Sökning: onr:"swepub:oai:DiVA.org:liu-169431" >
Classifying the cla...
Classifying the classifier : dissecting the weight space of neural networks
-
- Eilertsen, Gabriel, 1984- (författare)
- Linköpings universitet,Medie- och Informationsteknik,Tekniska fakulteten
-
- Jönsson, Daniel, 1984- (författare)
- Linköpings universitet,Medie- och Informationsteknik,Tekniska fakulteten
-
- Ropinski, Timo (författare)
- Institute of Media Informatics, Ulm University, Germany
-
visa fler...
-
- Unger, Jonas, 1978- (författare)
- Linköpings universitet,Medie- och Informationsteknik,Tekniska fakulteten
-
- Ynnerman, Anders, 1963- (författare)
- Linköpings universitet,Medie- och Informationsteknik,Tekniska fakulteten
-
visa färre...
-
(creator_code:org_t)
- IOS PRESS, 2020
- 2020
- Engelska.
-
Ingår i: Proceedings of the 24th European Conference on Artificial Intelligence (ECAI 2020). - : IOS PRESS. - 9781643681016 ; , s. 1119-1126
- Relaterad länk:
-
http://ebooks.iospre...
-
visa fler...
-
https://urn.kb.se/re...
-
https://doi.org/10.3...
-
visa färre...
Abstract
Ämnesord
Stäng
- This paper presents an empirical study on the weights of neural networks, where we interpret each model as a point in a high-dimensional space – the neural weight space. To explore the complex structure of this space, we sample from a diverse selection of training variations (dataset, optimization procedure, architecture,etc.) of neural network classifiers, and train a large number of models to represent the weight space. Then, we use a machine learning approach for analyzing and extracting information from this space. Most centrally, we train a number of novel deep meta-classifiers withthe objective of classifying different properties of the training setup by identifying their footprints in the weight space. Thus, the meta-classifiers probe for patterns induced by hyper-parameters, so that we can quantify how much, where, and when these are encoded through the optimization process. This provides a novel and complementary view for explainable AI, and we show how meta-classifiers can reveal a great deal of information about the training setup and optimization, by only considering a small subset of randomly selected consecutive weights. To promote further research on the weight space, we release the neural weight space (NWS) dataset – a collection of 320K weightsnapshots from 16K individually trained deep neural networks.
Ämnesord
- NATURVETENSKAP -- Data- och informationsvetenskap -- Datorseende och robotik (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Computer Vision and Robotics (hsv//eng)
- NATURVETENSKAP -- Data- och informationsvetenskap -- Datavetenskap (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Computer Sciences (hsv//eng)
Nyckelord
- machine learning
- deep learning
- ai
- computer vision
Publikations- och innehållstyp
- ref (ämneskategori)
- kon (ämneskategori)
Hitta via bibliotek
Till lärosätets databas