Skip to content 2026 Speakers Sponsor KGC Blog Learn About KGC KGC 2019 KGC 2020 KGC 2021 KGC 2022 KGC 2023 KGC 2024 KGC 2025 Register now Generic selectors Exact matches only Search in title Search in content Post Type Selectors Home / graph neural network Conference news KGC 2020 KGC 2022 KGC 2023 KGC 2024 KGC 2025 KGC Talks Knowledge Graph News Learning material KGC 2025 Share Your Thoughts & Ideas in the KGC Community Survey By François Scharffe & Thomas Deely August 6, 2025August 11, 2025 Thank you

Musings of a Computer Scientist.

Machine learning influences numerous aspects of modern society, empowers new technologies, from Alphago to ChatGPT, and increasingly materializes in consumer products such as smartphones and self-driving cars. Despite the vital role and broad applications of artificial neural networks, we lack systematic approaches, such as network science, to understand their underlying mechanism. The difficulty is rooted in many possible model configurations, each with different hyper-parameters and weighted architectures

Researchers have shown how to write any magnetic pattern desired onto nanowires, which could help computers mimic how the brain processes information.

MAEっぽく, パッチをマスクしたものと元画像の間でSiamese Network https://www.slideshare.net/DeepLearningJP2016/dlmasked-siamese-networks-for-labelefficient-learning

Despite their immense promise in performing a variety of learning tasks, a theoretical understanding of the limitations of Deep Neural Networks (DNNs) has so far eluded practitioners. This is partly due to the inability to determine the closed forms of the learned functions, making it harder to study their generalization properties on unseen datasets. Recent work has shown that randomly initialized DNNs in the infinite width limit converge to kernel machines relying on a Neural Tangent Kernel (NTK) with kno

Stochastic gradient descent with a large initial learning rate is widely used for training modern neural net architectures. Although a small initial learning rate allows for faster training and

NeurIPS Proceedings Search Self-Supervised Representation Learning on Neural Network Weights for Model Characteristic Prediction Konstantin Schürholt, Dimche Kostadinov, Damian Borth Advances in Neural Information Processing Systems 34 (NeurIPS 2021) Abstract Self-Supervised Learning (SSL) has been shown to learn useful and information-preserving representations. Neural Networks (NNs) are widely applied, yet their weight space is still not fully understood. Therefore, we propose to use SSL to learn hyper

Distributed Alignment Search: Identifying Causal Mechanisms Structure and Interpretation of Deep Networks Distributed Alignment Search: Identifying Causal Mechanisms October 31, 2024 • Lucas Laird The theory of causal abstraction provides a general framework for assessing the degree to which complex models (e.g. neural networks) implement a simpler interpretable causal model. The goal of causal abstraction is to identify high-level simple causal mechanisms that abstract away from the irrelevant low-level

Since people are talking about the NTK/GP hypothesis of neural nets again, I thought it might be worth bringing up some recent research in the area t