We present a new technique called “t-SNE” that visualizes high-dimensional data by giving each datapoint a location in a two or three-dimensional map. The technique is a variation of Stochastic Neighbor Embedding (Hinton and Roweis, 2002) that is much easier to optimize, and produces significantly better visualizations by reducing the tendency to crowd points together in the center of the map. t-SNE is better than existing techniques at creating a single map that reveals structure at many different scales. This is particularly important for high-dimensional data that lie on several different, but related, low-dimensional manifolds, such as images of objects from multiple classes seen from multiple viewpoints. For visualizing the structure of very large datasets, we show how t-SNE can use random walks on neighborhood graphs to allow the implicit structure of all of the data to influence the way in which a subset of the data is displayed. We illustrate the performance of t-SNE on a wide variety of datasets and compare it with many other non-parametric visualization techniques, including Sammon mapping, Isomap, and Locally Linear Embedding. The visualizations produced by t-SNE are significantly better than those produced by the other techniques on almost all of the datasets.

/pdf/visualizing-data-using-t-sne-2w1n304iq0.pdf

Visualizing Data using t-SNE

Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

https://axon.cs.byu.edu/~martinez/classes/678/Presentations/Yao.pdf

Extreme learning machine: Theory and applications

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining - Concepts and Techniques.

State-space models offer a powerful modelling tool for time series prediction. However, as most algorithms are not optimized for long-term prediction, it may be hard to achieve good prediction results. In this paper, we investigate Gaussian linear regression filters for parameter estimation in state-space models and we propose new long-term prediction strategies. Experiments using the EM-algorithm for training of nonlinear state-space models show that significant improvements are possible with no additional computational cost.

/pdf/long-term-prediction-of-time-series-using-state-space-models-1fbky4z93i.pdf

Long-Term prediction of time series using state-space models

Underwater object tracking is one of the most essential and fundamental tasks in ocean investigations recent years. In this paper, we try to capture multi-scale retinex (MSR) model as well as the partial least square (PLS) analysis for underwater object tracking. We first make use of multi-scale retinex model to evolve and enhance the partial color constancy from the underwater video sequences, which could provide a versatile automatic strategy to simultaneous sharpening, dynamic range compression and color rendition. The partial least square analysis is further taken to capture the trajectories of underwater objects by learning a set of underwater appearance models for adaptive discriminative object representation. The proposed object tracking algorithm exploits both the ground truth appearance information of the labeled underwater object in the first frame and the image sequences observed online, thereby alleviating the tracking drift problem caused by modeling update.

Underwater object tracking strategy via multi-scale retinex and partial least squares analysis

Steganalysis consists in classifying documents as steganographied or genuine. This paper presents a methodology for steganalysis based on a set of 193 features, with two main goals: determine a sufficient number of images for effective training of a classifier in the obtained high-dimensional space, and use feature selection to select most relevant features for the desired classification. Dimensionality reduction is performed using a forward selection and reduces the original 193 features set by a factor of 13, with overall same performance.

/pdf/avantages-de-la-selection-de-caracteristiques-pour-la-111o2euu2u.pdf

Avantages de la Sélection de Caractéristiques pour la Stéganalyse

Statistical fault isolation with PCA

The Minimal Learning Machine (MLM) has been recently proposed as a novel supervised learning method for regression problems aiming at reconstructing the mapping between input and output distance matrices Estimation of the response is then achieved from the geometrical configuration of the output points Thanks to its comprehensive formulation, the MLM is inherently capable of dealing with nonlinear problems and multidimensional output spaces In this paper, we introduce an extension of the MLM to classification tasks, thus providing a unified framework for multiresponse regression and classification problems On the basis of our experiments, the MLM achieves results that are comparable to many de facto standard methods for classification with the advantage of offering a computationally lighter alternative to such approaches

Amaury Lendasse

Papers

Long-Term prediction of time series using state-space models

Underwater object tracking strategy via multi-scale retinex and partial least squares analysis

Avantages de la Sélection de Caractéristiques pour la Stéganalyse

Statistical fault isolation with PCA

Extending the Minimal Learning Machine for Pattern Classification