arXiv subjects

Sebastiano Vascon

Publications and source records attributed to Sebastiano Vascon.

At least 19 recordsLinked to original sources

arXiv2026

Fuse To Align: Multimodal Representation Learning for Retrieval via Hypergraph Regularization

Multimodal retrieval must integrate complementary evidence from views that may be missing, noisy, or unequally informative. When a sample provides limited reliable evidence, learning from its views alone can hinder the acquisition of useful retrieval representations. Semantically related training samples may provide complementary observations, yet existing volume-based objectives jointly align modalities within each sample without explicitly enabling information exchange across samples. We therefore propose a training-time hypergraph fusion framework that connects observed modalities both within samples and across semantically related samples, allowing representation learning to draw on evidence beyond an individual sample while leaving the retrieval pipeline unchanged at inference. The framework includes HyVol as a special case and introduces two configurations: wHyVol learns view contribution weights and symmetric semantic affinities, whereas wRHG additionally models typed interactions and participant roles through relational propagation. Missing modalities are omitted rather than imputed, and the hypergraph is discarded after training, incurring no additional inference cost. Experiments on eleven zero-shot video-text and audio-text retrieval benchmarks demonstrate broad improvements with both GRAM and HyperGRAM. Across 17 video benchmark-modality settings evaluated in both retrieval directions, wRHG outperforms GRAM in all 34 comparisons and HyperGRAM in 33 of 34. Mean text-to-video R@1 increases from 63.4 to 68.2 with GRAM and from 64.8 to 68.4 with HyperGRAM, while the largest video-to-text gain reaches 13.3 percentage points on ActivityNet. Under missing-modality evaluation, both weighted configurations retain positive text-to-video gains over GRAM across seven benchmarks at every evaluated missingness rate, even when 90% of clips lack one modality.

cs.CV↗

arXiv2026

A Multi-Layer Navigation System for Smart Wheelchairs with a Limited Input Interface

Smart wheelchairs are robotic devices that gather information from the surroundings and allow people to safely drive through various environments. They usually rely on the interpretation of high-level user's intentions, typically decoded through human-robot interface (HRI) or brain-machine interface (BMI). However, navigating in an unknown environment and accurately understanding user commands remain significant challenges. Misinterpretation of commands or the environment can lead to user frustration, distrust of the system, or even unsafe situations. In this paper, we present a novel Multi-Layer Navigation System (MLNS) for a smart wheelchair that enables navigation in unknown environments while interpreting and safely contextualizing user's intentions even in the case of limited and faulty commands. The stack relies on a Virtual Sensors-Driven Potential Fields (ViSPF) method for obstacle detection that accounts for the wheelchair's actual size, a Model Predictive Control (MPC) module for trajectory generation, and an open-vocabulary computer vision algorithm for target identification. Finally, a system based on Priority Queue (PQ) was implemented for target selection. We validated our system on a real smart wheelchair with experiments involving 10 participants without disabilities, including scenarios with faulty input interfaces. The results demonstrate the robustness and effectiveness of our system in real-world environments, proving its potential for practical applications in assistive technologies.

cs.RO↗

arXiv2026

Min-Cost Flow Routing for Evidence Assembly in Long Multimodal Documents

Answering questions about long multimodal documents requires distributing a fixed evidence budget across relevant facets in text, tables, figures, and slides while avoiding near-duplicates. We present \flowreader, which formulates evidence selection as a single minimum-cost flow problem with capacity limits over a multimodal content graph. Spectral decomposition identifies latent aspects of query-relevant content and allocates the budget among them in proportion to their spectral energy. These capacity limits enforce aspect coverage during routing without requiring a language-model planning call. Query-conditioned costs prioritize chains of relevant, mutually consistent evidence. Decomposing the optimal flow produces short evidence chains, which a vision-language model reads in parallel and a reasoner reconciles. On VisDoMBench with Qwen3-VL-32B, \flowreader\ achieves the highest macro accuracy ($68.9$), surpassing the strongest prior system by $2.7$ points, leading on three of five subsets and attaining the highest worst-subset accuracy. It uses a measured $17.5$ content nodes per query and maintains its lead at $12.9$. Ablation studies with a fixed graph, scorer, reader, and judge show that cost design drives accuracy, capacity limits preserve it while using about three-quarters of the reader tokens required by shortest-path routing without these limits on the same network, and spectral aspects align with LLM-generated sub-questions without a planning call.

