Searcharxiv⌕ Search

arXiv subjects

Enzo Tartaglione

Publications and source records attributed to Enzo Tartaglione.

At least 19 recordsLinked to original sources
arXiv2026

Exploiting Biases in Attention for Transformers

Attention is known to be sparse in vision transformers. In particular, we analyze the unequal relevance of attention-matrix columns to the final computation and show that this phenomenon becomes increasingly pronounced with network depth. This can be interpreted as saying that tokens have an absolute importance for the attention process, independently of queries. Inspired by this, we propose a reformulation of the attention formula that makes this phenomenon explicit by adding an input-dependent bias to the query-key dot product. We show that this new term gives each token a scalar value indicating its anticipated influence on other tokens during the attention process, and can thus serve as a proxy for pruning attention maps. Across multiple transformer model sizes, this allows us to substantially reduce the computational cost of attention, in terms of both FLOPs and inference latency, while retaining most of the original classification accuracy. We publicly released the code and models at https://github.com/EIDOSLAB/AttentionWithBias.

cs.LG↗
arXiv2026

GoDe: Gaussians on Demand for Backbone-Agnostic Progressive Compression

Splat-based radiance fields have rapidly diversified in primitive type, kernel support, and appearance parameterization. This diversity makes compression difficult: most codecs are single-rate, requiring a separately optimized model for each target bitrate, while progressive codecs remain tied to representation-specific assumptions that do not transfer across backbones. We introduce GoDe (Gaussians on Demand), a progressive compression framework whose central contribution is backbone agnosticism. It reorganizes any trained differentiable splat representation into a single scalable bitstream without modifying the backbone or re-deriving the compression structure for each new primitive type. GoDe builds a nested level-of-detail hierarchy via masked-gradient importance ranking, refines all levels with one quantization-aware fine-tuning stage, and stores the base layer and enhancement layers as progressive chunks compressed with shared entropy dictionaries. Any decoded prefix is directly renderable, so one training pipeline provides multiple rate-distortion operating points without per-rate retraining. Across five heterogeneous backbones and three standard benchmarks, GoDe achieves competitive rate-distortion performance while enabling faster rendering and substantially faster progressive decoding than existing methods. Project page: https://gaussians-on-demand.github.io/

cs.CV↗
arXiv2026

Validity-Preserving Hierarchical RL for Joint Routing and Switch Placement in EDA

Routing and switch placement are fundamental combinatorial optimization problems in chip design, requiring the joint optimization of routing topology and physical placement under strict structural, geometric and logical constraints. Existing approaches typically rely on carefully engineered heuristics that incorporate strong problem-specific biases to navigate the enormous space of possible designs. In this work, we introduce a hierarchical reinforcement learning framework for joint routing and switch placement at the level of logical communication routes. Starting from a minimal routing graph, our method progressively constructs increasingly expressive solutions through three coupled operations: switch expansion, switch placement, and route refinement. These operations preserve routing validity by construction, restricting exploration to feasible configurations where every communicating initiator-target pair has one assigned loop-free route. We explore the induced solution space using Gumbel Monte Carlo Tree Search, showing that neural-guided search substantially improves solution quality over non-learning optimization methods. Furthermore, pretraining across floorplans provides a strong initialization for fine-tuning on unseen instances.

cs.LG↗
arXiv2026

From Deep to Shallow: Unconstrained and Efficient Layer Merging Strategy

Although Deep Neural Networks have become foundational in many areas of Machine Learning, high computational demands limit their application in resource-constrained environments. To address this issue, depth compression methods have been proposed to identify and linearize redundant activation functions, thereby allowing for the merging of layers without intermediate non-linearities. However, these methods face two key challenges: they cannot be directly applied to convolutions with padding due to the absence of an analytical solution for merging these layers, and they typically increase the kernel size of merged layers, thus limiting speed-up gains. To overcome these limitations, we propose an efficient strategy that enables merging of layers without an existing analytical solution, and also without increasing kernel size. We validate our approach across multiple architectures and datasets, and measure inference speed-up gains on real embedded platforms. We publicly released the code at https://github.com/ShulzhenkoPetr/deep-to-shallow.

