A5088912607- Consumer Behavior in Brand Consumption and Identification
- Muscle activation and electromyography studies
- EEG and Brain-Computer Interfaces
- Digital Marketing and Social Media
- Parallel Computing and Optimization Techniques
- Motor Control and Adaptation
- Transcranial Magnetic Stimulation Studies
- Embedded Systems Design Techniques
- Stroke Rehabilitation and Recovery
- Cultural Differences and Values
- Neuroscience and Neural Engineering
- Customer Service Quality and Loyalty
- Visual and Cognitive Learning Processes
- Real-Time Systems Scheduling
- Memory Processes and Influences
- Functional Brain Connectivity Studies
- Neural dynamics and brain function
- Model Reduction and Neural Networks
- Urinary Bladder and Prostate Research
- Behavioral Health and Interventions
- Meteorological Phenomena and Simulations
- Neural and Behavioral Psychology Studies
- Advanced Neural Network Applications
- Media Influence and Health
- Digital Communication and Language
University of Houston
2019-2025
University of Miami
2024-2025
St. Jude Children's Research Hospital
2024-2025
University of Minnesota
2002-2019
Nvidia (United States)
2018-2019
Nvidia (United Kingdom)
2018-2019
Oklahoma State University
2019
University of Minnesota System
1987-2014
Rapita Systems (United Kingdom)
2011-2013
Advanced Micro Devices (Canada)
2010-2011
The graphics processing unit (GPU) has become an integral part of today's mainstream computing systems. Over the past six years, there been a marked increase in performance and capabilities GPUs. modern GPU is not only powerful engine but also highly parallel programmable processor featuring peak arithmetic memory bandwidth that substantially outpaces its CPU counterpart. GPU's rapid both programmability capability spawned research community successfully mapped broad range computationally...
Deep neural networks have enabled progress in a wide variety of applications. Growing the size network typically results improved accuracy. As model sizes grow, memory and compute requirements for training these models also increases. We introduce technique to train deep using half precision floating point numbers. In our technique, weights, activations gradients are stored IEEE half-precision format. Half-precision numbers limited numerical range compared single-precision propose two...
Based on three explanations of imagery effects memory, hypotheses regarding the conditions under which pictorial ads are or not remembered better than verbal-only generated and tested. The memorability brand names semantically related to product class was tested in versus form various conditions. results indicate that picture superiority occurs both immediate delayed recall tasks when processing is directed at appearance features. Verbal-only stimuli recalled as well pictures but become...
Journal Article Measurement of Individual Differences in Visual Versus Verbal Information Processing Get access Terry L. Childers, Childers Search for other works by this author on: Oxford Academic PubMed Google Scholar Michael J. Houston, Houston Susan E. Heckler Consumer Research, Volume 12, Issue 2, September 1985, Pages 125–134, https://doi.org/10.1086/208501 Published: 01 1985 history Received: 1984 Revision received: January
We extract pixel-level masks of extreme weather patterns using variants Tiramisu and DeepLabv3+ neural networks. describe improvements to the software frameworks, input pipeline, network training algorithms necessary efficiently scale deep learning on Piz Daint Summit systems. The scales 5300 P100 GPUs with a sustained throughput 21.0 PF/s parallel efficiency 79.0%. up 27360 V100 325.8 90.7% in single precision. By taking advantage FP16 Tensor Cores, half-precision version achieves peak 1.13...
The rigorous application of static timing analysis requires a large and costly amount detail knowledge on the hardware software components system. Probabilistic Timing Analysis has potential for reducing weight that demand. In this paper, we present sound measurement-based probabilistic technique based Extreme Value Theory. all experiments made as part work, bounds determined by our were less than 15% pessimistic in comparison with tightest possible obtainable any technique. As point...
The authors report the results of a laboratory experiment examining effects meaningfulness brand names on recall advertising. findings indicate that name explicitly conveying product benefit (e.g., PicturePerfect televisions) leads to higher an advertised claim consistent in meaning with compared nonsuggestive Emporium televisions). Conversely, suggestive lower subsequently unrelated superior sound) name. discuss implications these for marketers respect advertising strategies and optimal use...
Many advertisers believe the pictorial and verbal components of an ad should convey same meaning. Based on theoretical empirical evidence from a variety areas, three experiments were con...