cs.IR↗

arXiv2026

Hypergraph-Regularized Gramian Volumes for Multimodal Retrieval

Volume-based multimodal retrieval jointly scores a text query with a candidate's video, audio, and subtitle embeddings. While this approach captures higher-order within-candidate alignment, the score remains candidate-local, and semantically related training samples primarily serve as contrastive negatives. This work introduces Hypergraph-Regularized Gramian Volumes (HyVol), a training-time module that incorporates these semantic relations prior to evaluating the original volume loss. Document hyperedges connect the observed modalities of each candidate, whereas semantic hyperedges link candidates whose detached captions are mutual top-k neighbors. A shallow gated hyper-graph network applies residual corrections to the modality embeddings. Presence masks exclude unavailable streams from message passing, and identity padding preserves the determinant of the observed Gram submatrix without feature imputation. As refinement operates on embeddings rather than scores, the same construction applies to both Gram and HyperGram. We remove the hypergraph after training, leaving the backbone-only architecture, original scoring function, and retrieval cost unchanged. We train both backbones on a 150K-clip subset of VAST-27M and evaluate zero-shot performance on six benchmarks. Under the paired protocol, HyVol improves R@1 across all five retrieval benchmarks, with video-to-text gains reaching +8.3 on MSR-VTT and +7.6 on VATEX. Under missing-modality masking, the V2T margin remains positive in all experimental settings, although the T2V margin becomes slightly negative in four.

cs.CV↗

arXiv2026

Query-Conditioned Spherical Centroid Aggregation for Multimodal Retrieval

Multimodal retrieval integrates video, audio, subtitles, and text; however, recent geometric aggregators, such as Gramian volumes, hyperbolic volumes, and spectral objectives, treat all modalities symmetrically. Under a unified evaluation protocol, their joint scores frequently lag behind the strongest single-modality pathway by 1.9 to 27.6 R@1. Controlled analyses attribute this outcome to uniform modality influence. This work introduces Spherical Centroid Aggregation with Learned Adaptive Relevance (SCALAR), a query-conditioned aggregator that assigns relevance-based weights to each available modality before computing a spherical centroid. SCALAR accommodates arbitrary modality subsets and is trained on masked, reduced-arity views using rank-8 LoRA adapters. Across five benchmarks, SCALAR achieves positive aggregation gain on four, reaching +4.0 R@1, while none of the evaluated prior aggregators is positive on more than one. A uniform-weight ablation reproduces the degradation observed with symmetric aggregation. With only 4.8 million trainable parameters, SCALAR attains the highest text-to-video R@1 on three and performs within seed variation of the best result on a fourth. Under test-time modality dropout, SCALAR's representation-stage score surpasses the released GRAM checkpoint at every evaluated masking rate and benchmark by 3.2 to 10.9 R@1. Finally, as modalities are removed, rerankers trained exclusively on complete modality sets increasingly converge toward their video-only pathways, diminishing these representation-level gains and underscoring a limitation of standard two-stage retrieval pipelines.

cs.CV↗

arXiv2026

CoRE: Consensus Rewards via Equilibrium for Test-Time Reinforcement Learning

On unlabeled test data, reinforcement learning lacks a ground-truth reward; test-time RL methods derive one from the model's own roll-outs, rewarding those that match the majority vote over $N$ sampled answers. That vote discards a correct answer whenever it is a minority and scores every majority-matching roll-out identically. We replace it with \emph{CoRE} (Consensus Rewards via Equilibrium): the $N$ roll-outs form a graph whose edges combine answer agreement, reasoning similarity, and generation confidence, and replicator dynamics extract its dominant set, yielding a refined pseudo-label, a graded per-roll-out reward, and a per-question cohesiveness gate. CoRE strictly generalizes voting: majority voting is recovered as a special case; a block-value analysis gives a sharp threshold for when consensus recovers a correct minority against a larger wrong plurality; and confidence calibration provably lowers that threshold multiplicatively. Across seven backbones and five benchmarks (42 model--benchmark cells, three seeds each), \emph{CoRE} improves the untrained base by $+21.7$ points on average versus $+20.4$ for majority-vote TTRL, wins wherever agreement is contestable with margins over the vote of up to $+7.5$ points, and reaches the voting baseline's plateau accuracy in $54$--$70$\% fewer steps. Consensus, not counting: treating the roll-out group as a graph rather than a ballot box turns a brittle vote into a calibrated, graded, self-supervised reward at no extra roll-out cost.

