Leon Palafox

Reverse Engineering of Gene Regulatory Networks Using Dissipative Particle Swarm Optimization

Proteins are composed by amino acids, which are created by genes. To understand how different genes interact to create different proteins, we need to model the gene regulatory networks (GRNs) of different organisms. There are different models that attempt to model GRNs. In this paper, we use the popular S-System to model small networks. This model has been solved with different evolutionary computation techniques, which have obtained good results; yet, there are no models that achieve a perfect reconstruction of the network. We implement a variation of particle swarm optimization (PSO), called dissipative PSO (DPSO), to optimize the model; we also research the use of an L1 regularizer and compare it with other evolutionary computing approaches. To the best of our knowledge, neither the DPSO nor L1 optimizer has been jointly used to solve the S-System. We find that the combination of S-System and DPSO offers advantages over previously used methods, and presents promising results for inferencing larger and more complex networks.

Reconstruction of Gene Regulatory Networks from Gene Expression Data Using Decoupled Recurrent Neural Network Model

In this work we used the decoupled version of the recurrent neural network (RNN) model for gene network inference from gene expression data. In the decoupled version, the global problem of estimating the full set of parameters for the complete network is divided into several sub-problems each of which corresponds to estimating the parameters associated with a single gene. Thus, the decoupling of the model decreases the problem dimensionality and makes the reconstruction of larger networks more feasible from the point of algorithmic perspective. We applied a well established evolutionary algorithm called differential evolution for inferring the underlying network structure as well as the regulatory parameters. We investigated the effectiveness of the reconstruction mechanism in analyzing the gene expression data collected from both synthetic and real gene networks. The proposed method was successful in inferring important gene interactions from expression profiles.

Multi-objective portfolio optimization and rebalancing using genetic algorithms with local search

The Portfolio Optimization problem is an example of a resource allocation problem with money as the resource to be allocated to assets. We first have to select the assets from a pool of them available in the market and then assign proper weights to them to maximize the return and minimize the risk associated with the Portfolio. In our work, we have introduced a new “greedy coordinate ascent mutation operator” and we have also included the trading volumes concept. We performed simulations with the past data of NASDAQ100 and DowJones30, concentrating mainly on the 2008 recession period. We also compared our results with the indices and the simple Genetic Algorithms approach.

Automated detection of geological landforms on Mars using Convolutional Neural Networks

The large volume of high-resolution images acquired by the Mars Reconnaissance Orbiter has opened a new frontier for developing automated approaches to detecting landforms on the surface of Mars. However, most landform classifiers focus on crater detection, which represents only one of many geological landforms of scientific interest. In this work, we use Convolutional Neural Networks (ConvNets) to detect both volcanic rootless cones and transverse aeolian ridges. Our system, named MarsNet, consists of five networks, each of which is trained to detect landforms of different sizes. We compare our detection algorithm with a widely used method for image recognition, Support Vector Machines (SVMs) using Histogram of Oriented Gradients (HOG) features. We show that ConvNets can detect a wide range of landforms and has better accuracy and recall in testing data than traditional classifiers based on SVMs.

On the use of Population Based Incremental Learning to do Reverse Engineering on Gene Regulatory Networks

Gene Regulatory Networks (GRNs) describe the interactions between different genes. One of the most important tasks in biology is to find the right regulations in a GRN given observed data. The problem, is that the data is often noisy and scarce, and we have to use models robust to noise and scalable to hundreds of genes. Recently, Recursive Neural Networks (RNNs) have been presented as a viable model for GRNs, which is robust to noise and can be scaled to larger networks. In this paper, to optimize the parameters of the RNN, we implement a classic Population Based Incremental Learning (PBIL), which in certain scenarios has outperformed classic GA and other evolutionary techniques like Particle Swarm Optimization (PSO). We test this implementation on a small and a large artificial networks. We further study the optimal tunning parameters and discuss the advantages of the method.

Human action recognition using 4W1H and Particle Swarm Optimization Clustering

Tracking and recording human activities have been a major interest in the iSpace, for this purpose different recognition and clustering techniques have been used, like using a Learning Classifier System and data Mining Techniques. These techniques share the common factor of database dependence and there was actually little effort into making the system to understand the way human were behaving in a given time in the space. Using Artificial Intelligence techniques, we present a work that reads and classifies user object activity

Evolving Genetic Networks for Synthetic Biology

The sibling disciplines, systems and synthetic biology, are engaged in unraveling the complexity of the biological networks. One is trying to understand the design principle of the existing networks while the other is trying to engineer artificial gene networks with predicted functions. The significant and important role that computational intelligence can play to steer the life engineering discipline towards its ultimate goal, has been acknowledged since its time of birth. However, as the field is facing many challenges in building complex modules/systems from the simpler parts/devices, whether from scratch or through redesign, the role of computational assistance becomes even more crucial. Evolutionary computation, falling under the broader domain of artificial intelligence, is well-acknowledged for its near optimal solution seeking capability for poorly known and partially understood problems. Since the post genome period, these natural-selection simulating algorithms are playing a noteworthy role in identifying, analyzing and optimizing different types of biological networks. This article calls attention to how evolutionary computation can help synthetic biologists in assembling larger network systems from the lego-like parts.

Human action recognition using wavelet signal analysis as an input in 4W1H

The human action recognition problem, for real time implementation have always been part of the research interest in the iSpace. Different techniques have been used in approaching a real time sensing and processing of the information to be able to deliver feedback to the user as well to monitoring systems. In this paper we apply wavelet processing techniques to solve the problem of real time processing, as well as ti filter the original signal in order to have better classification.

A movement profile detection system using self organized maps in the Intelligent Space

Detecting human activities in a controlled room such as Hashimoto Laboratory Intelligent Space is an important part in researching the interaction between robots and humans within the space. A number of algorithms is required in order to detect and classify effectively human activity. In this paper we implement a self organized map structure to data obtained from a sensor network in order to classify human activity in the space for a latter classification and use of it.

Gene regulatory network reverse engineering using population based incremental learning and K-means

Finding interactions among genes is one of the main problems in molecular biology. In this paper, we use a novel approach to model the genes regulations, or Gene Regulatory Networks (GRNs). We use a Recursive Neural Network (RNN) to model the networks, and then use Population Based Incremental Learning (PBIL) enhanced with K-means to find the optimum parameters of the Neural Network. We present experiments with real data, we compare our algorithm with others approaches by calculating different statistics for the solutions.