Engineering Commons^™

When Less Is More: How Increasing The Complexity Of Machine Learning Strategies For Geothermal Energy Assessments May Not Lead Toward Better Estimates, Stanley P. Mordensky, John Lipor, Jacob Deangelo, Erick R. Burns, Cary R. Lindsey

Electrical and Computer Engineering Faculty Publications and Presentations

Previous moderate- and high-temperature geothermal resource assessments of the western United States utilized data-driven methods and expert decisions to estimate resource favorability. Although expert decisions can add confidence to the modeling process by ensuring reasonable models are employed, expert decisions also introduce human and, thereby, model bias. This bias can present a source of error that reduces the predictive performance of the models and confidence in the resulting resource estimates.

Our study aims to develop robust data-driven methods with the goals of reducing bias and improving predictive ability. We present and compare nine favorability maps for geothermal resources in the …

Deep Learning Approach For Screening Autism Spectrum Disorder In Children With Facial Images And Analysis Of Ethnoracial Factors In Model Development And Application, Angelina Lu, Marek Perkowski

Electrical and Computer Engineering Faculty Publications and Presentations

Autism spectrum disorder (ASD) is a developmental disability that can cause significant social, communication, and behavioral challenges. Early intervention for children with ASD can help to improve their intellectual ability and reduces autistic symptoms. Multiple clinical researches have suggested that facial phenotypic differences exist between ASD children and typically developing (TD) children. In this research, we propose a practical ASD screening solution using facial images through applying VGG16 transfer learning-based deep learning to a unique ASD dataset of clinically diagnosed children that we collected. Our model produced a 95% classification accuracy and 0.95 F1-score. The only other reported study using …

Clustered Hyperspectral Target Detection, Sean Onufer Stalley

Dissertations and Theses

Aerial target detection is often used to search for relatively small things over large areas of land. Depending on the size and signature of the target, detection can be a very easy or very difficult task. By capturing images with several hundred color channels, hyperspectral sensors provide a new way of looking at this task, both literally and figuratively. Hyperspectral sensors can be used in many aerial target detection tasks such as identifying unhealthy trees in a forest, searching for minerals at a mining site, or finding the sources of chemical leaks at a factory. The high spectral resolution of …

Flight Simulator Modeling Using Recurrent Neural Networks, Nickolas Sabatini, Andreas Natsis

Undergraduate Research & Mentoring Program

Recurrent neural networks (RNNs) are a form of machine learning used to predict future values. This project uses RNNs tor predict future values for a flight simulator. Coded in Python using the Keras library, the model demonstrates training loss and validation loss, referring to the error when training the model.

An Application Of Deep Learning Models To Automate Food Waste Classification, Alejandro Zachary Espinoza

Dissertations and Theses

Food wastage is a problem that affects all demographics and regions of the world. Each year, approximately one-third of food produced for human consumption is thrown away. In an effort to track and reduce food waste in the commercial sector, some companies utilize third party devices which collect data to analyze individual contributions to the global problem. These devices track the type of food wasted (such as vegetables, fruit, boneless chicken, pasta) along with the weight. Some devices also allow the user to leave the food in a kitchen container while it is weighed, so the container weight must also …

Explanation Methods For Neural Networks, Jack H. Chen, Christof Teuscher

Student Research Symposium

Neural Networks (NNs) have become a basis of almost all state-of-the-art machine learning algorithms and classifiers. While NNs have been shown to generalize well to real-world examples, researchers have struggled to show why they work on an intuitive level. We designed several methods to explain the decisions of two state-of-the-art NN classifiers, ResNet and an All-CNN, in the context of the Japanese Society of Radiological Technology (JSRT) lung nodule dataset and the CIFAR-10 image dataset. Leading explanation methods LIME and Grad-CAM generate variations of heat maps which represent the regions of the input determined salient by the NN. We analyze …

Reliable Explanations Via Adversarial Examples On Robust Networks, Walt Woods, Jack H. Chen, Christof Teuscher

Student Research Symposium

Neural Networks (NNs) are increasingly used as the basis of advanced machine learning techniques in sensitive fields such as autonomous vehicles and medical imaging. However, NNs have been found vulnerable to a class of imperceptible attacks, called adversarial examples, which arbitrarily alter the output of the network. To close the schism between needing reliability in real-world applications and the fragility of NNs, we propose a new method for stabilizing networks, and show that as an added bonus, our technique results in reliable, high-fidelity explanations for the NN's decision. Compared to the state-of-the-art, this technique increased the area under the curve …

Spectral Clustering For Electrical Phase Identification Using Advanced Metering Infrastructure Voltage Time Series, Logan Blakely