cs.AI↗

arXiv2026

RAG-Audio: Retrieval-Augmented Generation for Faithful Brain-to-Audio Reconstruction

Brain-to-audio reconstruction is limited by \emph{prior domination}: when a pretrained generator is conditioned on a weak neural signal, it produces realistic but stimulus-inaccurate audio. We introduce RAG-Audio, which decodes fMRI into a semantic audio embedding, retrieves a matching real-audio exemplar, and initializes the frozen generator's sampling trajectory from that exemplar while retaining the decoded embedding as conditioning. On Brain2Music, RAG-Audio improves 10-way stimulus identification from $0.14$--$0.18$ for direct generation, near the $0.10$ chance level, to $0.40$--$0.43$, comparable to retrieval. It also reduces Fréchet Audio Distance by roughly an order of magnitude, from $13.49$ to $1.25$ for AudioLDM. RAG-Audio approaches nearest-neighbor retrieval in identification while remaining generative; its higher FAD is expected because retrieval directly replays real audio. An autoregressive negative control, which lacks an initializable latent trajectory, shows no comparable gain, attributing the improvement to trajectory initialization. These results suggest that retrieval-guided initialization can mitigate prior domination in brain-to-audio generation.

cs.SD↗

arXiv2026

Constrained Dominant Sets for Multimodal Document Question Answering

Long multimodal document question answering is limited by which evidence reaches the reader, rather than by the quantity retrieved. In lengthy documents, findings often recur across figures, captions, and introductory sentences, causing similarity based retrievers in modern multimodal retrieval-augmented generation (RAG) systems to allocate resources to near-duplicates while overlooking complementary evidence. This work introduces a retriever that selects evidence as a Constrained Dominant Set (CDS) on a query-augmented affinity graph, offering three advantages that similarity ranking does not. First, the query is encoded as a hard structural constraint, ensuring that every selected element is directly connected to the question through the cluster anchor. Second, the relevance-redundancy balance is determined automatically by a spectral bound, eliminating the need for manually tuned trade offs required by diversity-aware selectors. Third, the selection process achieves a global equilibrium via replicator dynamics, thereby avoiding the distortions introduced by greedy heuristics. The method is inherently graph-based and does not require training. Using a Qwen3-VL-32B reader, CDS establishes a new state of the art on VisDoMBench ($66.99$ average) and improves over the no-retrieval baseline by $37.1$ points on VisDoMBench and $4.8$ on MMLongBench-Doc.

cs.IR↗

arXiv2026

Solving Jigsaw Puzzles in the Wild: Human-Guided Reconstruction of Cultural Heritage Fragments

Reassembling real-world archaeological artifacts from fragmented pieces poses significant challenges due to erosion, missing regions, irregular shapes, and large-scale ambiguity. Traditional jigsaw puzzle solvers, often designed for clean synthetic scenarios, struggle under these conditions, especially when the number of fragments grows into the thousands, as in the RePAIR benchmark. In this paper, we propose a human-in-the-loop (HIL) puzzle solving framework designed to address the complexity and scale of real-world cultural heritage reconstruction. Our approach integrates an automatic relaxation-labeling solver with interactive human guidance, allowing users to iteratively lock verified placements, correct errors, and guide the system toward semantically and geometrically coherent assemblies. We introduce two complementary interaction strategies, Iterative Anchoring and Continuous Interactive Refinement, which support scalable reconstruction across varying levels of ambiguity and puzzle size. Experiments on several RePAIR groups demonstrate that our hybrid approach substantially outperforms both fully automatic and manual baselines in accuracy and efficiency, offering a practical solution for large-scale expert-in-the-loop artifact reassembly.

cs.CV↗

arXiv2025

ECAM: A Contrastive Learning Approach to Avoid Environmental Collision in Trajectory Forecasting

