Set Of Kinetic Data Research Articles

Laplacian eigenmaps based manifold regularized CNN for visual recognition

The paper proposes a novel Laplacian eigenmaps based manifold regularized CNN (LE-CNN) for action recognition. The proposed LE-CNN model incorporates Laplacian eigenmaps based manifold structure information of training samples into CNN layer by layer, which can keep adjacent samples as close as possible during space transformation. In addition, the Laplacian eigenmaps based manifold structure can accelerate convergence during the training process. Experiments are performed on two standard action datasets, UCF101 and HMDB51, to validate the proposed LE-CNN model. Furthermore, we extend experiments on a large-scale action recognition dataset (i.e. the Kinetics dataset) and compare the LE-CNN model with other advanced models. In addition to action recognition, we also apply the LE-CNN model to an image classification task on the CIFAR-10 dataset, to demonstrate the effectiveness of the LE-CNN model across different classification tasks.

Usefulness of insights into the kinetic compensation effects in the kinetic analysis of the coal gasification process

AbstractIn this paper, an attempt has been made to understand the kinetic compensation effects and their usefulness in the kinetic analysis of a lab‐based gasification study. The gasification experiment was carried out in two different gasifying environments, that is, 0.4%O₂ + 8%H₂O‐Ar and 8%H₂O‐Ar, for two different particle sizes of Loy Yang brown coal. Analysis of kinetic values with the change in particle size and gasifying environment was investigated. This provides information on the path of product gas formation and how the overall controlling factor affects the path of char gasification, including the rate‐limiting step. Furthermore, the results indicate that having multiple sets of kinetic parameters caused by the inclusion of the change in char properties into kinetics during solid–gas heterogeneous reactions opens up the scope for wider applications in chemical reaction engineering. This includes the design of a reactor with a proper kinetic model, optimization of feedstock, and process parameters with the identification of pathways for product gas formation, which ultimately plays a key role in scaling up technology from bench‐scale to plant‐scale. In contrast, the study of kinetics with having a single set of kinetic data based on the initial change in char properties limits its applications.

Measuring Kinetics under Vibrational Strong Coupling: Testing for a Change in the Nucleophilicity of Water and Alcohols.

Vibrational Strong Coupling (VSC) has been reported to change the rate of organic reactions. However, a lack of convenient and reliable methods to measure reaction kinetics under VSC makes it challenging to obtain mechanistic insight into its influence, hindering progress in the field. Here, we use recently developed fixed-width optical cavities to obtain large kinetic datasets under VSC with small errors (± 1-5%) in an operationally simple manner using UV-vis spectroscopy. The setup is used to test whether VSC changes a fundamental kinetic property of polar reactions, nucleophilicity, for water and alcohols, species commonly used in VSC-modified chemistry. We determined the rate constants for nucleophilic capture with a library of benzhydrilium ions as reference electrophiles with and without strong coupling of the nucleophile's key vibrations. For all investigated combinations of electrophiles and nucleophiles, only minor changes in the observed rate constants of the reactions were observed, independently of the coupled bands. These results indicate that VSC does not substantially alter the nucleophilicity of water and alcohols, suggesting that polar reactions are modified through other, presently unknown mechanisms. Fixed-width cavities allow for convenient and reproducible UV-vis kinetics, facilitating mechanistic studies of VSC-modified chemistry.

Measuring Kinetics under Vibrational Strong Coupling: Testing for a Change in the Nucleophilicity of Water and Alcohols

Vibrational Strong Coupling (VSC) has been reported to change the rate of organic reactions. However, a lack of convenient and reliable methods to measure reaction kinetics under VSC makes it challenging to obtain mechanistic insight into its influence, hindering progress in the field. Here, we use recently developed fixed‐width optical cavities to obtain large kinetic datasets under VSC with small errors (± 1‐5%) in an operationally simple manner using UV‐vis spectroscopy. The setup is used to test whether VSC changes a fundamental kinetic property of polar reactions, nucleophilicity, for water and alcohols, species commonly used in VSC‐modified chemistry. We determined the rate constants for nucleophilic capture with a library of benzhydrilium ions as reference electrophiles with and without strong coupling of the nucleophile’s key vibrations. For all investigated combinations of electrophiles and nucleophiles, only minor changes in the observed rate constants of the reactions were observed, independently of the coupled bands. These results indicate that VSC does not substantially alter the nucleophilicity of water and alcohols, suggesting that polar reactions are modified through other, presently unknown mechanisms. Fixed‐width cavities allow for convenient and reproducible UV‐vis kinetics, facilitating mechanistic studies of VSC‐modified chemistry.

