Real World Metamer Sets: Or How we Came to Love Noise

It is well known that color formation acts as a noise-reducing lossy compression mechanism that results in ambiguity, known as metamerism. Surfaces that match under one set of conditions-an illuminant and observer or capture device-can mismatch under others. The phenomenon has been studied extensively in the past, leading to important results like metamer mismatch volumes, color correction, reflectance estimation and the computation of metamer sets-sets of all possible reflectances that could result in a given sensor response. However, most of these approaches have three limitations: first, they simplify the problem and make assumptions about what reflectances can look like (i.e., being smooth, natural, residing in a subspace based on some measured data), second, they deal with strict mathematical metamerism and overlook noise or precision, and third, only isolated responses are considered without taking the context of a response into account. In this paper we address these limitations by outlining an approach that allows for the robust computation of approximate unconstrained metamer sets and exact unconstrained paramer sets. The notion of spatial or relational paramer sets that take neighboring responses into account, and applications to illuminant estimation and color constancy are also briefly discussed.

A Novel Robust Approach for Computing DE-9IM Matrices Based on Space Partition and Integer Coordinates

A novel approach for a robust computation of positional relations of two-dimensional geometric features is presented which guarantees reliable results, provided that the initial data is valid. The method is based on the use of integer coordinates and a method to generate a complete, gap-less and non-overlapping spatial decomposition. The spatial relationships of two geometric features are then represented using DE-9IM matrices. These allow the spatial relationships to be represented compactly. The DE-9IM matrices are based on the spatial decomposition using explicit neighborhood relations. No further geometric calculations are required for their computation. Based on comparative tests, it could be proven that this approach, up to a predictable limit, provides correct results and thus offers advantages over classical methods for the calculation of spatial relationships. This novel method can be used in all fields, especially where guaranteed reliable results are required.

Neuronal-Plasticity and Reward-Propagation Improved Recurrent Spiking Neural Networks

Different types of dynamics and plasticity principles found through natural neural networks have been well-applied on Spiking neural networks (SNNs) because of their biologically-plausible efficient and robust computations compared to their counterpart deep neural networks (DNNs). Here, we further propose a special Neuronal-plasticity and Reward-propagation improved Recurrent SNN (NRR-SNN). The historically-related adaptive threshold with two channels is highlighted as important neuronal plasticity for increasing the neuronal dynamics, and then global labels instead of errors are used as a reward for the paralleling gradient propagation. Besides, a recurrent loop with proper sparseness is designed for robust computation. Higher accuracy and stronger robust computation are achieved on two sequential datasets (i.e., TIDigits and TIMIT datasets), which to some extent, shows the power of the proposed NRR-SNN with biologically-plausible improvements.

Google Stock Movement

Human beings tend to make predictions about future events irrespective of probability of occurrence. We are fascinated to solve puzzles and patterns. One such area which intrigues many, full of complexity and unpredicted behavior, is the stock market. For the last decade or so, we have been trying to find patterns and understand the behavior of the stock market with the help of robust computation systems and new approaches to extract and analyze the huge amount of data. In this chapter, the authors have tried to understand stock price movement using a long short-term memory (LSTM) network and predict future behavior of stock price.