1 code implementation • 25 Nov 2021 • Zhuoran Liu, Zhengyu Zhao, Alex Kolmus, Tijn Berns, Twan van Laarhoven, Tom Heskes, Martha Larson
Recent work has shown that imperceptible perturbations can be applied to craft unlearnable examples (ULEs), i. e. images whose content cannot be used to improve a classifier during training.
1 code implementation • 1 Nov 2021 • Alex Kolmus, Grégory Baltus, Justin Janquart, Twan van Laarhoven, Sarah Caudill, Tom Heskes
Fast, highly accurate, and reliable inference of the sky origin of gravitational waves would enable real-time multi-messenger astronomy.
1 code implementation • 6 Jul 2021 • Iordan Ganev, Twan van Laarhoven, Robin Walters
We introduce a class of fully-connected neural networks whose activation functions, rather than being pointwise, rescale feature vectors by a function depending only on their norm.
no code implementations • 8 Nov 2020 • Kai Chen, Twan van Laarhoven, Elena Marchiori
The heavy tail and skewness characteristics of such distributions in the spectral domain allow to capture long-range covariance of the signal in the time domain.
no code implementations • 6 May 2019 • Sebastiano Vascon, Sinem Aslan, Alessandro Torcinovich, Twan van Laarhoven, Elena Marchiori, Marcello Pelillo
Unsupervised domain adaptation (UDA) amounts to assigning class labels to the unlabeled instances of a dataset from a target domain, using labeled instances of a dataset from a related source domain.
no code implementations • 7 Aug 2018 • Kai Chen, Twan van Laarhoven, Perry Groot, Jinsong Chen, Elena Marchiori
The resulting kernel is called Multi-Output Convolution Spectral Mixture (MOCSM) kernel.
no code implementations • 3 Aug 2018 • Kai Chen, Twan van Laarhoven, Elena Marchiori, Feng Yin, Shuguang Cui
The function interaction is modeled by using cross convolution of latent functions.
1 code implementation • 12 Apr 2018 • Twan van Laarhoven
By assuming that the network is uniform, we can approximate the structure of unobserved parts of the network to obtain a method for local community detection.
no code implementations • 12 Apr 2018 • Jeroen Manders, Twan van Laarhoven, Elena Marchiori
Under the assumption that features from pre-trained deep neural networks are transferable across related domains, domain adaptation reduces to aligning source and target domain at class prediction uncertainty level.
1 code implementation • 20 Mar 2018 • Twan van Laarhoven, Elena Marchiori
Despite their success, state of the art methods based on this approach are either involved or unable to directly scale to data with many features.
1 code implementation • 15 Nov 2017 • Jacopo Acquarelli, Elena Marchiori, Lutgarde M. C. Buydens, Thanh Tran, Twan van Laarhoven
2) How is the performance of hyperspectral image classification methods affected when using disjoint train and test sets?
no code implementations • 14 Sep 2017 • Nastaran Mohammadian Rad, Seyed Mostafa Kia, Calogero Zarbo, Twan van Laarhoven, Giuseppe Jurman, Paola Venuti, Elena Marchiori, Cesare Furlanello
Our results show that: 1) feature learning outperforms handcrafted features; 2) parameter transfer learning is beneficial in longitudinal settings; 3) using LSTM to learn the temporal dynamic of signals enhances the detection rate especially for skewed training data; 4) an ensemble of LSTMs provides more accurate and stable detectors.
no code implementations • 16 Jun 2017 • Twan van Laarhoven
We show that popular optimization methods such as ADAM only partially eliminate the influence of normalization on the learning rate.
no code implementations • 16 Jun 2017 • Twan van Laarhoven, Elena Marchiori
Unsupervised Domain Adaptation (DA) is used to automatize the task of labeling data: an unlabeled dataset (target) is annotated using a labeled dataset (source) from a related domain.
no code implementations • 21 Jan 2016 • Twan van Laarhoven, Elena Marchiori
We investigate the relation of conductance with weighted kernel k-means for a single community, which leads to the introduction of a new objective function, $\sigma$-conductance.
no code implementations • 22 Jul 2014 • Twan van Laarhoven, Elena Marchiori
We argue that this is a desirable property, provide conditions under which NMF quality functions are local, and propose a novel class of local probabilistic NMF quality functions for soft graph clustering.
no code implementations • 15 Aug 2013 • Twan van Laarhoven, Elena Marchiori
This motivates the derivation of a new family of quality functions, adaptive scale modularity, which does satisfy the proposed axioms.