Human trajectory forecasting is crucial in applications such as autonomous driving, robotics and surveillance. Accurate forecasting requires models to consider various factors, including social interactions, multi-modal predictions, pedestrian intention and environmental context. While existing methods account for these factors, they often overlook the impact of the environment, which leads to collisions with obstacles. This paper introduces ECAM (Environmental Collision Avoidance Module), a contrastive learning-based module to enhance collision avoidance ability with the environment. The proposed module can be integrated into existing trajectory forecasting models, improving their ability to generate collision-free predictions. We evaluate our method on the ETH/UCY dataset and quantitatively and qualitatively demonstrate its collision avoidance capabilities. Our experiments show that state-of-the-art methods significantly reduce (-40/50%) the collision rate when integrated with the proposed module. The code is available at https://github.com/CVML-CFU/ECAM.

cs.CV↗

arXiv2024

Backdoor Learning Curves: Explaining Backdoor Poisoning Beyond Influence Functions

Backdoor attacks inject poisoning samples during training, with the goal of forcing a machine learning model to output an attacker-chosen class when presented a specific trigger at test time. Although backdoor attacks have been demonstrated in a variety of settings and against different models, the factors affecting their effectiveness are still not well understood. In this work, we provide a unifying framework to study the process of backdoor learning under the lens of incremental learning and influence functions. We show that the effectiveness of backdoor attacks depends on: (i) the complexity of the learning algorithm, controlled by its hyperparameters; (ii) the fraction of backdoor samples injected into the training set; and (iii) the size and visibility of the backdoor trigger. These factors affect how fast a model learns to correlate the presence of the backdoor trigger with the target class. Our analysis unveils the intriguing existence of a region in the hyperparameter space in which the accuracy on clean test samples is still high while backdoor attacks are ineffective, thereby suggesting novel criteria to improve existing defenses.

cs.LG↗

arXiv2024

Re-assembling the past: The RePAIR dataset and benchmark for real world 2D and 3D puzzle solving

This paper proposes the RePAIR dataset that represents a challenging benchmark to test modern computational and data driven methods for puzzle-solving and reassembly tasks. Our dataset has unique properties that are uncommon to current benchmarks for 2D and 3D puzzle solving. The fragments and fractures are realistic, caused by a collapse of a fresco during a World War II bombing at the Pompeii archaeological park. The fragments are also eroded and have missing pieces with irregular shapes and different dimensions, challenging further the reassembly algorithms. The dataset is multi-modal providing high resolution images with characteristic pictorial elements, detailed 3D scans of the fragments and meta-data annotated by the archaeologists. Ground truth has been generated through several years of unceasing fieldwork, including the excavation and cleaning of each fragment, followed by manual puzzle solving by archaeologists of a subset of approx. 1000 pieces among the 16000 available. After digitizing all the fragments in 3D, a benchmark was prepared to challenge current reassembly and puzzle-solving methods that often solve more simplistic synthetic scenarios. The tested baselines show that there clearly exists a gap to fill in solving this computationally complex problem.

cs.CV↗

arXiv2024

Nash Meets Wertheimer: Using Good Continuation in Jigsaw Puzzles

Jigsaw puzzle solving is a challenging task for computer vision since it requires high-level spatial and semantic reasoning. To solve the problem, existing approaches invariably use color and/or shape information but in many real-world scenarios, such as in archaeological fresco reconstruction, this kind of clues is often unreliable due to severe physical and pictorial deterioration of the individual fragments. This makes state-of-the-art approaches entirely unusable in practice. On the other hand, in such cases, simple geometrical patterns such as lines or curves offer a powerful yet unexplored clue. In an attempt to fill in this gap, in this paper we introduce a new challenging version of the puzzle solving problem in which one deliberately ignores conventional color and shape features and relies solely on the presence of linear geometrical patterns. The reconstruction process is then only driven by one of the most fundamental principles of Gestalt perceptual organization, namely Wertheimer's {\em law of good continuation}. In order to tackle this problem, we formulate the puzzle solving problem as the problem of finding a Nash equilibrium of a (noncooperative) multiplayer game and use classical multi-population replicator dynamics to solve it. The proposed approach is general and allows us to deal with pieces of arbitrary shape, size and orientation. We evaluate our approach on both synthetic and real-world data and compare it with state-of-the-art algorithms. The results show the intrinsic complexity of our purely line-based puzzle problem as well as the relative effectiveness of our game-theoretic formulation.

cs.GT↗

arXiv2024

OpenNav: Efficient Open Vocabulary 3D Object Detection for Smart Wheelchair Navigation

