Laurens van der Maaten
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Laurens van der Maaten has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Arash Atashpendar
Computer Science and Communications Research Unit, University of Luxembourg, Luxembourg, 1359 Luxembourg
1 shared publication Awni Hannun
1 shared publication- Brandon Duderstadt
Johns Hopkins University
1 shared publication Brian H. Wang
1 shared publication- Cedric Christian Laczny
Luxembourg Centre for Systems Biomedicine, University of Luxembourg
1 shared publication Chuan Guo
1 shared publication
Latest Publications
- Open Access
NOMAD Projection
The rapid adoption of generative AI has driven an explosion in the size of datasets consumed and produced by AI models. Traditional methods for unstructured data visualization, such as t-SNE and UMAP, have not kept up with the pace of dataset scaling. This presents a significant challenge for AI explainability, which relies on methods such as t-SNE and UMAP for exploratory data analysis. In this paper, we introduce Negative Or Mean Affinity Discrimination (NOMAD) Projection, the first method for unstructured data visualization via nonlinear dimensionality reduction that can run on multiple GPUs at train time. We provide theory that situates NOMAD Projection as an approximate upper bound on the InfoNC-t-SNE loss, and empirical results that demonstrate NOMAD Projection's superior performance and speed profile compared to existing state-of-the-art methods. We demonstrate the scalability of NOMAD Projection by computing the first complete data map of Multilingual Wikipedia.
Open AccessNOMAD Projection
The rapid adoption of generative AI has driven an explosion in the size of datasets consumed and produced by AI models. Traditional methods for unstructured data visualization, such as t-SNE and UMAP, have not kept up with the pace of dataset scaling. This presents a significant challenge for AI explainability, which relies on methods such as t-SNE and UMAP for exploratory data analysis. In this paper, we introduce Negative Or Mean Affinity Discrimination (NOMAD) Projection, the first method for unstructured data visualization via nonlinear dimensionality reduction that can run on multiple GPUs at train time. We provide theory that situates NOMAD Projection as an approximate upper bound on the InfoNC-t-SNE loss, and empirical results that demonstrate NOMAD Projection's superior performance and speed profile compared to existing state-of-the-art methods. We demonstrate the scalability of NOMAD Projection by computing the first complete data map of Multilingual Wikipedia.
- Open Access
Certified Data Removal from Machine Learning Models
Good data stewardship requires removal of data at the request of the data's owner. This raises the question if and how a trained machine-learning model, which implicitly stores information about its training data, should be affected by such a removal request. Is it possible to "remove" data from a machine-learning model? We study this problem by defining certified removal: a very strong theoretical guarantee that a model from which data is removed cannot be distinguished from a model that never observed the data to begin with. We develop a certified-removal mechanism for linear classifiers and empirically study learning settings in which this mechanism is practical.
Open AccessCertified Data Removal from Machine Learning Models
Good data stewardship requires removal of data at the request of the data's owner. This raises the question if and how a trained machine-learning model, which implicitly stores information about its training data, should be affected by such a removal request. Is it possible to "remove" data from a machine-learning model? We study this problem by defining certified removal: a very strong theoretical guarantee that a model from which data is removed cannot be distinguished from a model that never observed the data to begin with. We develop a certified-removal mechanism for linear classifiers and empirically study learning settings in which this mechanism is practical.
- Open Access
Anytime Stereo Image Depth Estimation on Mobile Devices
Many applications of stereo depth estimation in robotics require the generation of accurate disparity maps in real time under significant computational constraints. Current state-of-the-art algorithms force a choice between either generating accurate mappings at a slow pace, or quickly generating inaccurate ones, and additionally these methods typically require far too many parameters to be usable on power- or memory-constrained devices. Motivated by these shortcomings, we propose a novel approach for disparity prediction in the anytime setting. In contrast to prior work, our end-to-end learned approach can trade off computation and accuracy at inference time. Depth estimation is performed in stages, during which the model can be queried at any time to output its current best estimate. Our final model can process 1242$ \times $375 resolution images within a range of 10-35 FPS on an NVIDIA Jetson TX2 module with only marginal increases in error -- using two orders of magnitude fewer parameters than the most competitive baseline. The source code is available at https://github.com/mileyan/AnyNet .
