Sindy Löwe, Sindy Lowe, Sindy Loewe

Sindy Löwe

PhD Candidate Machine Learning

I am a PhD student in Machine Learning at the University of Amsterdam, supervised by Prof. Max Welling.
I am interested in self-supervised representation learning, focusing on local learning approaches and on structured representations.

Selected Publications

Rotating Features
Rotating Features for Object Discovery
   Oral Presentation
NeurIPS 2023
Sindy Löwe, Phillip Lippe, Francesco Locatello, Max Welling
We introduce advancements over the Complex AutoEncoder (below) to create continuous and distributed object-centric representations of real-world images without supervision.
Complex AutoEncoder
Complex-Valued Autoencoders for Object Discovery
TMLR 2022
Sindy Löwe, Phillip Lippe, Maja Rudolph, Max Welling
We present the Complex AutoEncoder – an object discovery approach that takes inspiration from neuroscience to implement distributed object-centric representations. After introducing complex-valued activations into a convolutional autoencoder, it learns to encode feature information in the activations’ magnitudes and object affiliation in their phase values.
Amortized Causal Discovery
Amortized Causal Discovery: Learning to Infer Causal Graphs from Time-Series Data
CLeaR 2022
Sindy Löwe*, David Madras*, Richard Zemel, Max Welling
We propose a new framework for causal discovery in which we learn to infer causal relations across samples with different underlying causal graphs but shared dynamics.
Object-Centric Video Models by Contrasting Sets
Learning Object-Centric Video Models by Contrasting Sets
NeurIPS 2020 Workshop: Object Representations for Learning and Reasoning
Sindy Löwe, Klaus Greff, Rico Jonschkowski, Alexey Dosovitskiy, Thomas Kipf
We highlight a fundamental problem in the loss used by previous approaches for contrastive object discovery: it cannot differentiate between a model that represents all the different objects in a scene and a model that (re-)represents the same object over and over. Our proposed solution: a set-based contrastive loss.
Greedy Infomax
Putting An End to End-to-End: Gradient-Isolated Learning of Representations
   Honorable Mention for the Outstanding New Directions Paper Award
NeurIPS 2019
Sindy Löwe*, Peter O'Connor, Bastiaan S. Veeling*
We show that we can train a neural network without end-to-end backpropagation and achieve competitive performance.
Greedy Infomax
Greedy InfoMax for Self-Supervised Representation Learning
   University of Amsterdam Thesis Award 2020
   KNVI/KIVI Thesis Prize for Informatics and Information Science 2020
Master's Thesis (2019)
Sindy Löwe
This thesis resulted in the above publication: "Putting An End to End-to-End: Gradient-Isolated Learning of Representations"
Defect Segmentation with Structural Similarity
Improving Unsupervised Defect Segmentation by Applying Structural Similarity to Autoencoders
Paul Bergmann, Sindy Löwe, Michael Fauser, David Sattlegger, Carsten Steger
We illustrate the shortcomings of pixel-wise reconstruction errors when using autoencoders for unsupervised defect segmentation on image data and propose to use a perceptual loss based on structural similarity.
Temporal Predictability in Visual Cortex
Temporal Predictability of Visual Target Onset by Audition Leads to Decrease in Evoked Neural Activity in Mouse V1
Extended Abstract at Neuroscience 2015
Sindy Löwe, Masataka Watanabe, Nikos Logothetis, Laura Busse, Steffen Katzner
We investigate how temporal predictability affects the processing of a visual stimulus by measuring the responses of single neurons in the mouse primary visual cortex (V1).


Sindy Löwe

loewe [dot] sindy [at] gmail [dot] com

AMLab, University of Amsterdam

Science Park 904

1098 XH Amsterdam

The Netherlands