Electrochemical conversion of contaminant p-nitrophenol to p-aminophenol and high-value p-quinones: Mechanism, kinetics and modeling

Understanding the chemistry of contaminant p-nitrophenol (p-NP) in electrochemical systems is of importance to remove it from wastewater and to produce high-value products such as quinones. In this study, redox experiments in a system initially containing an aqueous solution of sodium chloride and p-NP at pH 5 and 25 °C were performed in an electrochemical reactor with low-cost steel electrodes. The reactor was operated at 0.1 and 0.2 A, and at initial p-NP concentrations of 4.04×10−5, 6.97×10−5 and 10.3×10−5 M. Air and argon were bubbled at the bottom of the cathode to have solutions saturated with O2 (7.7 mg L−1) and free of dissolved O2. At these conditions p-NP was completely converted to less toxic and more valuable p-hydroquinone and p-aminophenol in an hour. The reaction rates were individually controlled by the steps of electron transfer and adsorption of p-NP at the cathode surface. In the presence of dissolved O2, p-hydroquinone and p-aminophenol were intermediate species oxidized to p-benzoquinone, and p-quinoneimine in the greatest amount, respectively. The rate constants for the reactions were determined on the basis of kinetic results obtained by liquid chromatography and spectrophotometry. The suggested mechanism, and the innovative kinetic model involving Langmuir-Hinshelwood rate expressions for the surface reactions, were consistent with results from mass spectrometry and total organic carbon. For the first time for this kind of kinetic model, it was also able to reproduce a large set of kinetic data of electroreduction of p-NP with iron electrodes obtained by independent researchers at several different conditions.

Experimental investigation and modeling of mass transport properties of O2 in biomass chars

This study presents two sets of kinetic adsorption data for O2 on two different biomass chars at temperatures from (298 to 373) K and pressures from (26 to 300) kPa. Thereon, incorporating data of 21 adsorption kinetic curves on a hydrochar and 24 adsorption kinetic curves on a torrefied beechwood char, a parameterization of the pore-structure dependent kinetic adsorption (PSK) model is obtained for the two investigated chars. Extensive possibilities to predict time- and pore-structure-resolved mass transport properties using the PSK model are discussed. The mass transfer coefficient and the adsorption flow rate are direct outcomes and derivatives of the PSK model and hold for a physically sound description of the time-resolved mass transport of gaseous species in the porous structure of particles. The physical and extrapolative behavior of the modeled mass transfer coefficients and adsorption flow rates are discussed for temperatures up to 1100 K, allowing for new insights into the mass transfer.

Kinetics of Phosphate Ions and Phytase Activity Production for Lactic Acid-Producing Bacteria Utilizing Milling and Whitening Stages Rice Bran as Biopolymer Substrates.

A study evaluated nine kinetic data and four kinetic parameters related to growth, production of various phytase activities (PEact), and released phosphate ion concentration ([Pi]) from five lactic acid bacteria (LAB) strains cultivated in three types of media: phytate (IP6), milling stage rice bran (MsRB), and whitening stage rice bran (WsRB). Score ranking techniques were used, combining these kinetic data and parameters to select the most suitable LAB strain for each medium across three cultivation time periods (24, 48, and 72 h). In the IP6 medium, Lacticaseibacillus casei TISTR 1500 exhibited statistically significant highest (p ≤ 0.05) normalized summation scores using a 2:1 weighting between kinetic and parameter data sets. This strain also had the statistically highest levels (p ≤ 0.05) of produced phosphate ion concentration ([Pi]) (0.55 g/L) at 72 h and produced extracellular specific phytase activity (ExSp-PEact) (0.278 U/mgprotein) at 48 h. For the MsRB and WsRB media, Lactiplantibacillus plantarum TISTR 877 performed exceptionally well after 72 h of cultivation. It produced ([Pi], ExSp-PEact) pairs of (0.53 g/L, 0.0790 U/mgprotein) in MsRB and (0.85 g/L, 0.0593 U/mgprotein) in WsRB, respectively. Overall, these findings indicate the most promising LAB strains for each medium and cultivation time based on their ability to produce phosphate ions and extracellular specific phytase activity. The selection process utilized a combination of kinetic data and parameter analysis.