Dissertations and Theses

The increasing demand for and prevalence of distributed energy resources (DER) such as solar power, electric vehicles, and energy storage, present a unique set of challenges for integration into a legacy power grid, and accurate models of the low-voltage distribution systems are critical for accurate simulations of DER. Accurate labeling of the phase connections for each customer in a utility model is one area of grid topology that is known to have errors and has implications for the safety, efficiency, and hosting capacity of a distribution system. This research presents a methodology for the phase identification of customers solely using …

From A Locally Competitive Algorithm To Sensory Relevance Models, Walter Woods

Electrical and Computer Engineering PhD Day

This poster addresses the development of a new Machine Learning (ML) mechanism, the Sensory Relevance Model (SRM), as a means of splitting information processing tasks into two sub-tasks with more intuitive properties. Specifically, SRMs are a front-end for other ML techniques, re-mapping the input data to a similar space with significantly more sparsity, achieved through the transformation and suppression of inputs irrelevant to the task. Prior work has attempted to reveal this information for Neural Networks (NNs) either as a post-processing step via saliency maps or through a simple masking of the input achieved with a dot product (so-called ``attention'' …

Hierarchical Random Boolean Network Reservoirs, Sai Kiran Cherupally

Dissertations and Theses

Reservoir Computing (RC) is an emerging Machine Learning (ML) paradigm. RC systems contain randomly assembled computing devices and can be trained to solve complex temporal tasks. These systems are computationally cheaper to train than other ML paradigms such as recurrent neural networks, and they can also be trained to solve multiple tasks simultaneously. Further, hierarchical RC systems with fixed topologies, were shown to outperform monolithic RC systems by up to 40% when solving temporal tasks. Although the performance of monolithic RC networks was shown to improve with increasing network size, building large monolithic networks may be challenging, for example because …

Lithography Hotspot Detection, Jea Woo Park

Dissertations and Theses

The lithography process for chip manufacturing has been playing a critical role in keeping Moor's law alive. Even though the wavelength used for the process is bigger than actual device feature size, which makes it difficult to transfer layout patterns from the mask to wafer, lithographers have developed a various technique such as Resolution Enhancement Techniques (RETs), Multi-patterning, and Optical Proximity Correction (OPC) to overcome the sub-wavelength lithography gap.

However, as feature size in chip design scales down further to a point where manufacturing constraints must be applied to early design phase before generating physical design layout. Design for Manufacturing …

Memory And Information Processing In Recurrent Neural Networks, Alireza Goudarzi, Sarah Marzen, Peter Banda, Guy Feldman, Matthew R. Lakin, Christof Teuscher, Darko Stefanovic

Electrical and Computer Engineering Faculty Publications and Presentations

Recurrent neural networks (RNN) are simple dynamical systems whose computational power has been attributed to their short-term memory. Short-term memory of RNNs has been previously studied analytically only for the case of orthogonal networks, and only under annealed approximation, and uncorrelated input. Here for the first time, we present an exact solution to the memory capacity and the task-solving performance as a function of the structure of a given network instance, enabling direct determination of the function-structure relation in RNNs. We calculate the memory capacity for arbitrary networks with exponentially correlated input and further related it to the performance of …

Training Set Design For Test Removal Classication In Ic Test, Nagarjun Hassan Ranganath

Dissertations and Theses

This thesis reports the performance of a simple classifier as a function of its training data set. The classifier is used to remove analog tests and is named the Test Removal Classifier (TRC).

The thesis proposes seven different training data set designs that vary by the number of wafers in the data set, the source of the wafers and the replacement scheme of the wafers. The training data set size ranges from a single wafer to a maximum of five wafers. Three of the training data sets include wafers from the Lot Under Test (LUT). The training wafers in the …

Reward-Driven Training Of Random Boolean Network Reservoirs For Model-Free Environments, Padmashri Gargesa

Dissertations and Theses

Reservoir Computing (RC) is an emerging machine learning paradigm where a fixed kernel, built from a randomly connected "reservoir" with sufficiently rich dynamics, is capable of expanding the problem space in a non-linear fashion to a higher dimensional feature space. These features can then be interpreted by a linear readout layer that is trained by a gradient descent method. In comparison to traditional neural networks, only the output layer needs to be trained, which leads to a significant computational advantage. In addition, the short term memory of the reservoir dynamics has the ability to transform a complex temporal input state …

Quantum Inductive Learning And Quantum Logic Synthesis, Martin Lukac

Dissertations and Theses

Since Quantum Computer is almost realizable on large scale and Quantum Technology is one of the main solutions to the Moore Limit, Quantum Logic Synthesis (QLS) has become a required theory and tool for designing Quantum Logic Circuits. However, despite its growth, there is no any unified aproach to QLS as Quantum Computing is still being discovered and novel applications are being identified.

