Research
An animated map of my research — every paper is a node, attached to one or more topics. Drag to rearrange, hover for details, click a paper to open it, click a topic to isolate its cluster.
Topics are assigned via OpenAlex concepts and a keyword override list in _data/topics.yml. Papers can belong to multiple topics. Auto-rebuilt daily.
Browse as list (99 papers across 8 topics)
Compression & Pruning (31)
- Compression in 3D Gaussian Splatting: A Survey of Methods, Trends, and Future Directions
- GoDe: Gaussians on Demand for Progressive Level of Detail and Scalable Compression
- Lightweight Embedded FPGA Deployment of Learned Image Compression with Knowledge Distillation and Hybrid Quantization
- Security and real-time fpga integration for learned image compression
- The silence of the weights: an investigation of structural pruning strategies for attention-based audio signal architectures
- NEPENTHE: Entropy-Based Pruning as a Neural Network Depth's Reducer
- INSTANT: Compressing Gradients and Activations for Resource-Efficient Training
- Beyond Low-rank Decomposition: A Shortcut Approach for Efficient On-Device Learning
- Denoising Diffusion Probabilistic Model for Point Cloud Compression at Low Bit-Rates
- Efficient Progressive Image Compression with Variance-aware Masking
- ELMGS: Enhancing memory and computation scaLability through coMpression for 3D Gaussian Splatting
- STanH: Parametric Quantization for Variable Rate Learned Image Compression
- TEP-ones: A simple yet effective approach for transferability estimation of pruned backbones
- Till the Layers Collapse: Compressing a Deep Neural Network Through the Lenses of Batch Normalization Layers
- Activation Map Compression through Tensor Decomposition for Deep Learning
- ALICE: Adapt your Learnable Image Compression modEl for variable bitrates
- Domain Adaptation for Learned Image Compression with Supervised Adapters
- DSD$^2$: Can We Dodge Sparse Double Descent and Compress the Neural Network Worry-Free?
- Find the Lady: Permutation and Re-Synchronization of Deep Neural Networks
- Gabic: Graph-Based Attention Block for Image Compression
- Trimming the Fat: Efficient Compression of 3D Gaussian Splats through Pruning
- A Differentiable Entropy Model for Learned Image Compression
- Can Unstructured Pruning Reduce the Depth in Deep Neural Networks?
- Compressing Explicit Voxel Grid Representations: Fast NeRFs Become Also Small
- Learn how to Prune Pixels for Multi-view Neural Image-based Synthesis
- Shannon strikes again! entropy-based pruning in deep neural networks for transfer learning under extreme memory and computation budgets
- Unified Measures for the Rate-Distortion-Latency Trade-off
- The Rise of the Lottery Heroes: Why Zero-Shot Pruning is Hard
- HEMP: High-order entropy minimization for neural network compression
- On the role of structured pruning for neural network compression
- Pruning artificial neural networks: A way to find well-generalizing, high-entropy sharp minima
3D & Gaussian Splatting (11)
- Compression in 3D Gaussian Splatting: A Survey of Methods, Trends, and Future Directions
- RAVE: Rate-Adaptive Visual Encoding for 3D Gaussian Splatting
- Denoising Diffusion Probabilistic Model for Point Cloud Compression at Low Bit-Rates
- ELMGS: Enhancing memory and computation scaLability through coMpression for 3D Gaussian Splatting
- LayerFold: A Python library to reduce the depth of neural networks
- Boost Your NeRF: A Model-Agnostic Mixture of Experts Framework for High Quality and Efficient Rendering
- Trimming the Fat: Efficient Compression of 3D Gaussian Splats through Pruning
- Compressing Explicit Voxel Grid Representations: Fast NeRFs Become Also Small
- Learn how to Prune Pixels for Multi-view Neural Image-based Synthesis
- Two is Better than One: Achieving High-Quality 3D Scene Modeling with a NeRF Ensemble
- A Two-Step Radiologist-Like Approach for Covid-19 Computer-Aided Diagnosis from Chest X-Ray Images
Debiasing & Fairness (24)
- How I Met Your Bias: Investigating Bias Amplification in Diffusion Models
- I-INR: Iterative Implicit Neural Representations
- Study of Training Dynamics for Memory-Constrained Fine-Tuning
- Diffusing DeBias: Synthetic Bias Amplification for Model Debiasing
- Improving Generative Data Augmentation with Prior-Knowledge for Dysplasia Grading of Colorectal Polyps
- Say My Name: a Model's Bias Discovery Framework
- WaterMAS: Sharpness-Aware Maximization for Neural Network Watermarking
- Boost Your NeRF: A Model-Agnostic Mixture of Experts Framework for High Quality and Efficient Rendering
- Debiasing surgeon: fantastic weights and how to find them
- DSD$^2$: Can We Dodge Sparse Double Descent and Compress the Neural Network Worry-Free?
