Morteza Mohammad-Noori

Florent Foucaud, Narges Ghareghani, Lucas Lorieau et al.

In the GEODETIC SET problem, an input is a (di)graph $G$ and integer $k$, and the objective is to decide whether there exists a vertex subset $S$ of size $k$ such that any vertex in $V(G)\setminus S$ lies on a shortest (directed) path between two vertices in $S$. The problem has been studied on undirected and directed graphs from both algorithmic and graph-theoretical perspectives. We focus on directed graphs and prove that GEODETIC SET admits a polynomial-time algorithm on ditrees, that is, digraphs with possible 2-cycles when the underlying undirected graph is a tree (after deleting possible parallel edges). This positive result naturally leads us to investigate cases where the underlying undirected graph is "close to a tree". Towards this, we show that GEODETIC SET on digraphs without 2-cycles and whose underlying undirected graph has feedback edge set number $\textsf{fen}$, can be solved in time $2^{\mathcal{O}(\textsf{fen})} \cdot n^{\mathcal{O}(1)}$, where $n$ is the number of vertices. To complement this, we prove that the problem remains NP-hard on DAGs (which do not contain 2-cycles) even when the underlying undirected graph has constant feedback vertex set number. Our last result significantly strengthens the result of Araújo and Arraes [Discrete Applied Mathematics, 2022] that the problem is NP-hard on DAGs when the underlying undirected graph is either bipartite, cobipartite or split.

Alireza Morsali, MohammadJavad Vaez, Mohammadhossein Soltani et al.

Implicit Neural Representations (INRs) have emerged as a powerful framework for modeling continuous signals. The spectral bias of ReLU-based networks is a well-established limitation, restricting their ability to capture fine-grained details in target signals. While previous works have attempted to mitigate this issue through frequency-based encodings or architectural modifications, these approaches often introduce additional complexity and do not fully address the underlying challenge of learning high-frequency components efficiently. We introduce Sinusoidal Trainable Activation Functions (STAF), designed to directly tackle this limitation by enabling networks to adaptively learn and represent complex signals with higher precision and efficiency. STAF inherently modulates its frequency components, allowing for self-adaptive spectral learning. This capability significantly improves convergence speed and expressivity, making STAF highly effective for both signal representations and inverse problems. Through extensive evaluations across a range of tasks, including signal representation (shape, image, audio) and inverse problems (super-resolution, denoising), as well as neural radiance fields (NeRF), we demonstrate that STAF consistently outperforms state-of-the-art methods in accuracy and reconstruction fidelity. These results establish STAF as a robust solution to spectral bias and the capacity--convergence tradeoff, with broad applicability in computer vision and graphics. Our codebase is publicly accessible at https://github.com/AlirezaMorsali/STAF.