Open AccessAnytime Stereo Image Depth Estimation on Mobile Devices
Many applications of stereo depth estimation in robotics require the generation of accurate disparity maps in real time under significant computational constraints. Current state-of-the-art algorithms force a choice between either generating accurate mappings at a slow pace, or quickly generating inaccurate ones, and additionally these methods typically require far too many parameters to be usable on power- or memory-constrained devices. Motivated by these shortcomings, we propose a novel approach for disparity prediction in the anytime setting. In contrast to prior work, our end-to-end learned approach can trade off computation and accuracy at inference time. Depth estimation is performed in stages, during which the model can be queried at any time to output its current best estimate. Our final model can process 1242$ \times $375 resolution images within a range of 10-35 FPS on an NVIDIA Jetson TX2 module with only marginal increases in error -- using two orders of magnitude fewer parameters than the most competitive baseline. The source code is available at https://github.com/mileyan/AnyNet .
- Top Cited PapersOpen Access
- 20 January 2015
- Published by Springer Nature in Microbiome
- Vol. 3 (1) , 1-7
VizBin - an application for reference-independent visualization and human-augmented binning of metagenomic data
Metagenomics is limited in its ability to link distinct microbial populations to genetic potential due to a current lack of representative isolate genome sequences. Reference-independent approaches, which exploit for example inherent genomic signatures for the clustering of metagenomic fragments (binning), offer the prospect to resolve and reconstruct population-level genomic complements without the need for prior knowledge. We present VizBin, a Java™-based application which offers efficient and intuitive reference-independent visualization of metagenomic datasets from single samples for subsequent human-in-the-loop inspection and binning. The method is based on nonlinear dimension reduction of genomic signatures and exploits the superior pattern recognition capabilities of the human eye-brain system for cluster identification and delineation. We demonstrate the general applicability of VizBin for the analysis of metagenomic sequence data by presenting results from two cellulolytic microbial communities and one human-borne microbial consortium. The superior performance of our application compared to other analogous metagenomic visualization and binning methods is also presented. VizBin can be applied de novo for the visualization and subsequent binning of metagenomic datasets from single samples, and it can be used for the post hoc inspection and refinement of automatically generated bins. Due to its computational efficiency, it can be run on common desktop machines and enables the analysis of complex metagenomic datasets in a matter of minutes. The software implementation is available at https://claczny.github.io/VizBin under the BSD License (four-clause) and runs under Microsoft Windows™, Apple Mac OS X™ (10.7 to 10.10), and Linux.
- 20 January 2015
- Published by Springer Nature in Microbiome
- Vol. 3 (1) , 1-7
Top Cited PapersOpen AccessVizBin - an application for reference-independent visualization and human-augmented binning of metagenomic data
Metagenomics is limited in its ability to link distinct microbial populations to genetic potential due to a current lack of representative isolate genome sequences. Reference-independent approaches, which exploit for example inherent genomic signatures for the clustering of metagenomic fragments (binning), offer the prospect to resolve and reconstruct population-level genomic complements without the need for prior knowledge. We present VizBin, a Java™-based application which offers efficient and intuitive reference-independent visualization of metagenomic datasets from single samples for subsequent human-in-the-loop inspection and binning. The method is based on nonlinear dimension reduction of genomic signatures and exploits the superior pattern recognition capabilities of the human eye-brain system for cluster identification and delineation. We demonstrate the general applicability of VizBin for the analysis of metagenomic sequence data by presenting results from two cellulolytic microbial communities and one human-borne microbial consortium. The superior performance of our application compared to other analogous metagenomic visualization and binning methods is also presented. VizBin can be applied de novo for the visualization and subsequent binning of metagenomic datasets from single samples, and it can be used for the post hoc inspection and refinement of automatically generated bins. Due to its computational efficiency, it can be run on common desktop machines and enables the analysis of complex metagenomic datasets in a matter of minutes. The software implementation is available at https://claczny.github.io/VizBin under the BSD License (four-clause) and runs under Microsoft Windows™, Apple Mac OS X™ (10.7 to 10.10), and Linux.
- Open Access
Marginalizing Corrupted Features
The goal of machine learning is to develop predictors that generalize well to test data. Ideally, this is achieved by training on an almost infinitely large training data set that captures all variations in the data distribution. In practical learning settings, however, we do not have infinite data and our predictors may overfit. Overfitting may be combatted, for example, by adding a regularizer to the training objective or by defining a prior over the model parameters and performing Bayesian inference. In this paper, we propose a third, alternative approach to combat overfitting: we extend the training set with infinitely many artificial training examples that are obtained by corrupting the original training data. We show that this approach is practical and efficient for a range of predictors and corruption models. Our approach, called marginalized corrupted features (MCF), trains robust predictors by minimizing the expected value of the loss function under the corruption model. We show empirically on a variety of data sets that MCF classifiers can be trained efficiently, may generalize substantially better to test data, and are also more robust to feature deletion at test time.