cs.LG↗
arXiv2026

HYGENE: A Diffusion-based Hypergraph Generation Method

Hypergraphs are powerful mathematical structures that can model complex, high-order relationships in various domains, including social networks, bioinformatics, and recommender systems. However, generating realistic and diverse hypergraphs remains challenging due to their inherent complexity and lack of effective generative models. In this paper, we introduce a diffusion-based Hypergraph Generation (HYGENE) method that addresses these challenges through a progressive local expansion approach. HYGENE works on the bipartite representation of hypergraphs, starting with a single pair of connected nodes and iteratively expanding it to form the target hypergraph. At each step, nodes and hyperedges are added in a localized manner using a denoising diffusion process, which allows for the construction of the global structure before refining local details. Our experiments demonstrated the effectiveness of HYGENE, proving its ability to closely mimic a variety of properties in hypergraphs. To the best of our knowledge, this is the first attempt to employ deep learning models for hypergraph generation, and our work aims to lay the groundwork for future research in this area.

cs.LG↗
arXiv2026

Feature-Aware (Hyper)graph Generation via Next-Scale Prediction

Graph generative models perform well on small-scale structured data but struggle to scale to large, complex structures. Hierarchical approaches improve scalability but often ignore node and edge features, which are critical in real-world applications. In this paper, we propose FAHNES (feature-aware (hyper)graph generation via next-scale prediction), a hierarchical framework that jointly generates topology and features for graphs and hypergraphs. FAHNES progressively constructs the final sample through localized expansion and refinement, guided by a novel node budget controlling granularity and ensuring cross-scale consistency. Experiments on synthetic, 3D mesh, and graph point cloud datasets demonstrate competitive or state-of-the-art performance while uniquely scaling to large-scale graphs and hypergraphs with features. Our code is open source.

cs.LG↗
arXiv2026

CutClean: Neural Network Pruning for Privacy-Preserving Inference

Neural networks are increasingly deployed in high-stakes applications with growing privacy leakage concerns. We show that this privacy leakage can occur even in the absence of representation imbalances that lead to traditional dataset biases. This poses significant privacy risks when deploying models that process sensitive attributes. In this context, we propose CutClean, a privacy-aware pruning method that allows to reduce privacy information flow through the network, while increasing its sparsity. Our approach employs auxiliary linear privacy heads placed at each network's block to quantify information leakage, and further applies increasing levels of sparsity to remove the private attribute leakage, measured in terms of the accuracy of the privacy head attached to the last block. Experiments on synthetic and real-world datasets demonstrate that our approach effectively minimizes private information flow while achieving high sparsity rates and preserving classification target accuracy.

cs.LG↗
arXiv2026

Bias In, Bias Out? Finding Unbiased Subnetworks in Vanilla Models

The issue of algorithmic biases in deep learning has led to the development of various debiasing techniques, many of which perform complex training procedures or dataset manipulation. However, an intriguing question arises: is it possible to extract fair and bias-agnostic subnetworks from standard vanilla-trained models without relying on additional data, such as unbiased training set? In this work, we introduce Bias-Invariant Subnetwork Extraction (BISE), a learning strategy that identifies and isolates "bias-free" subnetworks that already exist within conventionally trained models, without retraining or finetuning the original parameters. Our approach demonstrates that such subnetworks can be extracted via pruning and can operate without modification, effectively relying less on biased features and maintaining robust performance. Our findings contribute towards efficient bias mitigation through structural adaptation of pre-trained neural networks via parameter removal, as opposed to costly strategies that are either data-centric or involve (re)training all model parameters. Extensive experiments on common benchmarks show the advantages of our approach in terms of the performance and computational efficiency of the resulting debiased model.

cs.LG↗
arXiv2026

Nix and Fix: Targeting 1000x Compression of 3D Gaussian Splatting with Diffusion Models

3D Gaussian Splatting (3DGS) revolutionized novel view rendering. Instead of inferring from dense spatial points, as implicit representations do, 3DGS uses sparse Gaussians. This enables real-time performance but increases space requirements, hindering rate-constrained applications. 3DGS compression emerged as a field aimed at alleviating this issue. While impressive progress has been made, at low rates, compression introduces artifacts that degrade visual quality significantly. We introduce NiFi, a method for extreme 3DGS compression through restoration via artifact-aware, diffusion-based one-step distillation. We show that our method achieves state-of-the-art perceptual quality at extremely low rates, down to 0.1 MB, and towards 1000x rate improvement over 3DGS at comparable perceptual performance. Code is available at: https://github.com/ceteke/nifi