- Find the Lady: Permutation and Re-Synchronization of Deep Neural Networks
- A hitchhiker's guide to white-box neural network watermarking robustness
- Mining bias-target Alignment from Voronoi Cells
- Packed Ensembles for efficient uncertainty estimation
- Information Removal at the bottleneck in Deep Neural Networks
- To update or not to update? Neurons at equilibrium in deep models
- Bridging the gap between debiasing and privacy for deep learning
- Delving in the loss landscape to embed robust watermarks into neural networks
- End: Entangling and disentangling deep representations for bias correction
- HEMP: High-order entropy minimization for neural network compression
- On the role of structured pruning for neural network compression
- Serene: Sensitivity-based regularization of neurons for structured sparsity in neural networks
- A non-discriminatory approach to ethical deep learning
- Take a ramble into solution spaces for classification problems in neural networks
Regularization (17)
- Till the Layers Collapse: Compressing a Deep Neural Network Through the Lenses of Batch Normalization Layers
- Unified Pipeline for Generalized Mental State Detection Using EEG Signals
- DSD$^2$: Can We Dodge Sparse Double Descent and Compress the Neural Network Worry-Free?
- Can we avoid Double Descent in Deep Neural Networks?
- Disentangling private classes through regularization
- Mining bias-target Alignment from Voronoi Cells
- Shannon strikes again! entropy-based pruning in deep neural networks for transfer learning under extreme memory and computation budgets
- Sparse Double Descent in Vision Transformers: real or phantom threat?
- Loss-based sensitivity regularization: towards deep sparse neural networks
- End: Entangling and disentangling deep representations for bias correction
- Serene: Sensitivity-based regularization of neurons for structured sparsity in neural networks
- A non-discriminatory approach to ethical deep learning
- Pruning artificial neural networks: A way to find well-generalizing, high-entropy sharp minima
- Unveiling covid-19 from chest x-ray with deep learning: a hurdles race with small data
- Post-synaptic potential regularization has potential
- Learning sparse neural networks via sensitivity-driven regularization
- Role of synaptic stochasticity in training low-precision neural networks
Privacy (5)
- Disentangling private classes through regularization
- SCoTTi: Save Computation at Training Time with an adaptive framework
- Information Removal at the bottleneck in Deep Neural Networks
- Bridging the gap between debiasing and privacy for deep learning
- A non-discriminatory approach to ethical deep learning
Efficient & Frugal AI (19)
- Lightweight Embedded FPGA Deployment of Learned Image Compression with Knowledge Distillation and Hybrid Quantization
- Efficient Resource-Constrained Training of Vision Transformers via Subspace Optimization
- INSTANT: Compressing Gradients and Activations for Resource-Efficient Training
- Beyond Low-rank Decomposition: A Shortcut Approach for Efficient On-Device Learning
- Efficient adaptation of deep neural networks for semantic segmentation in space applications
- Efficient Progressive Image Compression with Variance-aware Masking
- Improving Generative Data Augmentation with Prior-Knowledge for Dysplasia Grading of Colorectal Polyps
- Boost Your NeRF: A Model-Agnostic Mixture of Experts Framework for High Quality and Efficient Rendering
- The Simpler The Better: An Entropy-Based Importance Metric To Reduce Neural Networks' Depth
- Towards On-device Learning on the Edge: Ways to Select Neurons to Update under a Budget Constraint
- Trimming the Fat: Efficient Compression of 3D Gaussian Splats through Pruning
