% and 52.2%, respectively.

Giulio Malenza, Francesco Targa, Adriano Marques Garcia et al.

In today's era of rapid technological advancement, artificial intelligence (AI) applications require large-scale, high-performance, and data-intensive computations, leading to significant energy demands. Addressing this challenge necessitates a combined approach involving both hardware and software innovations. Hardware manufacturers are developing new, efficient, and specialized solutions, with the RISC-V architecture emerging as a prominent player due to its open, extensible, and energy-efficient instruction set architecture (ISA). Simultaneously, software developers are creating new algorithms and frameworks, yet their energy efficiency often remains unclear. In this study, we conduct a comprehensive benchmark analysis of machine learning (ML) applications on the 64-core SOPHON SG2042 RISC-V architecture. We specifically analyze the energy consumption of deep learning inference models across three leading AI frameworks: PyTorch, ONNX Runtime, and TensorFlow. Our findings show that frameworks using the XNNPACK back-end, such as ONNX Runtime and TensorFlow, consume less energy compared to PyTorch, which is compiled with the native OpenBLAS back-end.

Iacopo Colonnelli, Barbara Cantalupo, Concetto Spampinato et al.

HPC is an enabling platform for AI. The introduction of AI workloads in the HPC applications basket has non-trivial consequences both on the way of designing AI applications and on the way of providing HPC computing. This is the leitmotif of the convergence between HPC and AI. The formalized definition of AI pipelines is one of the milestones of HPC-AI convergence. If well conducted, it allows, on the one hand, to obtain portable and scalable applications. On the other hand, it is crucial for the reproducibility of scientific pipelines. In this work, we advocate the StreamFlow Workflow Management System as a crucial ingredient to define a parametric pipeline, called "CLAIRE COVID-19 Universal Pipeline," which is able to explore the optimization space of methods to classify COVID-19 lung lesions from CT scans, compare them for accuracy, and therefore set a performance baseline. The universal pipeline automatizes the training of many different Deep Neural Networks (DNNs) and many different hyperparameters. It, therefore, requires a massive computing power, which is found in traditional HPC infrastructure thanks to the portability-by-design of pipelines designed with StreamFlow. Using the universal pipeline, we identified a DNN reaching over 90% accuracy in detecting COVID-19 lesions in CT scans.

Paolo Viviani, Maurizio Drocco, Marco Aldinucci

This work aims to assess the state of the art of data parallel deep neural network training, trying to identify potential research tracks to be exploited for performance improvement. Beside, it presents a design for a practical C++ library dedicated at implementing and unifying the current state of the art methodologies for parallel training in a performance-conscious framework, allowing the user to explore novel strategies without departing significantly from its usual work-flow.