Visual multimedia have become an inseparable part of our digital social lives, and they often capture moments tied with deep affections. Automated visual sentiment analysis tools can provide a means of extracting the rich feelings and latent dispositions embedded in these media. In this work, we explore how Convolutional Neural Networks (CNNs), a now de facto computational machine learning tool particularly in the area of Computer Vision, can be specifically applied to the task of visual sentiment prediction. We accomplish this through fine-tuning experiments using a state-of-the-art CNN and via rigorous architecture analysis, we present several modifications that lead to accuracy improvements over prior art on a dataset of images from a popular social media platform. We additionally present visualizations of local patterns that the network learned to associate with image sentiment for insight into how visual positivity (or negativity) is perceived by the model.

No abstract

No abstract

Abstract:In this paper, we study the geodetic convexity of graphs focusing on the problem of the complexity to compute inclusion-minimum hull set of a graph in several graph classes.For any two vertices u, v ∈ V of a connected graph G = (V, E), the closed interval I [u, v] of u and v is the the set of vertices that belong to some shortestIn other words, a subset S is convex if, for any u, v ∈ S and for any shortest (u, v)-path P, V (P) ⊆ S. Given a subset S ⊆ V , the convex hull I h [S] of S is the smallest convex set that contains S. We say that S is a hull set of G if I h [S] = V . The size of a minimum hull set of G is the hull number of G, denoted by hn(G). The HULL NUMBER problem is to decide whether hn(G) ≤ k, for a given graph G and an integer k. Dourado et al. showed that this problem is NP-complete in general graphs.In this paper, we answer an open question of Dourado et al. [12] by showing that the HULL NUMBER problem is NP-hard even when restricted to the class of bipartite graphs. Then, we design polynomial time algorithms to solve the HULL NUMBER problem in several graph classes. First, we deal with the class of complements of bipartite graphs. Then, we generalize some results in [1] to the class of (q, q − 4)-graphs and to the class of cacti. Finally, we prove tight upper bounds on the hull numbers. In particular, we show that the hull number of an n-node graph G without simplicial vertices is at most 1 + ⌈ Le nombre enveloppe de quelques classes de graphesRésumé : Dans cet article nousétudions une notion de convexité dans les graphes. Nous nous concentrons sur la question de la compléxité du calcul de l'enveloppe minimum d'un graphe dans le cas de diverses classes de graphes. Etant donné un graphe G = (V, E), l'intervalle I [u, v] entre deux sommets u, v ∈ V est l'ensemble des sommets qui appartiennentà un plus court chemin entre u et v. Pour un ensemble S ⊆ V , on note I [S] l'ensemble u,v∈S I[u, v].Nous montrons que décider si hn(G) ≤ k est un problème NP-complet dans la classe des graphes bipartis et nous prouvons que hn(G) peutêtre calculé en temps polynomial pour les cobipartis, (q, q − 4)-graphes et cactus. Nous montrons aussi des bornes supérieures du nombre enveloppe des graphes en général, des graphes sans triangles et des graphes réguliers.

No abstract