Article
Robotics
Antoine Richard, Stephanie Aravecchia, Thomas Schillaci, Matthieu Geist, Cedric Pradalier
Summary: This letter demonstrates using Deep RL and Domain Randomization to successfully solve a navigation task in a natural environment, showing the model's ability to adapt and perform well without real world training. Additionally, the RL agent is proven to be more robust, faster, and more accurate compared to other methods.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2021)
Article
Computer Science, Artificial Intelligence
Alina Kloss, Georg Martius, Jeannette Bohg
Summary: This work investigates the advantages of differentiable filters (DFs) over other learning approaches, provides practical guidance, and compares DFs with various underlying filtering algorithms through extensive experiments.
Article
Computer Science, Artificial Intelligence
Sung-Wook Park, Jun-Yeong Kim, Jun Park, Se-Hoon Jung, Chun-Bo Sim
Summary: This paper provides a comprehensive review of the latest transfer learning methods for generative adversarial networks (GANs) and proposes an effective method of fixing some layers of the generator and discriminator to address training issues. The experiment uses StyleGAN and evaluates performance using Frechet Inception Distance (FID), coverage, and density. Results show that the proposed method avoids overfitting and achieves better performance on various datasets compared to existing methods.
APPLIED INTELLIGENCE
(2023)
Review
Immunology
Rory Doolan, Namitha Putananickal, Lucienne Tritten, Tiffany Bouchery
Summary: Soil-transmitted helminths affect approximately 1.5 billion people worldwide, but no human vaccine is currently available. The current strategy for elimination relies on preventive chemotherapy. Traditional vaccine approaches focusing on peptide antigens have had limited success, and mucosal and cellular-based vaccines may be a better alternative in fighting against helminth infection.
FRONTIERS IN IMMUNOLOGY
(2023)
Article
Forestry
Miguel Garcia-Hidalgo, Jose Miguel Olano, Jose Reyes-Lopez, Susana Gomez Martinez, Victor Alonso Gomez, Gabriel Sanguesa-Barreda
Summary: CaptuRING is a reliable and affordable tool that transforms tree-ring samples into digital images using open source software and a DIY approach. It operates through a Raspberry Pi controlled by an Arduino board, capturing high-resolution images. Three video tutorials are available for constructing and installing CaptuRING from scratch.
Article
Mechanics
Pawan Negi, Prabhu Ramachandran
Summary: In this paper, the convergence properties of commonly used solid, inlet, and outlet boundary implementations are verified using the method of manufactured solutions. A convergent WCSPH scheme along with suitable methods for implementing the boundary conditions are proposed based on the convergence offered by these methods. The accuracy of the proposed scheme is demonstrated by solving the flow past a circular cylinder, paving the way for the accurate and efficient simulation of fluid flows using the SPH method.
Article
Robotics
Julian Ibarz, Jie Tan, Chelsea Finn, Mrinal Kalakrishnan, Peter Pastor, Sergey Levine
Summary: Deep reinforcement learning has shown promise in enabling physical robots to learn complex skills in the real world, which presents numerous challenges in perception and movement. Real-world robotics provides a unique domain for evaluating deep RL algorithms, addressing challenges that are often overlooked in mainstream RL research.
INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH
(2021)
Editorial Material
Chemistry, Multidisciplinary
Hai -Na Feng
Summary: Hai-Na Feng is a second-year PhD candidate at East China Normal University. She started her research on molecular nanotopology in the group of Prof. David Leigh and Prof. Liang Zhang. From 2017 to 2019, she worked with Prof. Julius Rebek and Yang Yu at Shanghai University, focusing on the host-guest chemistry of water-soluble cavitand. After completing her master's studies, she joined the current group as a research assistant in 2019. Her current research interest lies in the synthesis of topologically complex molecules and molecularly woven materials.
Article
Astronomy & Astrophysics
Azim Ahmadzadeh, Berkay Aydin, Manolis K. Georgoulis, Dustin J. Kempton, Sushant S. Mahajan, Rafal A. Angryk
Summary: This study presents a case study of solar flare forecasting using metadata feature time series, addressing prominent issues of class-imbalance and temporal coherence. Effective solutions for these challenges are provided, along with discussions on the impact of data manipulation tasks on performance.
ASTROPHYSICAL JOURNAL SUPPLEMENT SERIES
(2021)
Article
Mathematics
Theresa Buchter, Andreas Eichler, Nicole Steib, Karin Binder, Katharina Bocherer-Linder, Stefan Krauss, Markus Vogel
Summary: Bayesian Reasoning, a fundamental idea in probability, is crucial for evaluating uncertain situations. It involves calculating conditional probabilities, assessing parameter changes on results, and explaining formula outcomes, and is particularly important in non-mathematical fields like medicine and law. Developing training courses on Bayesian Reasoning specific to these professions is necessary. Evidence-based research shows that students from medicine and law backgrounds can improve their Bayesian Reasoning skills through such courses, enhancing their professional expertise.