cs.CV↗
arXiv2026

I-INR: Iterative Implicit Neural Representations

Implicit Neural Representations (INRs) have revolutionized signal processing and computer vision by modeling signals as continuous, differentiable functions parameterized by neural networks. However, INRs are prone to the spectral bias problem, limiting their ability to retain high-frequency information, and often struggle with noise robustness. Motivated by recent trends in iterative refinement processes, we propose Iterative Implicit Neural Representations (I-INRs). This novel plug-and-play framework iteratively refines signal reconstructions to restore high-frequency details, improve noise robustness, and enhance generalization, ultimately delivering superior reconstruction quality. I-INRs integrate seamlessly into existing INR architectures with only a 0.5-2% increase in parameters. During reconstruction, the iterative refinement adds just 0.8-1.6% additional FLOPs over the baseline while delivering a substantial performance boost of up to +2.0 PSNR. Extensive experiments demonstrate that I-INRs consistently outperform WIRE, SIREN, and Gauss across various computer vision tasks, including image fitting, image denoising, and object occupancy prediction. The code is available at github.com/optimizer077/I-INR.

cs.CV↗
arXiv2026

The silence of the weights: a structural pruning strategy for attention-based audio signal architectures with second order metrics

Transformer-based models have become the state of the art across multiple domains, from natural language processing to machine listening, thanks to the attention mechanisms. However, the attention layers require a large number of parameters and high-end hardware for both training and inference. We propose a novel channel-pruning technique explicitly targeted at the attention mechanism, decoupling the pruning of each head and the four layers in the attention block: query, key, value, and output projection matrices, employing a second-order metric to score the network's parameters. We compare our technique against head-pruning strategies and magnitude-driven scoring metrics, investigating the effects of pruning on Audio Spectrogram Transformer (AST) and Whisper. Our results show that even after pruning 50\% of the parameters in the attention block, performance is largely preserved.

cs.SD↗
arXiv2026

Efficient Resource-Constrained Training of Transformers via Subspace Optimization

As AI increasingly shapes daily life, energy consumption and data privacy have become pressing concerns. On-device learning trains models directly on edge devices, cutting energy consumption and safeguarding data privacy. However, the expanding scale of modern neural networks creates a major obstacle for on-device training. Although prior work has concentrated on compact convolutional architectures, we instead apply subspace-based training to transformer models. Motivated by the idea that a model's essential information lies in a fixed subspace, we introduce Weight-Activation Subspace Iteration (WASI), a method that mitigates the memory bottleneck of backpropagation and boosts inference efficiency in transformer models by restricting training to this subspace. Our results demonstrate that WASI maintains accuracy comparable to vanilla training while reducing memory usage by up to $62\times$ and computational cost (FLOPs) by up to $2\times$. On a Raspberry Pi 5, WASI achieves roughly $1.4\times$ faster training and inference than vanilla training. The code is available at https://github.com/Le-TrungNguyen/ICLR2026-WASI.git.

cs.LG↗
arXiv2026

Layer Collapse Can be Induced by Unstructured Pruning

Unstructured pruning is a popular compression method for efficiently reducing model parameters. However, while it effectively decreases the number of parameters, it is commonly believed that unstructured pruning cannot shorten the computational critical path, i.e., the maximum number of layers traversed during forward propagation. In this paper, we study when and how unstructured pruning can yield structural effects. For rectifier-activated networks, we introduce the notion of neuron entropy, which quantifies the degree of nonlinearity utilization. We show that magnitude-based pruning naturally lowers this entropy, sometimes down to zero-entropy layers that become linearizable and can thus be removed. Building on this insight, we propose a method that leverages "unstructured" pruning to favor sparsity in low-entropy layers, enabling their complete removal. We validate the phenomenon across CNNs, Vision Transformers, and NLP models: unstructured pruning can induce effective layer removal with little or no performance degradation in over-parameterized networks.