- All predictions matter: an online video prediction approach
- Compressing Explicit Voxel Grid Representations: Fast NeRFs Become Also Small
- Mental State Classification Using EEG Signals: Ethics, Law and Challenges
- Mental state classification using EEG signals: ethics, law and challenges
- Packed Ensembles for efficient uncertainty estimation
- SCoTTi: Save Computation at Training Time with an adaptive framework
- Unified Measures for the Rate-Distortion-Latency Trade-off
- On the role of structured pruning for neural network compression
Continual Learning (17)
- Study of Training Dynamics for Memory-Constrained Fine-Tuning
- Ask and Remember: A Questions-Only Replay Strategy for Continual Visual Question Answering
- Efficient adaptation of deep neural networks for semantic segmentation in space applications
- Enhancing Plasticity for First Session Adaptation Continual Learning
- Neural Velocity for hyperparameter tuning
- Domain Adaptation for Learned Image Compression with Supervised Adapters
- Mini but Mighty: Finetuning ViTs with Mini Adapters
- Unsupervised Learning of Unbiased Visual Representations
- Weighted Ensemble Models Are Strong Continual Learners
- Mining bias-target Alignment from Voronoi Cells
- Shannon strikes again! entropy-based pruning in deep neural networks for transfer learning under extreme memory and computation budgets
- Capsule Networks with Routing Annealing
- End: Entangling and disentangling deep representations for bias correction
- A non-discriminatory approach to ethical deep learning
- Pruning artificial neural networks: A way to find well-generalizing, high-entropy sharp minima
- Unveiling covid-19 from chest x-ray with deep learning: a hurdles race with small data
- Post-synaptic potential regularization has potential
Other (20)
- Feature-aware Hypergraph Generation via Next-Scale Prediction
- Memory Constrained Dynamic Subnetwork Update for Transfer Learning
- Capsule Networks Do Not Need to Model Everything
- FOLDER: Accelerating Multi-modal Large Language Models with Enhanced Performance
- HYGENE: A Diffusion-based Hypergraph Generation Method
- LaCoOT: Layer Collapse through Optimal Transport
- WiGNet: Windowed Vision Graph Neural Network
- Memory-Optimized Once-For-All Network
- AIoT-based Neural Decoding and Neurofeedback for Accelerated Cognitive Training: Vision, Directions and Preliminary Results
- Neural network-derived perfusion maps: A model-free approach to computed tomography perfusion in patients with acute ischemic stroke
- Optimized preprocessing and Tiny ML for Attention State Classification
- Unbiased Supervised Contrastive Learning
- Applications of AI and HPC in the Health Domain
- Marco Aldinucci, David Atienza, Federico Bolelli, Monica Caballero, Iacopo Colonnelli, Jose Flich, Jon A. Gomez, David Gonzalez, Costantino Grana, Marco Grangetto, Simone Leo, Pedro Lopez, Dana Oniga, Roberto Paredes, Luca Pireddu, Eduardo Quinones, Tatiana Silva
- The DeepHealth HPC Infrastructure: Leveraging Heterogenous HPC and Cloud-Computing Infrastructures for IA-Based Medical Solutions
- The DeepHealth Toolkit: A Key European Free and Open-Source Software for Deep Learning and Computer Vision Ready to Exploit Heterogeneous HPC and Cloud Architectures
- A Hitchhiker’s Guide towards Transactive Memory System Modeling in Small Group Interactions
- Dysplasia grading of colorectal polyps through convolutional neural network analysis of whole slide images
- UniToPatho, a labeled histopathological dataset for colorectal polyps classification and adenoma dysplasia grading
- Communication scheduling and buslet synthesis for low-interconnect HLS designs