Sports competition tactical analysis model of cross-modal transfer learning intelligent robot based on Swin Transformer and CLIP.

This paper presents an innovative Intelligent Robot Sports Competition Tactical Analysis Model that leverages multimodal perception to tackle the pressing challenge of analyzing opponent tactics in sports competitions. The current landscape of sports competition analysis necessitates a comprehensive understanding of opponent strategies. However, traditional methods are often constrained to a single data source or modality, limiting their ability to capture the intricate details of opponent tactics. Our system integrates the Swin Transformer and CLIP models, harnessing cross-modal transfer learning to enable a holistic observation and analysis of opponent tactics. The Swin Transformer is employed to acquire knowledge about opponent action postures and behavioral patterns in basketball or football games, while the CLIP model enhances the system's comprehension of opponent tactical information by establishing semantic associations between images and text. To address potential imbalances and biases between these models, we introduce a cross-modal transfer learning technique that mitigates modal bias issues, thereby enhancing the model's generalization performance on multimodal data. Through cross-modal transfer learning, tactical information learned from images by the Swin Transformer is effectively transferred to the CLIP model, providing coaches and athletes with comprehensive tactical insights. Our method is rigorously tested and validated using Sport UV, Sports-1M, HMDB51, and NPU RGB+D datasets. Experimental results demonstrate the system's impressive performance in terms of prediction accuracy, stability, training time, inference time, number of parameters, and computational complexity. Notably, the system outperforms other models, with a remarkable 8.47% lower prediction error (MAE) on the Kinetics dataset, accompanied by a 72.86-second reduction in training time. The presented system proves to be highly suitable for real-time sports competition assistance and analysis, offering a novel and effective approach for an Intelligent Robot Sports Competition Tactical Analysis Model that maximizes the potential of multimodal perception technology. By harnessing the synergies between the Swin Transformer and CLIP models, we address the limitations of traditional methods and significantly advance the field of sports competition analysis. This innovative model opens up new avenues for comprehensive tactical analysis in sports, benefiting coaches, athletes, and sports enthusiasts alike.

EMO-MoviNet: Enhancing Action Recognition in Videos with EvoNorm, Mish Activation, and Optimal Frame Selection for Efficient Mobile Deployment.

The primary goal of this study is to develop a deep neural network for action recognition that enhances accuracy and minimizes computational costs. In this regard, we propose a modified EMO-MoviNet-A2* architecture that integrates Evolving Normalization (EvoNorm), Mish activation, and optimal frame selection to improve the accuracy and efficiency of action recognition tasks in videos. The asterisk notation indicates that this model also incorporates the stream buffer concept. The Mobile Video Network (MoviNet) is a member of the memory-efficient architectures discovered through Neural Architecture Search (NAS), which balances accuracy and efficiency by integrating spatial, temporal, and spatio-temporal operations. Our research implements the MoviNet model on the UCF101 and HMDB51 datasets, pre-trained on the kinetics dataset. Upon implementation on the UCF101 dataset, a generalization gap was observed, with the model performing better on the training set than on the testing set. To address this issue, we replaced batch normalization with EvoNorm, which unifies normalization and activation functions. Another area that required improvement was key-frame selection. We also developed a novel technique called Optimal Frame Selection (OFS) to identify key-frames within videos more effectively than random or densely frame selection methods. Combining OFS with Mish nonlinearity resulted in a 0.8-1% improvement in accuracy in our UCF101 20-classes experiment. The EMO-MoviNet-A2* model consumes 86% fewer FLOPs and approximately 90% fewer parameters on the UCF101 dataset, with a trade-off of 1-2% accuracy. Additionally, it achieves 5-7% higher accuracy on the HMDB51 dataset while requiring seven times fewer FLOPs and ten times fewer parameters compared to the reference model, Motion-Augmented RGB Stream (MARS).

The intracellular parasite Anncaliia algerae induces a massive miRNA down-regulation in human cells.