cs.LG↗
arXiv2026

Study of Training Dynamics for Memory-Constrained Fine-Tuning

Memory-efficient training of deep neural networks has become increasingly important as models grow larger while deployment environments impose strict resource constraints. We propose TraDy, a novel transfer learning scheme leveraging two key insights: layer importance for updates is architecture-dependent and determinable a priori, while dynamic stochastic channel selection provides superior gradient approximation compared to static approaches. We introduce a dynamic channel selection approach that stochastically resamples channels between epochs within preselected layers. Extensive experiments demonstrate TraDy achieves state-of-the-art performance across various downstream tasks and architectures while maintaining strict memory constraints, achieving up to 99% activation sparsity, 95% weight derivative sparsity, and 97% reduction in FLOPs for weight derivative computation.

cs.LG↗
arXiv2026

HFMCA: Orthonormal Feature Learning for EEG-based Brain Decoding

Electroencephalography (EEG) analysis is critical for brain-computer interfaces and neuroscience, but the intrinsic noise and high dimensionality of EEG signals hinder effective feature learning. We propose a self-supervised framework based on the Hierarchical Functional Maximal Correlation Algorithm (HFMCA), which learns orthonormal EEG representations by enforcing feature decorrelation and reducing redundancy. This design enables robust capture of essential brain dynamics for various EEG recognition tasks. We validate HFMCA on two benchmark datasets, SEED and BCIC-2A, where pretraining with HFMCA consistently outperforms competitive self-supervised baselines, achieving notable gains in classification accuracy. Across diverse EEG tasks, our method demonstrates superior cross-subject generalization under leave-one-subject-out validation, advancing state-of-the-art by 2.71\% on SEED emotion recognition and 2.57\% on BCIC-2A motor imagery classification.

eess.SP↗
arXiv2026

RAVE: Rate-Adaptive Visual Encoding for 3D Gaussian Splatting

Recent advances in neural scene representations have transformed immersive multimedia, with 3D Gaussian Splatting (3DGS) enabling real-time photorealistic rendering. Despite its efficiency, 3DGS suffers from large memory requirements and costly training procedures, motivating efforts toward compression. Existing approaches, however, operate at fixed rates, limiting adaptability to varying bandwidth and device constraints. In this work, we propose a flexible compression scheme for 3DGS that supports interpolation at any rate between predefined bounds. Our method is computationally lightweight, requires no retraining for any rate, and preserves rendering quality across a broad range of operating points. Experiments demonstrate that the approach achieves efficient, high-quality compression while offering dynamic rate control, making it suitable for practical deployment in immersive applications. The code is available at https://github.com/inspiros/RAVE.

cs.CV↗
arXiv2025

How I Met Your Bias: Investigating Bias Amplification in Diffusion Models

Diffusion-based generative models demonstrate state-of-the-art performance across various image synthesis tasks, yet their tendency to replicate and amplify dataset biases remains poorly understood. Although previous research has viewed bias amplification as an inherent characteristic of diffusion models, this work provides the first analysis of how sampling algorithms and their hyperparameters influence bias amplification. We empirically demonstrate that samplers for diffusion models -- commonly optimized for sample quality and speed -- have a significant and measurable effect on bias amplification. Through controlled studies with models trained on Biased MNIST, Multi-Color MNIST and BFFHQ, and with Stable Diffusion, we show that sampling hyperparameters can induce both bias reduction and amplification, even when the trained model is fixed. Source code is available at https://github.com/How-I-met-your-bias/how_i_met_your_bias.

cs.LG↗
arXiv2025

Diffusing DeBias: Synthetic Bias Amplification for Model Debiasing

Deep learning model effectiveness in classification tasks is often challenged by the quality and quantity of training data whenever they are affected by strong spurious correlations between specific attributes and target labels. This results in a form of bias affecting training data, which typically leads to unrecoverable weak generalization in prediction. This paper aims at facing this problem by leveraging bias amplification with generated synthetic data: we introduce Diffusing DeBias (DDB), a novel approach acting as a plug-in for common methods of unsupervised model debiasing exploiting the inherent bias-learning tendency of diffusion models in data generation. Specifically, our approach adopts conditional diffusion models to generate synthetic bias-aligned images, which replace the original training set for learning an effective bias amplifier model that we subsequently incorporate into an end-to-end and a two-step unsupervised debiasing approach. By tackling the fundamental issue of bias-conflicting training samples memorization in learning auxiliary models, typical of this type of techniques, our proposed method beats current state-of-the-art in multiple benchmark datasets, demonstrating its potential as a versatile and effective tool for tackling bias in deep learning models. Code is available at https://github.com/Malga-Vision/DiffusingDeBias

cs.LG↗