“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

The field of meta-learning, or learning-to-learn, has seen a dramatic rise in interest in recent years. Contrary to conventional approaches to AI where tasks are solved from scratch using a fixed learning algorithm, meta-learning aims to improve the learning algorithm itself, given the experience of multiple learning episodes. This paradigm provides an opportunity to tackle many conventional challenges of deep learning, including data and computation bottlenecks, as well as generalization. This survey describes the contemporary meta-learning landscape. We first discuss definitions of meta-learning and position it with respect to related fields, such as transfer learning and hyperparameter optimization. We then propose a new taxonomy that provides a more comprehensive breakdown of the space of meta-learning methods today. We survey promising applications and successes of meta-learning such as few-shot learning and reinforcement learning. Finally, we discuss outstanding challenges and promising areas for future research.

Meta-Learning in Neural Networks: A Survey

https://digital.library.adelaide.edu.au/dspace/bitstream/2440/101132/3/hdl_101132.pdf

Bidirectional communication between cumulus cells and the oocyte: Old hands and new players?

Noncoding RNAs (ncRNAs) have long been known to play vital roles in eukaryotic gene regulation. Studies conducted over a decade ago revealed that maturation of spliced, polyadenylated coding mRNA occurs by reactions involving small nuclear RNAs and small nucleolar RNAs; mRNA translation depends on activities mediated by transfer RNAs and ribosomal RNAs, subject to negative regulation by micro RNAs; transcriptional competence of sex chromosomes and some imprinted genes is regulated in cis by ncRNAs that vary by species; and both small-interfering RNAs and piwi-interacting RNAs bound to Argonaute-family proteins regulate post-translational modifications on chromatin and local gene expression states. More recently, gene-regulating noncoding RNAs have been identified, such as long intergenic and long noncoding RNAs (collectively referred to as lncRNAs)--a class totaling more than 100,000 transcripts in humans, which include some of the previously mentioned RNAs that regulate dosage compensation and imprinted gene expression. Here, we provide an overview of lncRNA activities, and then review the role of lncRNAs in processes vital to reproduction, such as germ cell specification, sex determination and gonadogenesis, sex hormone responses, meiosis, gametogenesis, placentation, non-genetic inheritance, and pathologies affecting reproductive tissues. Results from many species are presented to illustrate the evolutionarily conserved processes lncRNAs are involved in.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/mrd.22581

Long Non-Coding RNA Regulation of Reproduction and Development

So far, the characteristics of a good quality egg have been elusive, similar to the nature of the physiological, cellular, and molecular cues leading to its production both in vivo and in vitro. Current understanding highlights a strong and complex interdependence between the follicular cells and the gamete. Secreted factors induce cellular responses in the follicular cells, and direct exchange of small molecules from the cumulus cells to the oocyte through gap junctions controls meiotic arrest. Studying the interconnection between the cumulus cells and the oocyte, we previously demonstrated that the somatic cells also contribute transcripts to the gamete. Here, we show that these transcripts can be visualized moving down the transzonal projections (TZPs) to the oocyte, and that a time course analysis revealed progressive RNA accumulation in the TZPs, indicating that RNA transfer occurs before the initiation of meiosis resumption under a timetable fitting with the acquisition of developmental competence. A comparison of the identity of the nascent transcripts trafficking in the TZPs, with those in the oocyte increasing in abundance during maturation, and that are present on the oocyte's polyribosomes, revealed transcripts common to all three fractions, suggesting the use of transferred transcripts for translation. Furthermore, the removal of potential RNA trafficking by stripping the cumulus cells caused a significant reduction in maturation rates, indicating the need for the cumulus cell RNA transfer to the oocyte. These results offer a new perspective to the determinants of oocyte quality and female fertility, as well as provide insight that may eventually be used to improve in vitro maturation conditions.

https://bioone.org/journals/biology-of-reproduction/volume-94/issue-1/biolreprod.114.127571/Cumulus-Cell-Transcripts-Transit-to-the-Bovine-Oocyte-in-Preparation/10.1095/biolreprod.114.127571.pdf

Cumulus Cell Transcripts Transit to the Bovine Oocyte in Preparation for Maturation

Even after several decades of quiescent storage in the ovary, the female germ cell is capable of reinitiating transcription to build the reserves that are essential to support early embryonic development. In the current model of mammalian oogenesis, there exists bilateral communication between the gamete and the surrounding cells that is limited to paracrine signaling and direct transfer of small molecules via gap junctions existing at the end of the somatic cells' projections that are in contact with the oolemma. The purpose of this work was to explore the role of cumulus cell projections as a means of conductance of large molecules, including RNA, to the mammalian oocyte. By studying nascent RNA with confocal and transmission electron microscopy in combination with transcript detection, we show that the somatic cells surrounding the fully grown bovine oocyte contribute to the maternal reserves by actively transferring large cargo, including mRNA and long noncoding RNA. This occurrence was further demonstrated by the reconstruction of cumulus-oocyte complexes with transfected cumulus cells transferring a synthetic transcript. We propose selective transfer of transcripts occurs, the delivery of which is supported by a remarkable synapselike vesicular trafficking connection between the cumulus cells and the gamete. This unexpected exogenous contribution to the maternal stores offers a new perspective on the determinants of female fertility.