This research investigates the learning that occurs throughout several information acquisition and choice experiences. The effects of three factors may naturally vary in consumer experiences are studied: a consumer's goals, how much knows about product's features prior to choice, content feedback received after choice. Results show consumers learn is organized memory around goal(s) drives Further, higher levels feature knowledge result more accurate experience, but, contrary predictions,...
The Merasa project aims to achieve a breakthrough in hardware design, hard real-time support system software, and worst-case execution time analysis tools for embedded multicore processors. focuses on developing processor designs systems techniques guarantee the analyzability timing predictability of every feature provided by processor.
Static timing analysis is the state-of-the-art practice of ascertaining behavior current-generation real-time embedded systems. The adoption more complex hardware to respond increasing demand for computing power in next-generation systems exacerbates some limitations static analysis. In particular, effort acquiring (1) detailed information on develop an accurate model its execution latency as well (2) knowledge program presence varying conditions, such those dependent history previously...
The authors report the results of a laboratory experiment examining effects meaningfulness brand names on recall advertising. findings indicate that name explicitly conveying product benefit (e.g., PicturePerfect televisions) leads to higher an advertised claim consistent in meaning with compared nonsuggestive Emporium televisions). Conversely, suggestive lower subsequently unrelated superior sound) name. discuss implications these for marketers respect advertising strategies and optimal use...
Journal Article Exemplars or Beliefs? The Impact of Self-View on the Nature and Relative Influence Brand Associations Get access Sharon Ng, Ng Search for other works by this author on: Oxford Academic PubMed Google Scholar Michael J. Houston Consumer Research, Volume 32, Issue 4, March 2006, Pages 519–529, https://doi.org/10.1086/500482 Published: 01 2006
Modern processors are evolving into hybrid, heterogeneous with both CPU and GPU cores used for general purpose computation. Several languages such as Brook, CUDA, more recently OpenCL being developed to fully harness the potential of these processors. These typically involve control code running on performance-critical, data-parallel kernel GPUs.
We extract pixel-level masks of extreme weather patterns using variants Tiramisu and DeepLabv3+ neural networks. describe improvements to the software frameworks, input pipeline, network training algorithms necessary efficiently scale deep learning on Piz Daint Summit systems. The scales 5300 P100 GPUs with a sustained throughput 21.0 PF/s parallel efficiency 79.0%. up 27360 V100 325.8 90.7% in single precision. By taking advantage FP16 Tensor Cores, half-precision version achieves peak 1.13...
Hybrid brain computer interfaces (BCI) utilizing the high temporal resolution of electroencephalography (EEG) and spatial near-infrared spectroscopy (fNIRS) are preferred over single-modal BCIs. However, due to large dimensionality multi-class statistical features commonly used in fNIRS signals, it is easy cause overfitting EEG-fNIRS hybrid BCI classifier. Therefore, a low-dimensional feature extraction method for based on EEG-informed general linear model (GLM) analysis proposed this paper....
The proliferation of biological sequence data has motivated the need for an extremely fast probabilistic search. One method performing this search involves evaluating Viterbi probability a hidden Markov model (HMM) desired family each in protein database. However, one difficulties with current implementations is time required to large databases. Many and upcoming architectures offering amounts compute power are designed data-parallel execution streaming mind. We present algorithm HMM's...
Segmentation of structures from measured volume data, such as anatomy in medical imaging, is a challenging data-dependent task. In this paper, we present segmentation method that leverages the parallel processing capabilities modern programmable graphics hardware order to run significantly faster than previous methods. addition, collocating algorithm computation with visualization on circumvents need transfer data across system bus, allowing for and interaction. This unique it utilizes...
Many advertisers believe the pictorial and verbal components of an ad should convey same meaning. Based on theoretical empirical evidence from a variety areas, three experiments were conducted that show superior recall for ads in which picture copy discrepant information about product attributes when brand name are linked interactively. An elaborative processing explanation effect is supported by finding this superiority diminishes if consumers have less opportunity to process form...
Accurate anatomical matching for patient-specific electromyographic (EMG) mapping is crucial yet technically challenging in various medical disciplines. The fixed electrode construction of multielectrode arrays (MEAs) makes it nearly impossible to match an individual's unique muscle anatomy. This mismatch between the MEAs and target muscles leads missing relevant activity, highly redundant data, complicated placement optimization, inaccuracies classification algorithms. Here, we present...