Browse
Submit Data
Databases
API
Help

Dataset Information

0 Views

0 Connections

0 Citations

0 Reanalyses

0 Downloads

Omics score: 0

Bioinspired Architecture Selection for Multitask Learning.

ABSTRACT: Faced with a new concept to learn, our brain does not work in isolation. It uses all previously learned knowledge. In addition, the brain is able to isolate the knowledge that does not benefit us, and to use what is actually useful. In machine learning, we do not usually benefit from the knowledge of other learned tasks. However, there is a methodology called Multitask Learning (MTL), which is based on the idea that learning a task along with other related tasks produces a transfer of information between them, what can be advantageous for learning the first one. This paper presents a new method to completely design MTL architectures, by including the selection of the most helpful subtasks for the learning of the main task, and the optimal network connections. In this sense, the proposed method realizes a complete design of the MTL schemes. The method is simple and uses the advantages of the Extreme Learning Machine to automatically design a MTL machine, eliminating those factors that hinder, or do not benefit, the learning process of the main task. This architecture is unique and it is obtained without testing/error methodologies that increase the computational complexity. The results obtained over several real problems show the good performances of the designed networks with this method.

SUBMITTER: Bueno-Crespo A

PROVIDER: S-EPMC5480319 | biostudies-other | 2017

REPOSITORIES: biostudies-other

ACCESS DATA

Json Xml

Similar Datasets

Bioinspired Breathable Architecture for Water Harvesting.

Project description:Thuja plicata is a coniferous tree which displays remarkable water channelling properties. In this article, an easily fabricated mesh inspired by the hierarchical macro surface structure of Thuja plicata branchlets is described which emulates this efficient water collection behaviour. The key parameters are shown to be the pore size, pore angle, mesh rotation, tilt angle (branch droop) and layering (branch overlap). Envisaged societal applications include water harvesting and low cost breathable architecture for developing countries.

| S-EPMC4649541 | biostudies-literature

Bioinspired Membrane Interfaces: Controlling Actomyosin Architecture and Contractility.

Project description:The creation of biologically inspired artificial lipid bilayers on planar supports provides a unique platform to study membrane-confined processes in a well-controlled setting. At the plasma membrane of mammalian cells, the linkage of the filamentous (F)-actin network is of pivotal importance leading to cell-specific and dynamic F-actin architectures, which are essential for the cell's shape, mechanical resilience, and biological function. These networks are established through the coordinated action of diverse actin-binding proteins and the presence of the plasma membrane. Here, we established phosphatidylinositol-4,5-bisphosphate (PtdIns[4,5]P2)-doped supported planar lipid bilayers to which contractile actomyosin networks were bound via the membrane-actin linker ezrin. This membrane system, amenable to high-resolution fluorescence microscopy, enabled us to analyze the connectivity and contractility of the actomyosin network. We found that the network architecture and dynamics are not only a function of the PtdIns[4,5]P2 concentration but also depend on the presence of negatively charged phosphatidylserine (PS). PS drives the attached network into a regime, where low but physiologically relevant connectivity to the membrane results in strong contractility of the actomyosin network, emphasizing the importance of the lipid composition of the membrane interface.

| S-EPMC9999349 | biostudies-literature

Bioinspired Molecular Factories with Architecture and In Vivo Functionalities as Cell Mimics.

Project description:Despite huge need in the medical domain and significant development efforts, artificial cells to date have limited composition and functionality. Although some artificial cells have proven successful for producing therapeutics or performing in vitro specific reactions, they have not been investigated in vivo to determine whether they preserve their architecture and functionality while avoiding toxicity. Here, these limitations are overcome and customizable cell mimic is achieved-molecular factories (MFs)-by supplementing giant plasma membrane vesicles derived from donor cells with nanometer-sized artificial organelles (AOs). MFs inherit the donor cell's natural cytoplasm and membrane, while the AOs house reactive components and provide cell-like architecture and functionality. It is demonstrated that reactions inside AOs take place in a close-to-nature environment due to the unprecedented level of complexity in the composition of the MFs. It is further demonstrated that in a zebrafish vertebrate animal model, these cell mimics show no apparent toxicity and retain their integrity and function. The unique advantages of highly varied composition, multicompartmentalized architecture, and preserved functionality in vivo open new biological avenues ranging from the study of biorelevant processes in robust cell-like environments to the production of specific bioactive compounds.

| S-EPMC7029636 | biostudies-literature

Higher Order Architecture of Designer Peptides Forms Bioinspired 10?nm siRNA Delivery System.