The intent of this study is to experimentally explore principles of Quantum Logic Synthesis and its applications to Inductive Machine Learning. Based on algorithmic approach, I first design a Genetic Algorithm for Quantum Logic Synthesis that is …

Efficient Decomposition Of Large Fuzzy Functions And Relations, Paul Burkey, Marek Perkowski

Electrical and Computer Engineering Faculty Publications and Presentations

This paper presents a new approach to decomposition of fuzzy functions. A tutorial background on fuzzy logic representations is first given to emphasize next the simplicity and generality of this new approach. Ashenhurst-like decomposition of fuzzy functions was discussed in [3] but it was not suitable for programming and was not programmed. In our approach, fuzzy functions are converted to multiple-valued functions and decomposed using an mv decomposer. Then the decomposed multiple-valued functions are converted back to fuzzy functions. This approach allows for Curtis-like decompositions with arbitrary number of intermediate fuzzy variables, that have been not presented for fuzzy functions …

A New Approach To Robot’S Imitation Of Behaviors By Decomposition Of Multiple-Valued Relations, Uland Wong, Marek Perkowski

Electrical and Computer Engineering Faculty Publications and Presentations

Relation decomposition has been used for FPGA mapping, layout optimization, and data mining. Decision trees are very popular in data mining and robotics. We present relation decomposition as a new general-purpose machine learning method which generalizes the methods of inducing decision trees, decision diagrams and other structures. Relation decomposition can be used in robotics also in place of classical learning methods such as Reinforcement Learning or Artificial Neural Networks. This paper presents an approach to imitation learning based on decomposition. A Head/Hand robot learns simple behaviors using features extracted from computer vision, speech recognition and sensors.

Implicit Algorithms For Multi-Valued Input Support Manipulation, Alan Mishchenko, Craig Files, Marek Perkowski, Bernd Steinbach, Christina Dorotska

Electrical and Computer Engineering Faculty Publications and Presentations

We present an implicit approach to solve problems arising in decomposition of incompletely specified multi-valued functions and relations. We introduce a new representation based on binaryencoded multi-valued decision diagrams (BEMDDs). This representation shares desirable properties of MDDs, in particular, compactness, and is applicable to weakly-specified relations with a large number of output values. This makes our decomposition approach particularly useful for data mining and machine learning. Using BEMDDs to represent multi-valued relations we have developed two complementary input support minimization algorithms. The first algorithm is efficient when the resulting support contains almost all initial variables; the second is efficient when …

Decomposition Of Relations: A New Approach To Constructive Induction In Machine Learning And Data Mining -- An Overview, Marek Perkowski, Stanislaw Grygiel

Electrical and Computer Engineering Faculty Publications and Presentations

This is a review paper that presents work done at Portland State University and associated groups in years 1989 - 2001 in the area of functional decomposition of multivalued functions and relations, as well as some applications of these methods.

Constructive Induction Machines For Data Mining, Marek Perkowski, Stanislaw Grygiel, Qihong Chen, Dave Mattson

Electrical and Computer Engineering Faculty Publications and Presentations

"Learning Hardware" approach involves creating a computational network based on feedback from the environment (for instance, positive and negative examples from the trainer), and realizing this network in an array of Field Programmable Gate Arrays (FPGAs). Computational networks can be built based on incremental supervised learning (Neural Net training) or global construction (Decision Tree design). Here we advocate the approach to Learning Hardware based on Constructive Induction methods of Machine Learning (ML) using multivalued functions. This is contrasted with the Evolvable Hardware (EHW) approach in which learning/evolution is based on the genetic algorithm only.

Constructive Induction Machines For Data Mining, Marek Perkowski, Stanislaw Grygiel, Qihong Chen, Dave Mattson

Electrical and Computer Engineering Faculty Publications and Presentations

"Learning Hardware" approach involves creating a computational network based on feedback from the environment (for instance, positive and negative examples from the trainer), and realizing this network in an array of Field Programmable Gate Arrays (FPGAs). Computational networks can be built based on incremental supervised learning (Neural Net training) or global construction (Decision Tree design). Here we advocate the approach to Learning Hardware based on Constructive Induction methods of Machine Learning (ML) using multivalued functions. This is contrasted with the Evolvable Hardware (EHW) approach in which learning/evolution is based on the genetic algorithm only. Various approaches to supervised inductive learning …