Anncaliia algerae belongs to microsporidia, a group of obligate intracellular parasites related to fungi. These parasites are largely spread in water and food-webs and can infect a wide variety of hosts ranging from invertebrates to vertebrates including humans. In humans, microsporidian infections are mainly opportunistic as immunocompetent hosts can clear parasites naturally. Recent studies however have reported persistent microsporidian infections and have highlighted them as a risk factor in colon cancer. This may be a direct result of cell infection or may be an indirect effect of the infectious microenvironment and the host's response. In both cases, this raises the question of the effects of microsporidian infection at the host and host-cell levels. We aimed to address the question of human host intracellular response to microsporidian infection through a transcriptomic kinetic study of human foreskin fibroblasts (HFF) infected with A.algerae, a human infecting microsporidia with an exceptionally wide host range. We focused solely on host response studying both coding and small non-coding miRNA expression. Our study revealed a generalized down-regulation of cell miRNAs throughout infection with up to 547 different miRNAs downregulated at some timepoints and also transcriptomic dysregulations that could facilitate parasite development with immune and lipid metabolism genes modulation. We also hypothesize possible small nucleic acid expropriation explaining the miRNA downregulation. This work contributes to a better understanding of the dialogue that can occur between an intracellular parasite and its host at the cellular level, and can guide future studies on microsporidian infection biology to unravel the mode of action of these minimalist parasites at the tissue or host levels.We have also generated a kinetic and comprehensive transcriptomic data set of an infectious process that can help support comparative studies in the broader field of parasitology. Lastly, these results may warrant for caution regarding microsporidian exposure and persistent infections.

Factor analysis for signal modeling and noise characterization in spectro-kinetic data

Measurement signals and noise (heteroscedastic and homoscedastic) in different evolutionary spectral kinetic data sets are modeled and characterized by applying maximum likelihood common factor analysis (MLFA) and principal axis factoring (PAF), as factor-based techniques. They decompose data into common and specific factors, that are assumed to be constant heteroscedastic. The digital filter is also used for characterization of noise in data and to confirm the obtained results from factor analysis techniques. No one of the applied methods requires the use of replicated measurements. Inverted Blackman windowed sinc (BWS) smoothing coefficients as high-pass digital filters are employed to provide point and bin estimates of noise from measurement vectors without any assumption about noise structure. The proportionality of noise to signals and their characteristics are also reported and are comparable when applying different methods.The proposed methods are compared using both simulated and experimental kinetic data sets. Experimental data includes fluorescence and UV–vis absorbance kinetic spectra. A simple solution for problem of MLFA application to fat data is proposed and successfully applied. Different types of independent noise including identical independent distribution (iid), constant heteroscedastic, and general heteroscedastic are added to the simulated data, in various levels.In the case of iid noise results from all methods are almost the same. In the presence of heteroscedastic noise results from MLFA and PAF are preferred to PCA regarding subspace angles and discrepancy function values. Employment of methods for highly overlapped, skinny, and fat data shows the superiority of MLFA to PAF and PCA. The independent noise characteristics obtained from factor analysis methods are comparable to that from digital filters.

Surface electromyogram, kinematic, and kinetic dataset of lower limb walking for movement intent recognition

Surface electromyogram (sEMG) offers a rich set of motor information for decoding limb motion intention that serves as a control input to Intelligent human-machine synergy systems (IHMSS). Despite growing interest in IHMSS, the current publicly available datasets are limited and can hardly meet the growing demands of researchers. This study presents a novel lower limb motion dataset (designated as SIAT-LLMD), comprising sEMG, kinematic, and kinetic data with corresponding labels acquired from 40 healthy humans during 16 movements. The kinematic and kinetic data were collected using a motion capture system and six-dimensional force platforms and processed using OpenSim software. The sEMG data were recorded using nine wireless sensors placed on the subjects’ thigh and calf muscles on the left limb. Besides, SIAT-LLMD provides labels to classify the different movements and different gait phases. Analysis of the dataset verified the synchronization and reproducibility, and codes for effective data processing are provided. The proposed dataset can serve as a new resource for exploring novel algorithms and models for characterizing lower limb movements.

Estimation of Lower Extremity Joint Moments and 3D Ground Reaction Forces Using IMU Sensors in Multiple Walking Conditions: A Deep Learning Approach.