Project description:The higher-order architecture observed in biological systems, like viruses, is very effective in nucleic acid transport. The replications of this system has been attempted with both synthetic and naturally occurring polymers with mixed results. Here we describe a peptide/siRNA quaternary complex that functions as an siRNA delivery system. The rational design of a peptide assembly is inspired by the viral capsids, but not derived from them. We selected the collagen peptide (COL) to provide the structural stability and the folding framework, and hybridize it with the cell penetrating peptide (CPP) that allows for effective penetration of biological barriers. The peptide/siRNA quaternary complex forms stoichiometric, 10?nm nanoparticles, that show fast cellular uptake (<30?min), effective siRNA release, and gene silencing. The complex provides capsid-like protection for siRNA against nucleases without being immunostimulatory, or cytotoxic. Our data suggests that delivery vehicles based on synthetic quaternary structures that exhibit higher-order architecture may be effective in improving delivery and release of nucleic acid cargo.

| S-EPMC6856157 | biostudies-literature

In Situ Remodeling Overrules Bioinspired Scaffold Architecture of Supramolecular Elastomeric Tissue-Engineered Heart Valves.

Project description:In situ tissue engineering that uses resorbable synthetic heart valve scaffolds is an affordable and practical approach for heart valve replacement; therefore, it is attractive for clinical use. This study showed no consistent collagen organization in the predefined direction of electrospun scaffolds made from a resorbable supramolecular elastomer with random or circumferentially aligned fibers, after 12 months of implantation in sheep. These unexpected findings and the observed intervalvular variability highlight the need for a mechanistic understanding of the long-term in situ remodeling processes in large animal models to improve predictability of outcome toward robust and safe clinical application.

| S-EPMC7775962 | biostudies-literature

Collagenic architecture and morphotraits in a marine basal metazoan as a model for bioinspired applied research.

Project description:In some Porifera (Demospongiae: Keratosa), prototypes of the connective system are almost exclusively based on collagenic networks. We studied the topographic distribution, spatial layout, microtraits, and/or morphogenesis of these collagenic structures in Ircinia retidermata (Dictyoceratida: Irciniidae). Analyses were carried out on a clonal strain from sustainable experimental mariculture by using light and scanning electron microscopy. Histology revealed new insights on the widely diversified and complex hierarchical assemblage of collagenic structures. Key evolutionary novelties in the organization of sponge connective system were found out. The aquiferous canals are shaped as corrugate-like pipelines conferring plasticity to the water circulation system. Compact clusters of elongated cells are putatively involved in a nutrient transferring system. Knob-ended filaments are characterized by a banding pattern and micro-components. Ectosome and outer endosome districts are the active fibrogenetic areas, where exogenous material constitutes an axial condensation nucleus for the ensuing morphogenesis. The new data can be useful to understand not only the evolutionary novelties occurring in the target taxon but also the morpho-functional significance of its adaptive collagenic anatomical traits. In addition, data may give insights on both marine collagen sustainable applied researches along with evolutionary and phylogenetic analyses, thus highlighting sponges as a key renewable source for inspired biomaterials. Therefore, we also promote bioresources sustainable exploitation with the aim to provide new donors of marine collagen, thereby supporting conservation of wild populations/species.

| S-EPMC9306819 | biostudies-literature

The network architecture of value learning.

Project description:Value guides behavior. With knowledge of stimulus values and action consequences, behaviors that maximize expected reward can be selected. Prior work has identified several brain structures critical for representing both stimuli and their values. Yet, it remains unclear how these structures interact with one another and with other regions of the brain to support the dynamic acquisition of value-related knowledge. Here, we use a network neuroscience approach to examine how BOLD functional networks change as 20 healthy human subjects learn the values of novel visual stimuli over the course of four consecutive days. We show that connections between regions of the visual, frontal, and cingulate cortices become stronger as learning progresses, with some of these changes being specific to the type of feedback received during learning. These results demonstrate that functional networks dynamically track behavioral improvement in value judgments, and that interactions between network communities form predictive biomarkers of learning.

| S-EPMC6130435 | biostudies-literature

Learning from Northern clingfish (Gobiesox maeandricus): bioinspired suction cups attach to rough surfaces.