https://bioone.org/journals/Biology-of-Reproduction/volume-91/issue-4/biolreprod.114.119867/The-Gametic-Synapse-RNA-Transfer-to-the-Bovine-Oocyte1/10.1095/biolreprod.114.119867.pdf

The Gametic Synapse; RNA Transfer to the Bovine Oocyte

Deep kernel learning provides an elegant and principled framework for combining the structural properties of deep learning algorithms with the flexibility of kernel methods. By means of a deep neural network, we learn a parametrized kernel operator that can be combined with a differentiable kernel algorithm during inference. While previous work within this framework has focused on learning a single kernel for large datasets, we learn a kernel family for a variety of few-shot regression tasks. Compared to single deep kernel learning, our algorithm enables the identification of the appropriate kernel for each task during inference. As such, it is well adapted for complex task distributions in a few-shot learning setting, which we demonstrate by comparing against existing state-of-the-art algorithms using real-world, few-shot regression tasks related to the field of drug discovery.

Adaptive Deep Kernel Learning.

Recent work in graph neural networks (GNNs) has led to improvements in molecular activity and property prediction tasks. Unfortunately, GNNs often fail to capture the relative importance of interactions between molecular substructures, in part due to the absence of efficient intermediate pooling steps. To address these issues, we propose LaPool (Laplacian Pooling), a novel, data-driven, and interpretable hierarchical graph pooling method that takes into account both node features and graph structure to improve molecular representation. We benchmark LaPool on molecular graph prediction and understanding tasks and show that it outperforms recent GNNs. Interestingly, LaPool also remains competitive on non-molecular tasks. Both quantitative and qualitative assessments are done to demonstrate LaPool's improved interpretability and highlight its potential benefits in drug design. Finally, we demonstrate LaPool's utility for the generation of valid and novel molecules by incorporating it into an adversarial autoencoder.

Towards Interpretable Sparse Graph Representation Learning with Laplacian Pooling.

Retinal oximetry is a non-invasive technique to investigate the hemodynamics, vasculature and health of the eye. Current techniques for retinal oximetry have been plagued by quantitatively inconsistent measurements and this has greatly limited their adoption in clinical environments. To become clinically relevant oximetry measurements must become reliable and reproducible across studies and locations. To this end, we have developed a convolutional neural network algorithm for multi-wavelength oximetry, showing a greatly improved calculation performance in comparison to previously reported techniques. The algorithm is calibration free, performs sensing of the four main hemoglobin conformations with no prior knowledge of their characteristic absorption spectra and, due to the convolution-based calculation, is invariable to spectral shifting. We show, herein, the dramatic performance improvements in using this algorithm to deduce effective oxygenation (SO2), as well as the added functionality to accurately measure fractional oxygenation ([Formula: see text]). Furthermore, this report compares, for the first time, the relative performance of several previously reported multi-wavelength oximetry algorithms in the face of controlled spectral variations. The improved ability of the algorithm to accurately and independently measure hemoglobin concentrations offers a high potential tool for disease diagnosis and monitoring when applied to retinal spectroscopy.

/pdf/convolutional-neural-networks-for-spectroscopic-analysis-in-3ry4up8kfm.pdf

Convolutional Neural Networks for Spectroscopic Analysis in Retinal Oximetry.

This work introduces a geometric framework and a novel network architecture for creating correspondences between samples of different conditions. Under this formalism, the latent space is a fiber bundle stratified into a base space encoding conditions, and a fiber space encoding the variations within conditions. Furthermore, this latent space is endowed with a natural pull-back metric. The correspondences between conditions are obtained by minimizing an energy functional, resulting in diffeomorphism flows between fibers. 
We illustrate this approach using MNIST and Olivetti and benchmark its performances on the task of batch correction, which is the problem of integrating multiple biological datasets together.

Prudencio Tossou

Papers

The Gametic Synapse; RNA Transfer to the Bovine Oocyte

Adaptive Deep Kernel Learning.

Towards Interpretable Sparse Graph Representation Learning with Laplacian Pooling.

Convolutional Neural Networks for Spectroscopic Analysis in Retinal Oximetry.

Geodesics in fibered latent spaces: A geometric approach to learning correspondences between conditions.