Computational Foundations of Natural Intelligence

Gerven, Marcel van

doi:10.1101/166785

Cited by 6 publications

(4 citation statements)

References 280 publications

(260 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nevertheless, DNNs are of increasing interest for natural vision (Cox and Dean, 2014;Khaligh-Razavi and Kriegeskorte, 2014;Yamins et al, 2014;Güçlü and van Gerven, 2015b, a;Kriegeskorte, 2015;Cichy et al, 2016;Wen et al, 2016Wen et al, , 2017Eickenberg et al, 2017;Horikawa and Kamitani, 2017;Seeliger et al, 2017). Recent studies have shown that DNNs, especially convolutional neural networks for image recognition (Krizhevsky et al, 2012;Simonyan and Zisserman, 2014;He et al, 2016), preserve the representational geometry in object-sensitive visual areas (Khaligh-Razavi and Kriegeskorte, 2014;Yamins et al, 2014;Cichy et al, 2016), and predicts neural and fMRI responses to natural picture or video stimuli (Güçlü and van Gerven, 2015b, a;Wen et al, 2016Wen et al, , 2017Eickenberg et al, 2017;Seeliger et al, 2017), suggesting their close relevance to how the brain organizes and processes visual information (Cox and Dean, 2014;Kriegeskorte, 2015;Yamins and DiCarlo, 2016;Kietzmann et al, 2017;van Gerven, 2017). DNNs also open new opportunities for mapping the visual cortex, including the cortical hierarchy of spatial and temporal processing (Güçlü and van Gerven, 2015b, a;Cichy et al, 2016;Wen et al, 2016;Eickenberg et al, 2017), category representation and organization (Khaligh-Razavi and Kriegeskorte, 2014;Wen et al, 2017), visual-field maps (Wen et al...…”

Section: Discussionmentioning

confidence: 99%

“…Visual semantics explain the responses in the ventral temporal cortex but not at lower visual areas (Naselaris et al, 2009;Huth et al, 2012). On the other hand, braininspired deep neural networks (DNN) (LeCun et al, 2015), mimic the feedforward computation along the visual hierarchy (Kriegeskorte, 2015;Yamins and DiCarlo, 2016;Kietzmann et al, 2017;van Gerven, 2017), match human performance in image recognition (Krizhevsky et al, 2012;Simonyan and Zisserman, 2014;Szegedy et al, 2015;He et al, 2016), and explain cortical activity over nearly the entire visual cortex in response to natural visual stimuli (Yamins et al, 2014;Güçlü and van Gerven, 2015b, a;Wen et al, 2016Wen et al, , 2017Eickenberg et al, 2017;Seeliger et al, 2017).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Transferring and Generalizing Deep-Learning-based Neural Encoding Models across Subjects

Wen

Shi

Chen

et al. 2017

Preprint

View full text Add to dashboard Cite

Recent studies have shown the value of using deep learning models for mapping and characterizing how the brain represents and organizes information for natural vision. However, modeling the relationship between deep learning models and the brain (or encoding models), requires measuring cortical responses to large and diverse sets of natural visual stimuli from single subjects. This requirement limits prior studies to few subjects, making it difficult to generalize findings across subjects or for a population. In this study, we developed new methods to transfer and generalize encoding models across subjects. To train encoding models specific to a subject, the models trained for other subjects were used as the prior models and were refined efficiently using Bayesian inference with a limited amount of data from the specific subject. To train encoding models for a population, the models were progressively trained and updated with incremental data from different subjects. For the proof of principle, we applied these methods to functional magnetic resonance imaging (fMRI) data from three subjects watching tens of hours of naturalistic videos, while deep residual neural network driven by image recognition was used to model the visual cortical processing. Results demonstrate that the methods developed herein provide an efficient and effective strategy to establish subject-specific or populationwide predictive models of cortical representations of high-dimensional and hierarchical visual features.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Transferring and Generalizing Deep-Learning-based Neural Encoding Models across Subjects

Wen

Shi

Chen

et al. 2017

Preprint

View full text Add to dashboard Cite

show abstract

“…Artificial neural networks (ANNs) are computational models that are loosely inspired by the neurosciences and are mainly composed of simple processing units and their interconnections, so ANNs can learn from experience through modifying these interconnections [25]. Various ANN models have been studied in the past and recently researched in different applications.…”

Section: Prediction Model Based On Cnnmentioning

confidence: 99%

CNN-Based Volume Flow Rate Prediction of Oil–Gas–Water Three-Phase Intermittent Flow from Multiple Sensors