Human kinetics, specifically joint moments and ground reaction forces (GRFs) can provide important clinical information and can be used to control assistive devices. Traditionally, collection of kinetics is mostly limited to the lab environment because it relies on data that are measured from a motion capture system and floor-embedded force plates to calculate the dynamics via musculoskeletal models. This spatially limited method makes it extremely challenging to measure kinetics outside the laboratory in a variety of walking conditions due to the expensive device setup and large space required. Recently, employing machine learning with IMU sensors are suggested as an alternative method for biomechanical analyses. Although these methods enable estimating human kinetic data outside the laboratory by linking IMU sensor data with kinetics dataset, they are limited to show inaccurate kinetic estimates even in highly repeatable single walking conditions due to the employment of generic deep learning algorithms. Thus, this paper proposes a novel deep learning model, Kinetics-FM-DLR-Ensemble-Net for single limb prediction of hip, knee, and ankle joint moments and 3 dimensional Ground Reaction Forces (GRFs) using three IMU sensors on the thigh, shank, and foot under several representatives walking conditions in daily living, such as treadmill, level-ground, stair, and ramp. This is the first study that implements both joint moments and GRFs in multiple walking conditions using IMU sensors via deep learning. Our deep learning model is versatile and accurate for identifying human kinetics across diverse subjects and walking conditions and our model outperforms state-of-the-art deep learning model for kinetics estimation by a large margin.

Relational Action Bank with Semantic–Visual Attention for Few-Shot Action Recognition

Recently, few-shot learning has attracted significant attention in the field of video action recognition, owing to its data-efficient learning paradigm. Despite the encouraging progress, identifying ways to further improve the few-shot learning performance by exploring additional or auxiliary information for video action recognition remains an ongoing challenge. To address this problem, in this paper we make the first attempt to propose a relational action bank with semantic–visual attention for few-shot action recognition. Specifically, we introduce a relational action bank as the auxiliary library to assist the network in understanding the actions in novel classes. Meanwhile, the semantic–visual attention is devised to adaptively capture the connections to the foregone actions via both semantic correlation and visual similarity. We extensively evaluate our approach via two backbone models (ResNet-50 and C3D) on HMDB and Kinetics datasets, and demonstrate that the proposed model can obtain significantly better performance compared against state-of-the-art methods. Notably, our results demonstrate an average improvement of about 6.2% when compared to the second-best method on the Kinetics dataset.

Using BlazePose on Spatial Temporal Graph Convolutional Networks for Action Recognition

The ever-growing available visual data (i.e., uploaded videos and pictures by internet users) has attracted the research community's attention in the computer vision field. Therefore, finding efficient solutions to extract knowledge from these sources is imperative. Recently, the BlazePose system has been released for skeleton extraction from images oriented to mobile devices. With this skeleton graph representation in place, a Spatial-Temporal Graph Convolutional Network can be implemented to predict the action. We hypothesize that just by changing the skeleton input data for a different set of joints that offers more information about the action of interest, it is possible to increase the performance of the Spatial-Temporal Graph Convolutional Network for HAR tasks. Hence, in this study, we present the first implementation of the BlazePose skeleton topology upon this architecture for action recognition. Moreover, we propose the Enhanced-BlazePose topology that can achieve better results than its predecessor. Additionally, we propose different skeleton detection thresholds that can improve the accuracy performance even further. We reached a top-1 accuracy performance of 40.1% on the Kinetics dataset. For the NTU-RGB+D dataset, we achieved 87.59% and 92.1% accuracy for Cross-Subject and Cross-View evaluation criteria, respectively.

Self-Supervised Learning from Untrimmed Videos via Hierarchical Consistency.

Natural untrimmed videos provide rich visual content for self-supervised learning. Yet most previous efforts to learn spatio-temporal representations rely on manually trimmed videos, such as Kinetics dataset (Carreira and Zisserman 2017), resulting in limited diversity in visual patterns and limited performance gains. In this work, we aim to improve video representations by leveraging the rich information in natural untrimmed videos. For this purpose, we propose learning a hierarchy of temporal consistencies in videos, i.e., visual consistency and topical consistency, corresponding respectively to clip pairs that tend to be visually similar when separated by a short time span, and clip pairs that share similar topics when separated by a long time span. Specifically, we present a Hierarchical Consistency (HiCo++) learning framework, in which the visually consistent pairs are encouraged to share the same feature representations by contrastive learning, while topically consistent pairs are coupled through a topical classifier that distinguishes whether they are topic-related, i.e., from the same untrimmed video. Additionally, we impose a gradual sampling algorithm for the proposed hierarchical consistency learning, and demonstrate its theoretical superiority. Empirically, we show that HiCo++ can not only generate stronger representations on untrimmed videos, but also improve the representation quality when applied to trimmed videos. This contrasts with standard contrastive learning, which fails to learn powerful representations from untrimmed videos. Source code will be made available here.