Open AccessMarginalizing Corrupted Features
The goal of machine learning is to develop predictors that generalize well to test data. Ideally, this is achieved by training on an almost infinitely large training data set that captures all variations in the data distribution. In practical learning settings, however, we do not have infinite data and our predictors may overfit. Overfitting may be combatted, for example, by adding a regularizer to the training objective or by defining a prior over the model parameters and performing Bayesian inference. In this paper, we propose a third, alternative approach to combat overfitting: we extend the training set with infinitely many artificial training examples that are obtained by corrupting the original training data. We show that this approach is practical and efficient for a range of predictors and corruption models. Our approach, called marginalized corrupted features (MCF), trains robust predictors by minimizing the expected value of the loss function under the corruption model. We show empirically on a variety of data sets that MCF classifiers can be trained efficiently, may generalize substantially better to test data, and are also more robust to feature deletion at test time.
- Open Access
Barnes-Hut-SNE
The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data objects, and it uses a variant of the Barnes-Hut algorithm - an algorithm used by astronomers to perform N-body simulations - to approximate the forces between the corresponding points in the embedding. Our experiments show that the new algorithm, called Barnes-Hut-SNE, leads to substantial computational advantages over standard t-SNE, and that it makes it possible to learn embeddings of data sets with millions of objects.
Open AccessBarnes-Hut-SNE
The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data objects, and it uses a variant of the Barnes-Hut algorithm - an algorithm used by astronomers to perform N-body simulations - to approximate the forces between the corresponding points in the embedding. Our experiments show that the new algorithm, called Barnes-Hut-SNE, leads to substantial computational advantages over standard t-SNE, and that it makes it possible to learn embeddings of data sets with millions of objects.
- 31 May 2012
- Published by Elsevier in Ecological Informatics
- Vol. 9, 1-10
Analyzing floristic inventories with multiple maps
- 31 May 2012
- Published by Elsevier in Ecological Informatics
- Vol. 9, 1-10
Analyzing floristic inventories with multiple maps
- Top Cited PapersOpen Access
- 17 December 2011
- Published by Springer Nature in Machine Learning
- Vol. 87 (1) , 33-55
Visualizing non-metric similarities in multiple maps
- 17 December 2011
- Published by Springer Nature in Machine Learning
- Vol. 87 (1) , 33-55
Top Cited PapersOpen AccessVisualizing non-metric similarities in multiple maps
- Open Access
- 12 October 2011
- Published by Springer Nature in Cognitive Processing
- Vol. 13 (S2) , 507-518
Action unit classification using active appearance models and conditional random fields
In this paper, we investigate to what extent modern computer vision and machine learning techniques can assist social psychology research by automatically recognizing facial expressions. To this end, we develop a system that automatically recognizes the action units defined in the facial action coding system (FACS). The system uses a sophisticated deformable template, which is known as the active appearance model, to model the appearance of faces. The model is used to identify the location of facial feature points, as well as to extract features from the face that are indicative of the action unit states. The detection of the presence of action units is performed by a time series classification model, the linear-chain conditional random field. We evaluate the performance of our system in experiments on a large data set of videos with posed and natural facial expressions. In the experiments, we compare the action units detected by our approach with annotations made by human FACS annotators. Our results show that the agreement between the system and human FACS annotators is higher than 90% and underlines the potential of modern computer vision and machine learning techniques to social psychology research. We conclude with some suggestions on how systems like ours can play an important role in research on social signals.
- 12 October 2011
- Published by Springer Nature in Cognitive Processing
- Vol. 13 (S2) , 507-518
Open AccessAction unit classification using active appearance models and conditional random fields
In this paper, we investigate to what extent modern computer vision and machine learning techniques can assist social psychology research by automatically recognizing facial expressions. To this end, we develop a system that automatically recognizes the action units defined in the facial action coding system (FACS). The system uses a sophisticated deformable template, which is known as the active appearance model, to model the appearance of faces. The model is used to identify the location of facial feature points, as well as to extract features from the face that are indicative of the action unit states. The detection of the presence of action units is performed by a time series classification model, the linear-chain conditional random field. We evaluate the performance of our system in experiments on a large data set of videos with posed and natural facial expressions. In the experiments, we compare the action units detected by our approach with annotations made by human FACS annotators. Our results show that the agreement between the system and human FACS annotators is higher than 90% and underlines the potential of modern computer vision and machine learning techniques to social psychology research. We conclude with some suggestions on how systems like ours can play an important role in research on social signals.
- 12 August 2011
- Published by Springer Nature in Vegetation History and Archaeobotany
- Vol. 20 (5) , 435-445
Identification of Myosotis seeds by means of digital image analysis
- 12 August 2011
- Published by Springer Nature in Vegetation History and Archaeobotany
- Vol. 20 (5) , 435-445
Identification of Myosotis seeds by means of digital image analysis