Chen

et al. 2021

Sensors

View full text Add to dashboard Cite

In this paper, we propose a deep-learning-based method using a convolutional neural network (CNN) to predict the volume flow rates of individual phases in the oil–gas–water three-phase intermittent flow simultaneously by analyzing the measurement data from multiple sensors, including a temperature sensor, a pressure sensor, a Venturi tube and a microwave sensor. To build datasets, a series of experiments for the oil–gas–water three-phase intermittent flow in a horizontal pipe, in which gas volume fraction and water-in-liquid ratio ranges are 23.77–94.45% and 14.95–86.97%, respectively, and gas flow superficial velocity and liquid flow superficial velocity ranges are 0.66–5.23 and 0.27–2.14 m/s, respectively, have been carried out on a test loop pipeline. The preliminary results indicate that the model can provide relative prediction errors on the testing-1 dataset for the volume flow rates of oil-phase, gas-phase and water-phase within ±10% with 94.49%, 92.56% and 95.71% confidence levels, respectively. Additionally, the prediction results on the testing-2 dataset also demonstrate the generalization ability of the model. The consuming time of a prediction with one sample is 0.43 s on an Intel Xeon CPU E5-2678 v3, and 0.01 s on an NVIDIA GeForce GTX 1080 Ti GPU. Hence, the proposed CNN-based prediction model, which can fulfill the real-time application requirements in the petroleum industry, reveals the potential of using deep learning to obtain accurate results in the multiphase flow measurement field.

show abstract

“…127 Britain had become one of the main exporters of crops to the Rhine army, which further increased demand for produce while manpower remained the same. 128 Additionally, the church as a powerful new landowner has to be considered, negotiating land distribution, tenancy and taxation. 129 Ultimately the rural population felt the strain, as the part of society from which resources are extracted is most affected by political instability, economic change and food shortages.…”

Section: Diseases Of Deprivation?mentioning

confidence: 99%

Listening to the Kids: The Value of Childhood Palaeopathology for the Study of Rural Roman Britain

Rohnbogner

2017

Britannia

View full text Add to dashboard Cite

Childhood palaeopathology remains an underutilised resource in the study of Roman Britain, particularly for exploring the lives of the rural population. Lesions in child skeletons provide unique insights into past lifeways and population health, as adverse environmental conditions translate more readily into the osteological record of these vulnerable members of society. To demonstrate the range of information gleaned from the children, 1,279 non-adults (0-17 years) from 26 first-to fifth-century urban and rural settlements were analysed, comparing morbidity and mortality in the most comprehensive study to date. The distribution of ages-atdeath suggests migration between country and town, the latter presenting a stressful and unsanitary environment. However, as demonstrated by high rates of metabolic disease and infections, life in the countryside was hampered by demanding physical labour and potentially oppressive conditions with restricted access to resources. stress; Roman rural settlement INTRUCTION: ROMANO-BRITISH CHILDHOODHuman skeletal remains present the most direct form of evidence for studying past populations. Numerous palaeopathological studies of Romano-British adults have provided insight into population health, diet and lifeways, albeit with a bias towards urban cemeteries. 1 More recently, several studies have reported on health in both urban and rural contexts, yet the children did not receive emphasis. 2 Roman archaeology is no different from other branches of the discipline in its research concerns, therefore the childhood experience has not yet fully evolved as a dedicated subject in Roman scholarship. While stable isotope analysis of Romano-British children has received considerable attention, 3 the actual health status of children remains relatively unexplored and a comprehensive picture of growing up in Roman Britain is still lacking. As children are growing, they reflect adverse environmental conditions more readily than their parents. 4 Child health is therefore a very powerful indicator of overall population health and dynamics, making non-adult (0-17 years) palaeopathology a primary source of evidence for past lifestyles and living environments.The osteological term 'non-adult' is not without problems but allows us to label, observe and discuss immature individuals in the past without involuntarily attaching social age norms. 5The cut-off age of 17 years is a product of the most accurate osteological ageing methods used and allows comparison between bioarchaeological datasets. Alongside age-at-death, a range of acquired conditions of ill-health can be observed in non-adult skeletons, which can provide information regarding cultural change, environmental pressures and life histories. The classical literature on Romano-British children is limited and, as so often in the study of Romano-British lifeways, we are forced to look to Rome for analogies and guidance. The classical literature, epigraphic and iconographic references provide a wealth of information on the Roman childhood experience...

show abstract

Computational Foundations of Natural Intelligence

Cited by 6 publications

References 280 publications

Transferring and Generalizing Deep-Learning-based Neural Encoding Models across Subjects

Transferring and Generalizing Deep-Learning-based Neural Encoding Models across Subjects

CNN-Based Volume Flow Rate Prediction of Oil–Gas–Water Three-Phase Intermittent Flow from Multiple Sensors

Listening to the Kids: The Value of Childhood Palaeopathology for the Study of Rural Roman Britain

Contact Info

Product

Resources

About