17837 research outputs found
Sort by
Facial Emotion Recognition from Feature Loss Media:Human versus Machine Learning Algorithms
Full text linkA dynamic programming algorithm for the maximum s-club problem on trees
Full text linkComputing cliques in an undirected graph G = (V_G, E_G) is a fundamental problem in social network analysis. However, in some cases, the strict definition of a clique (a subset of vertices pairwise adjacent in G) often limits its applicability in real-world settings. To address this issue, we study the s-club: a clique relaxation that induces a subgraph of diameter at most s. Note that a clique is simply a 1-club. Computing a maximum s-club is a computationally challenging problem, as it is NP-hard for any positive integer s in arbitrary graphs. Thus, this paper presents a simple dynamic programming algorithm that efficiently computes a maximum s-club on an n-vertex tree in O(s.n) time. This algorithm outperforms existing algorithms for trees in theory and practice. This approach is a stepping stone towards computing maximum s-clubs on tree-like graphs
The Gestural Potential of Music: Identifying Musical Meaning, Engagement and Immersion in Video Game Music
Full text linkVideogame music engages players, summoning us into virtual worlds and soundscapes, encouraging us to adhere to the ludic parameters and play. This article establishes a new gestural analytical framework tailored to the playful audiovisual individualities of videogame design to reveal how players might become engaged in games. Case studies examined here include Super Mario World (1990) and Super Metroid (1994). I present a new analytical theory, graphically mapping gestures to determine the ways in which videogame music can successfully engage players to feel part of the ludo-narrative journey through a concept I term the ‘gestural potential’ of music
Wasserstein Distance-based Graph Contrastive Learning for Recommendation
Full text linkGraph contrastive learning (GCL) is able to learn augmentation-invariant representations from raw data and reduce the dependence on labeled data. In the field of recommendation systems, traditional GCL models become a potential solution to insufficient supervision signals by augmenting the user-item interaction graph and optimizing InfoNCE loss to learn user and item representations. However, existing GCL-based recommendation models are limited by dimensional collapse, causing the sub-optimal performance of recommendation models. To tackle this problem, we propose a Wasserstein Distance-based Graph Contrastive Learning model, namely WGCL. Specifically, we integrate the Wasserstein loss into contrastive learning-based recommendation models to align the user/item representations distribution with the isotropic Gaussian distribution, which makes the real distribution of representations more uniform, thereby alleviating dimensional collapse. In fact, Wasserstein loss measures the distinction between the real distribution of entities’ representations and the desired distribution of representations by computing the covariance of representations learned from the augmented views. As a result, Wasserstein distance metric not only enables the representations more uniformly distributed on the hypersphere, but also better preserves the original semantic information of entities. Extensive experiments conducted on three widely used datasets demonstrate that WGCL outperforms traditional recommendation models. Our code is released at https://github.com/Sodapease/WGC
Data sources:
Access Repository Dashboard
Do you manage Royal Holloway - Pure? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!