Výsledky vyhledávání - "Herrmann, Jan"

Report

A Machine Learning and Explainable AI Framework Tailored for Unbalanced Experimental Catalyst Discovery

Autor: Semnani, Parastoo, Bogojeski, Mihail, Bley, Florian, Zhang, Zizheng, Wu, Qiong, Kneib, Thomas, Herrmann, Jan, Weisser, Christoph, Patcas, Florina, Müller, Klaus-Robert

The successful application of machine learning (ML) in catalyst design relies on high-quality and diverse data to ensure effective generalization to novel compositions, thereby aiding in catalyst discovery. However, due to complex interactions, catal

Externí odkaz: http://arxiv.org/abs/2407.18935

Zobrazit plný text záznamu

Conference

Use-Case Studie eines auf der Nutzung von Handlungsfehlern basierenden AR-Lernsystems zur kritischen Reflexion der technischen Umsetzbarkeit

Autor: Kobelt, Dennis, Goppold, Marvin, Atanasyan, Alexander, Herrmann, Jan-Phillip, Tackenberg, Sven, Frenz, Martin, Gamber, Thilo

Empirische Untersuchungen haben gezeigt, dass das Verursachen von Fehlern und das Erleben von Fehlerkonsequenzen eine effektive Lerngelegenheit darstellen (Harteis, Bauer & Gruber, 2008, Kapur, 2015). Daher wird ein Lernsystem für die gewerblichtech

Externí odkaz: https://tud.qucosa.de/id/qucosa%3A73599
https://tud.qucosa.de/api/qucosa%3A73599/attachment/ATT-0/

Zobrazit plný text záznamu

Report

Towards Fixing Clever-Hans Predictors with Counterfactual Knowledge Distillation

Autor: Bender, Sidney, Anders, Christopher J., Chormai, Pattarawatt, Marxfeld, Heike, Herrmann, Jan, Montavon, Grégoire

This paper introduces a novel technique called counterfactual knowledge distillation (CFKD) to detect and remove reliance on confounders in deep learning models with the help of human expert feedback. Confounders are spurious features that models ten

Externí odkaz: http://arxiv.org/abs/2310.01011

Zobrazit plný text záznamu

Report

Disentangled Explanations of Neural Network Predictions by Finding Relevant Subspaces

Autor: Chormai, Pattarawat, Herrmann, Jan, Müller, Klaus-Robert, Montavon, Grégoire

Explainable AI aims to overcome the black-box nature of complex ML models like neural networks by generating explanations for their predictions. Explanations often take the form of a heatmap identifying input features (e.g. pixels) that are relevant

Externí odkaz: http://arxiv.org/abs/2212.14855

Zobrazit plný text záznamu