Open vocabulary 3D object detection (OV3D) allows precise and extensible object recognition crucial for adapting to diverse environments encountered in assistive robotics. This paper presents OpenNav, a zero-shot 3D object detection pipeline based on RGB-D images for smart wheelchairs. Our pipeline integrates an open-vocabulary 2D object detector with a mask generator for semantic segmentation, followed by depth isolation and point cloud construction to create 3D bounding boxes. The smart wheelchair exploits these 3D bounding boxes to identify potential targets and navigate safely. We demonstrate OpenNav's performance through experiments on the Replica dataset and we report preliminary results with a real wheelchair. OpenNav improves state-of-the-art significantly on the Replica dataset at mAP25 (+9pts) and mAP50 (+5pts) with marginal improvement at mAP. The code is publicly available at this link: https://github.com/EasyWalk-PRIN/OpenNav.

cs.CV↗

arXiv2023

Reassembling Broken Objects using Breaking Curves

Reassembling 3D broken objects is a challenging task. A robust solution that generalizes well must deal with diverse patterns associated with different types of broken objects. We propose a method that tackles the pairwise assembly of 3D point clouds, that is agnostic on the type of object, and that relies solely on their geometrical information, without any prior information on the shape of the reconstructed object. The method receives two point clouds as input and segments them into regions using detected closed boundary contours, known as breaking curves. Possible alignment combinations of the regions of each broken object are evaluated and the best one is selected as the final alignment. Experiments were carried out both on available 3D scanned objects and on a recent benchmark for synthetic broken objects. Results show that our solution performs well in reassembling different kinds of broken objects.

cs.CV↗

arXiv2023

Wild Patterns Reloaded: A Survey of Machine Learning Security against Training Data Poisoning

The success of machine learning is fueled by the increasing availability of computing power and large training datasets. The training data is used to learn new models or update existing ones, assuming that it is sufficiently representative of the data that will be encountered at test time. This assumption is challenged by the threat of poisoning, an attack that manipulates the training data to compromise the model's performance at test time. Although poisoning has been acknowledged as a relevant threat in industry applications, and a variety of different attacks and defenses have been proposed so far, a complete systematization and critical review of the field is still missing. In this survey, we provide a comprehensive systematization of poisoning attacks and defenses in machine learning, reviewing more than 100 papers published in the field in the last 15 years. We start by categorizing the current threat models and attacks, and then organize existing defenses accordingly. While we focus mostly on computer-vision applications, we argue that our systematization also encompasses state-of-the-art attacks and defenses for other data modalities. Finally, we discuss existing resources for research in poisoning, and shed light on the current limitations and open research questions in this research field.

cs.LG↗

arXiv2022

Geolocation of Cultural Heritage using Multi-View Knowledge Graph Embedding

Knowledge Graphs (KGs) have proven to be a reliable way of structuring data. They can provide a rich source of contextual information about cultural heritage collections. However, cultural heritage KGs are far from being complete. They are often missing important attributes such as geographical location, especially for sculptures and mobile or indoor entities such as paintings. In this paper, we first present a framework for ingesting knowledge about tangible cultural heritage entities from various data sources and their connected multi-hop knowledge into a geolocalized KG. Secondly, we propose a multi-view learning model for estimating the relative distance between a given pair of cultural heritage entities, based on the geographical as well as the knowledge connections of the entities.

cs.LG↗

arXiv2022

The Group Loss++: A deeper look into group loss for deep metric learning

Deep metric learning has yielded impressive results in tasks such as clustering and image retrieval by leveraging neural networks to obtain highly discriminative feature embeddings, which can be used to group samples into different classes. Much research has been devoted to the design of smart loss functions or data mining strategies for training such networks. Most methods consider only pairs or triplets of samples within a mini-batch to compute the loss function, which is commonly based on the distance between embeddings. We propose Group Loss, a loss function based on a differentiable label-propagation method that enforces embedding similarity across all samples of a group while promoting, at the same time, low-density regions amongst data points belonging to different groups. Guided by the smoothness assumption that "similar objects should belong to the same group", the proposed loss trains the neural network for a classification task, enforcing a consistent labelling amongst samples within a class. We design a set of inference strategies tailored towards our algorithm, named Group Loss++ that further improve the results of our model. We show state-of-the-art results on clustering and image retrieval on four retrieval datasets, and present competitive results on two person re-identification datasets, providing a unified framework for retrieval and re-identification.

cs.CV↗