Project description:While artificial suction cups only attach well to smooth surfaces, the Northern clingfish can attach to surfaces ranging from nanoscale smooth to rough stone. This ability is highly desirable for technical applications. The morphology of the fish's suction disc and its ability to attach to rough and slimy surfaces have been described before, and here we aim to close gaps in the biomechanical understanding, and transfer the biomechanical principles to technical suction cups. We demonstrate that the margin of the suction disc is the critical feature enabling attachment to rough surfaces. Second, friction measurements show that friction of the disc rim is increased on rough substrates and contributes to high tenacity. Increased friction causes a delay in failure of the suction cup and increases the attachment force. We were able to implement these concepts to develop the first suction cups bioinspired by Northern clingfish. These cups attach with tenacities up to 70 kPa on surfaces as rough as 270 µm grain size. The application of this technology is promising in fields such as surgery, industrial production processes and whale tagging. This article is part of the theme issue 'Transdisciplinary approaches to the study of adhesion and adhesives in biological systems'.

| S-EPMC6745484 | biostudies-literature

Predicting future learning from baseline network architecture.

Project description:Human behavior and cognition result from a complex pattern of interactions between brain regions. The flexible reconfiguration of these patterns enables behavioral adaptation, such as the acquisition of a new motor skill. Yet, the degree to which these reconfigurations depend on the brain's baseline sensorimotor integration is far from understood. Here, we asked whether spontaneous fluctuations in sensorimotor networks at baseline were predictive of individual differences in future learning. We analyzed functional MRI data from 19 participants prior to six weeks of training on a new motor skill. We found that visual-motor connectivity was inversely related to learning rate: sensorimotor autonomy at baseline corresponded to faster learning in the future. Using three additional scans, we found that visual-motor connectivity at baseline is a relatively stable individual trait. These results suggest that individual differences in motor skill learning can be predicted from sensorimotor autonomy at baseline prior to task execution.

| S-EPMC5910215 | biostudies-literature

Deep Learning Architecture Reduction for fMRI Data.

Project description:In recent years, deep learning models have demonstrated an inherently better ability to tackle non-linear classification tasks, due to advances in deep learning architectures. However, much remains to be achieved, especially in designing deep convolutional neural network (CNN) configurations. The number of hyper-parameters that need to be optimized to achieve accuracy in classification problems increases with every layer used, and the selection of kernels in each CNN layer has an impact on the overall CNN performance in the training stage, as well as in the classification process. When a popular classifier fails to perform acceptably in practical applications, it may be due to deficiencies in the algorithm and data processing. Thus, understanding the feature extraction process provides insights to help optimize pre-trained architectures, better generalize the models, and obtain the context of each layer's features. In this work, we aim to improve feature extraction through the use of a texture amortization map (TAM). An algorithm was developed to obtain characteristics from the filters amortizing the filter's effect depending on the texture of the neighboring pixels. From the initial algorithm, a novel geometric classification score (GCS) was developed, in order to obtain a measure that indicates the effect of one class on another in a classification problem, in terms of the complexity of the learnability in every layer of the deep learning architecture. For this, we assume that all the data transformations in the inner layers still belong to a Euclidean space. In this scenario, we can evaluate which layers provide the best transformations in a CNN, allowing us to reduce the weights of the deep learning architecture using the geometric hypothesis.

| S-EPMC8870362 | biostudies-literature

OmicsDI is part of the ELIXIR infrastructure

OmicsDI is an Elixir interoperability service. Learn more ›

Tweets

OmicsDI Databases

PRIDE
PeptideAtlas
MassIVE
JPOST Repository
Physiome Model Repository

EGA
EVA
ENA
LINCS
PAXDB
Cell Collective

MetaboLights
Metabolomics Workbench
MetabolomeExpress
GNPS
BioModels
FAIRDOMHub

ArrayExpress
dbGaP
ExpressionAtlas
GEO
NODE

Information

Databases
Help
API
Contact us
Code on GitHub
Terms of use
Submit Data