Why does the metabolic cost of walking increase on compliant substrates?

Walking on compliant substrates requires more energy than walking on hard substrates but the biomechanical factors that contribute to this increase are debated. Previous studies suggest various causative mechanical factors, including disruption to pendular energy recovery, increased muscle work, decreased muscle efficiency and increased gait variability. We test each of these hypotheses simultaneously by collecting a large kinematic and kinetic dataset of human walking on foams of differing thickness. This allowed us to systematically characterize changes in gait with substrate compliance, and, by combining data with mechanical substrate testing, drive the very first subject-specific computer simulations of human locomotion on compliant substrates to estimate the internal kinetic demands on the musculoskeletal system. Negative changes to pendular energy exchange or ankle mechanics are not supported by our analyses. Instead we find that the mechanistic causes of increased energetic costs on compliant substrates are more complex than captured by any single previous hypothesis. We present a model in which elevated activity and mechanical work by muscles crossing the hip and knee are required to support the changes in joint (greater excursion and maximum flexion) and spatio-temporal kinematics (longer stride lengths, stride times and stance times, and duty factors) on compliant substrates.

Predicting plant Rubisco kinetics from RbcL sequence data using machine learning.

Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) is responsible for the conversion of atmospheric CO2 to organic carbon during photosynthesis, and often acts as a rate limiting step in the later process. Screening the natural diversity of Rubisco kinetics is the main strategy used to find better Rubisco enzymes for crop engineering efforts. Here, we demonstrate the use of Gaussian processes (GPs), a family of Bayesian models, coupled with protein encoding schemes, for predicting Rubisco kinetics from Rubisco large subunit (RbcL) sequence data. GPs trained on published experimentally obtained Rubisco kinetic datasets were applied to over 9000 sequences encoding RbcL to predict Rubisco kinetic parameters. Notably, our predicted kinetic values were in agreement with known trends, e.g. higher carboxylation turnover rates (Kcat) for Rubisco enzymes from C4 or crassulacean acid metabolism (CAM) species, compared with those found in C3 species. This is the first study demonstrating machine learning approaches as a tool for screening and predicting Rubisco kinetics, which could be applied to other enzymes.

High accuracy barrier heights, enthalpies, and rate coefficients for chemical reactions

Quantitative chemical reaction data, including activation energies and reaction rates, are crucial for developing detailed kinetic mechanisms and accurately predicting reaction outcomes. However, such data are often difficult to find, and high-quality datasets are especially rare. Here, we use CCSD(T)-F12a/cc-pVDZ-F12//ωB97X-D3/def2-TZVP to obtain high-quality single point calculations for nearly 22,000 unique stable species and transition states. We report the results from these quantum chemistry calculations and extract the barrier heights and reaction enthalpies to create a kinetics dataset of nearly 12,000 gas-phase reactions. These reactions involve H, C, N, and O, contain up to seven heavy atoms, and have cleaned atom-mapped SMILES. Our higher-accuracy coupled-cluster barrier heights differ significantly (RMSE of ∼5 kcal mol−1) relative to those calculated at ωB97X-D3/def2-TZVP. We also report accurate transition state theory rate coefficients {k}_{infty }(T) between 300 K and 2000 K and the corresponding Arrhenius parameters for a subset of rigid reactions. We believe this data will accelerate development of automated and reliable methods for quantitative reaction prediction.

From YouTube to the brain: Transfer learning can improve brain-imaging predictions with deep learning

Deep learning has recently achieved best-in-class performance in several fields, including biomedical domains such as X-ray images. Yet, data scarcity poses a strict limit on training successful deep learning systems in many, if not most, biomedical applications, including those involving brain images. In this study, we translate state-of-the-art transfer learning techniques for single-subject prediction of simpler (sex and age) and more complex phenotypes (number of people in household, household income, fluid intelligence and smoking behavior). We fine-tuned 2D and 3D ResNet-18 convolutional neural networks for target phenotype predictions from brain images of ∼40,000 UK Biobank participants, after pretraining on YouTube videos from the Kinetics dataset and natural images from the ImageNet dataset. Transfer learning was effective on several phenotypes, especially sex and age classification. Additionally, transfer learning in particular outperformed deep learning models trained from scratch especially on smaller sample sizes. The out-of-sample performance using transfer learning from previously learned knowledge based on real-world images and videos could unlock the potential in many areas of imaging neuroscience where deep learning solutions are currently infeasible.