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

From the Publisher: 
The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

The paper investigates the acceleration of t-SNE--an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots--using two tree-based algorithms. In particular, the paper develops variants of the Barnes-Hut algorithm and of the dual-tree algorithm that approximate the gradient used for learning t-SNE embeddings in O(N log N). Our experiments show that the resulting algorithms substantially accelerate t-SNE, and that they make it possible to learn embeddings of data sets with millions of objects. Somewhat counterintuitively, the Barnes-Hut variant of t-SNE appears to outperform the dual-tree variant.

Accelerating t-SNE using tree-based algorithms

New high-dimensional, single-cell technologies offer unprecedented resolution in the analysis of heterogeneous tissues. However, because these technologies can measure dozens of parameters simultaneously in individual cells, data interpretation can be challenging. Here we present viSNE, a tool that allows one to map high-dimensional cytometry data onto two dimensions, yet conserve the high-dimensional structure of the data. viSNE plots individual cells in a visual similar to a scatter plot, while using all pairwise distances in high dimension to determine each cell's location in the plot. We integrated mass cytometry with viSNE to map healthy and cancerous bone marrow samples. Healthy bone marrow automatically maps into a consistent shape, whereas leukemia samples map into malformed shapes that are distinct from healthy bone marrow and from each other. We also use viSNE and mass cytometry to compare leukemia diagnosis and relapse samples, and to identify a rare leukemia population reminiscent of minimal residual disease. viSNE can be applied to any multi-dimensional single-cell technology.

viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia

Sparsity preserving projections with applications to face recognition

In recent years, a variety of nonlinear dimensionality reduction techniques have been proposed that aim to address the limitations of traditional techniques such as PCA and classical scaling. The paper presents a review and systematic comparison of these techniques. The performances of the nonlinear techniques are investigated on artificial and natural tasks. The results of the experiments reveal that nonlinear techniques perform well on selected artificial tasks, but that this strong performance does not necessarily extend to real-world tasks. The paper explains these results by identifying weaknesses of current nonlinear techniques, and suggests how the performance of nonlinear dimensionality reduction techniques may be improved.

/pdf/dimensionality-reduction-a-comparative-review-4vhn8oy7jr.pdf

Dimensionality Reduction: A Comparative Review

In this paper we investigate whether a personality profile can be determined by observing a player's behavior in a game. Five personality traits are used to define a personality profile. They are adopted from the Five Factor Model of personality. The five traits are measured by the NEO-PI-R questionnaire. For our purpose, we developed a game module for the game Neverwinter Nights. The module automatically stores a player's behavioral data. Experimental trials were run measuring the behavior of 44 participants. The experiment produced game behavior scores for 275 game variables per player. Correlation analysis shows relationships between all five personality traits and the video game data. From these results, we may conclude that a video game can be used to create an adequate personality profile of a player.

Games as personality profiling tools

Computer analysis of Van Gogh's complementary colours

In many applications of computer algebra large expressions must be simplified to make repeated numerical evaluations tractable. Previous works presented heuristically guided improvements, e.g., for Horner schemes. The remaining expression is then further reduced by common subexpression elimination. A recent approach successfully applied a relatively new algorithm, Monte Carlo Tree Search (MCTS) with UCT as the selection criterion, to find better variable orderings. Yet, this approach is fit for further improvements since it is sensitive to the so-called ``exploration-exploitation'' constant $C_p$ and the number of tree updates $N$. In this paper we propose a new selection criterion called Simulated Annealing UCT (SA-UCT) that has a \emph{dynamic} exploration-exploitation parameter, which decreases with the iteration number $i$ and thus reduces the importance of exploration over time. First, we provide an intuitive explanation in terms of the exploration-exploitation behavior of the algorithm. Then, we test our algorithm on three large expressions of different origins. We observe that SA-UCT widens the interval of good initial values $C_p$ where best results are achieved. The improvement is large (more than a tenfold) and facilitates the selection of an appropriate $C_p$.

Combining Simulated Annealing and Monte Carlo Tree Search for Expression Simplification

The goal of adaptive game AI is to enhance computer-controlled game-playing agents with (1) the ability to self-correct mistakes, and (2) creativity in responding to new situations. Dynamic scripting is a reinforcement learning technique that realises fast and reliable online adaptation of game AI. It employs knowledge bases which contain rules that can be included in game scripts. To be successful, dynamic scripting requires a mechanism to order the rules that are selected for scripts. So far, rule ordering was achieved by a manually-tuned priority value for each rule. In the present research, we propose three mechanisms to order rules automatically for dynamic scripting. We performed experiments in which we let dynamic scripting, using each of the three mechanisms, play against manually-designed tactics. Our results show that dynamic scripting with automatic rule ordering generates game AI that is at least as effective as dynamic scripting with manually-tuned priority values. Moreover, it has the ability to generate novel game AI with significantly increased effectiveness. The costs are a slight decrease in learning efficiency. So, we may conclude that automatic rule ordering is a valuable enhancement for dynamic scripting.

/pdf/automatic-rule-ordering-for-dynamic-scripting-3ax9ouqo14.pdf

Jaap van den Herik

Papers

Dimensionality Reduction: A Comparative Review

Games as personality profiling tools

Computer analysis of Van Gogh's complementary colours

Combining Simulated Annealing and Monte Carlo Tree Search for Expression Simplification

Automatic rule ordering for dynamic scripting