arXiv subjects

S. Palazzo

Publications and source records attributed to S. Palazzo.

6 recordsLinked to original sources

arXiv2025

radio-llava: Advancing Vision-Language Models for Radio Astronomical Source Analysis

The advent of next-generation radio telescopes is set to transform radio astronomy by producing massive data volumes that challenge traditional processing methods. Deep learning techniques have shown strong potential in automating radio analysis tasks, yet are often constrained by the limited availability of large annotated datasets. Recent progress in self-supervised learning has led to foundational radio vision models, but adapting them for new tasks typically requires coding expertise, limiting their accessibility to a broader astronomical community. Text-based AI interfaces offer a promising alternative by enabling task-specific queries and example-driven learning. In this context, Large Language Models (LLMs), with their remarkable zero-shot capabilities, are increasingly used in scientific domains. However, deploying large-scale models remains resource-intensive, and there is a growing demand for AI systems that can reason over both visual and textual data in astronomical analysis. This study explores small-scale Vision-Language Models (VLMs) as AI assistants for radio astronomy, combining LLM capabilities with vision transformers. We fine-tuned the LLaVA VLM on a dataset of 59k radio images from multiple surveys, enriched with 38k image-caption pairs from the literature. The fine-tuned models show clear improvements over base models in radio-specific tasks, achieving ~30% F1-score gains in extended source detection, but they underperform vision-only classifiers and exhibit ~20% drop on general multimodal tasks. Inclusion of caption data and LoRA fine-tuning enhances instruction-following and helps recover ~10% accuracy on multimodal benchmarks. This work lays the foundation for future advancements in radio VLMs, highlighting their potential and limitations, such as the need for better multimodal alignment, higher-quality datasets, and mitigation of catastrophic forgetting.

astro-ph.IM↗

arXiv2024

Self-supervised contrastive learning of radio data for source detection, classification and peculiar object discovery

New advancements in radio data post-processing are underway within the SKA precursor community, aiming to facilitate the extraction of scientific results from survey images through a semi-automated approach. Several of these developments leverage deep learning (DL) methodologies for diverse tasks, including source detection, object or morphology classification, and anomaly detection. Despite substantial progress, the full potential of these methods often remains untapped due to challenges associated with training large supervised models, particularly in the presence of small and class-unbalanced labelled datasets. Self-supervised learning has recently established itself as a powerful methodology to deal with some of the aforementioned challenges, by directly learning a lower-dimensional representation from large samples of unlabelled data. The resulting model and data representation can then be used for data inspection and various downstream tasks if a small subset of labelled data is available. In this work, we explored contrastive learning methods to learn suitable radio data representation from unlabelled images taken from the ASKAP EMU and SARAO MeerKAT GPS surveys. We evaluated trained models and the obtained data representation over smaller labelled datasets, also taken from different radio surveys, in selected analysis tasks: source detection and classification, and search for objects with peculiar morphology. For all explored downstream tasks, we reported and discussed the benefits brought by self-supervised foundational models built on radio data.

astro-ph.IM↗

arXiv2023

BioTrak: A Blockchain-based Platform for Food Chain Logistics Traceability

The food supply chain, following its globalization, has become very complex. Such complexities, introduce factors that influence adversely the quality of intermediate and final products. Strict constraints regarding parameters such as maintenance temperatures and transportation times must be respected in order to ensure top quality and reduce to a minimum the detrimental effects to public health. This is a multi-factorial endeavor and all of the involved stakeholders must accept and manage the logistics burden to achieve the best possible results. However, such burden comes together with additional complexities and costs regarding data storage, business process management and company specific standard operating procedures and as such, automated methods must be devised to reduce the impact of such intrusive operations. For the above reasons, in this paper we present BioTrak: a platform capable of registering and visualizing the whole chain of transformation and transportation processes including the monitoring of cold chain logistics of food ingredients starting from the raw material producers until the final product arrives to the end-consumer. The platform includes Business Process Modelling methods to aid food supply chain stakeholders to optimize their processes and also integrates a blockchain for guaranteeing the integrity, transparency and accountability of the data.

cs.CY↗

arXiv2019

Adversarial Framework for Unsupervised Learning of Motion Dynamics in Videos

Human behavior understanding in videos is a complex, still unsolved problem and requires to accurately model motion at both the local (pixel-wise dense prediction) and global (aggregation of motion cues) levels. Current approaches based on supervised learning require large amounts of annotated data, whose scarce availability is one of the main limiting factors to the development of general solutions. Unsupervised learning can instead leverage the vast amount of videos available on the web and it is a promising solution for overcoming the existing limitations. In this paper, we propose an adversarial GAN-based framework that learns video representations and dynamics through a self-supervision mechanism in order to perform dense and global prediction in videos. Our approach synthesizes videos by 1) factorizing the process into the generation of static visual content and motion, 2) learning a suitable representation of a motion latent space in order to enforce spatio-temporal coherency of object trajectories, and 3) incorporating motion estimation and pixel-wise dense prediction into the training procedure. Self-supervision is enforced by using motion masks produced by the generator, as a co-product of its generation process, to supervise the discriminator network in performing dense prediction. Performance evaluation, carried out on standard benchmarks, shows that our approach is able to learn, in an unsupervised way, both local and global video dynamics. The learned representations, then, support the training of video object segmentation methods with sensibly less (about 50%) annotations, giving performance comparable to the state of the art. Furthermore, the proposed method achieves promising performance in generating realistic videos, outperforming state-of-the-art approaches especially on motion-related metrics.

cs.CV↗

arXiv2018

Construction techniques and performances of a full-size prototype Micromegas chamber for the ATLAS muon spectrometer upgrade

A full-size prototype of a Micromegas precision tracking chamber for the upgrade of the ATLAS detector at the LHC Collider has been built between October 2015 and April 2016. This paper describes in detail the procedures used in constructing the single modules of the chamber in various INFN laboratories and the final assembly at the Laboratori Nazionali di Frascati (LNF). Results of the chamber exposure to the CERN SPS/H8 beam line in June 2016 are also presented. The performances achieved in the construction and the results of the test beam are compared with the requirements, which are imposed by the severe environment during the data-taking of the LHC foreseen for the next years.

physics.ins-det↗

arXiv2018

A Saliency-based Convolutional Neural Network for Table and Chart Detection in Digitized Documents

Deep Convolutional Neural Networks (DCNNs) have recently been applied successfully to a variety of vision and multimedia tasks, thus driving development of novel solutions in several application domains. Document analysis is a particularly promising area for DCNNs: indeed, the number of available digital documents has reached unprecedented levels, and humans are no longer able to discover and retrieve all the information contained in these documents without the help of automation. Under this scenario, DCNNs offers a viable solution to automate the information extraction process from digital documents. Within the realm of information extraction from documents, detection of tables and charts is particularly needed as they contain a visual summary of the most valuable information contained in a document. For a complete automation of visual information extraction process from tables and charts, it is necessary to develop techniques that localize them and identify precisely their boundaries. In this paper we aim at solving the table/chart detection task through an approach that combines deep convolutional neural networks, graphical models and saliency concepts. In particular, we propose a saliency-based fully-convolutional neural network performing multi-scale reasoning on visual cues followed by a fully-connected conditional random field (CRF) for localizing tables and charts in digital/digitized documents. Performance analysis carried out on an extended version of ICDAR 2013 (with annotated charts as well as tables) shows that our approach yields promising results, outperforming existing models.

cs.CV↗