Institution profile

Universidade de Lisboa

Academic institutioneurope · pt
Official website
Research library315linked papers
Opportunities0open roles
Selected work

Representative Papers

New Perspectives on Semiring Applications to Dynamic Programming

Dec 03, 2025

Counting minimum-cost solutions to NP-hard combinatorial optimization problems—such as Connected Dominating Set and Constraint Satisfaction—is computationally challenging due to the interplay between cost minimization and solution enumeration. Method: We propose a unified dynamic programming framework grounded in semiring algebra, centered on a novel Δ-product operation that relaxes the idempotence requirement of classical semirings, enabling efficient counting of minimum-cost solutions over non-idempotent semirings. The framework is parameterized by treewidth and clique-width to ensure tractability on structured inputs. Contribution/Results: We establish fixed-parameter tractable (FPT) enumeration for minimum-cost solutions under bounded treewidth or clique-width, proving polynomial-time solvability of the counting problem in these parameters. Our approach significantly extends classical DP’s expressiveness, unifying cost optimization and solution counting within a single algebraic model—thereby enabling rigorous analysis and efficient computation for previously intractable enumeration tasks.

1 citationsRead paper

A Probability-Guided Sampler for Neural Implicit Surface Rendering

Jun 10, 2025European Conference on Computer Vision

To address inefficient ray sampling and insufficient reconstruction fidelity for foreground implicit surfaces in Neural Radiance Fields (NeRF), this paper proposes an adaptive sampling framework explicitly targeting foreground implicit surfaces. The method models a differentiable probability density function (PDF) directly in the image projection space to guide dense ray sampling within regions of interest. Furthermore, it introduces a novel surface reconstruction loss that jointly optimizes the implicit surface and radiance field by integrating near-surface geometric priors with free-space constraints. Crucially, the approach requires no additional supervision or pretraining and consistently improves mainstream NeRF variants. It significantly enhances geometric accuracy and detail fidelity in target regions while reducing redundant sampling overhead. Experiments demonstrate substantial gains across standard metrics—PSNR, SSIM, and LPIPS—particularly in complex scenes.

1 citationsRead paper
Recent publications

Latest Papers