Review
Agriculture, Dairy & Animal Science
Megan C. Edwards, Caitlin Ford, Julia M. Hoy, Sean FitzGibbon, Peter J. Murray
Summary: Many studies have shown that translocation projects using captive or captive-bred animals often face low success rates, leading to the need for pre-release behavioural conditioning to encourage natural behaviours and enhance survival post-release. Predator avoidance training is a common technique used to improve prey responses to predators by pairing predator cues with unpleasant stimuli.
APPLIED ANIMAL BEHAVIOUR SCIENCE
(2021)
Article
Computer Science, Software Engineering
Roland Kaminski, Javier Romero, Torsten Schaub, Philipp Wanko
Summary: ASP, a popular and sophisticated approach to declarative problem solving, is easy to use but difficult to modify due to its underlying technology. This tutorial aims at helping users build their own ASP-based systems by extending ASP using metaprogramming and traditional programming techniques. By showcasing examples and case studies, it illustrates how clingo can be customized and extended for specific purposes.
THEORY AND PRACTICE OF LOGIC PROGRAMMING
(2021)
Article
Mechanics
Pawan Negi, Prabhu Ramachandran
Summary: The paper introduces a method called manufactured solutions (MMS) to comprehensively test the convergence and accuracy of a WCSPH-based solver. Using MMS, one can identify solver problems efficiently and test boundary conditions effectively.
Editorial Material
Chemistry, Multidisciplinary
Aitor Bermejo-Lopez
Summary: Dr. Aitor Bermejo Lopez graduated with a degree in chemistry and obtained a doctoral degree in organic chemistry. His research focuses on the development of new methods in organometallic catalysis and the study of their mechanisms. He recently defended his doctoral thesis and looks forward to his future growth in the scientific field.
Editorial Material
Cell Biology
Odeta Mece, Diede Houbaert, Patrizia Agostinis
Summary: Lymphatic endothelial cells (LECs) rely on fatty acid oxidation (FAO) and the transcription factor PROX1 for growth and maintaining their identity. The loss of ATG5 in LECs prevents injury-induced lymphangiogenesis by impairing the degradation of lipid droplets (LDs) and disrupting mitochondrial fitness. This leads to reduced mitochondrial FAO and acetyl-CoA levels, affecting PROX1-mediated epigenetic regulation and key lymphatic markers. Supplementing with acetate, a fatty acid precursor, rescues defective inflammation-driven lymphangiogenesis in LEC-specific atg5 knockout mice.
Article
Biochemical Research Methods
Carsen Stringer, Tim Wang, Michalis Michaelos, Marius Pachitariu
Summary: Cellpose is a generalist, deep learning-based segmentation method that can precisely segment cells from various types of images without requiring retraining or parameter adjustments. It was trained on a highly diverse image dataset and supports a three-dimensional extension. Software has been developed to support community contributions to the training data.
Article
Biochemistry & Molecular Biology
Carsen Stringer, Michalis Michaelos, Dmitri Tsyboulski, Sarah E. Lindo, Marius Pachitariu
Summary: The study found that individual neurons in the visual cortex provide unreliable estimates of visual features, and it is unknown if single-neuron variability is correlated across large neural populations. By recording and measuring a large number of neurons from the mouse visual cortex, it was discovered that behavioral variability during a sensory discrimination task could not be explained by neural variability in V1.
Article
Neurosciences
Anne-Kathrin Eiselt, Susu Chen, Jim Chen, Jon Arnold, Tahnbee Kim, Marius Pachitariu, Scott M. Sternson
Summary: The study found that mice can confuse hunger and thirst, similar to humans who sometimes overeat due to this confusion. The results show that decision-making in evaluating physiological need states tends to direct individuals towards restoring homeostasis. Instead of being guided by interoceptive knowledge of hunger and thirst states, need states are identified through outcome evaluation after food and water consumption.
NATURE NEUROSCIENCE
(2021)
Article
Multidisciplinary Sciences
Nicholas A. Steinmetz, Cagatay Aydin, Anna Lebedeva, Michael Okun, Marius Pachitariu, Marius Bauza, Maxime Beau, Jai Bhagat, Claudia Bohm, Martijn Broux, Susu Chen, Jennifer Colonell, Richard J. Gardner, Bill Karsh, Fabian Kloosterman, Dimitar Kostadinov, Carolina Mora-Lopez, John O'Callaghan, Junchol Park, Jan Putzeys, Britton Sauerbrei, Rik J. J. van Daal, Abraham Z. Vollan, Shiwei Wang, Marleen Welkenhuysen, Zhiwen Ye, Joshua T. Dudman, Barundeb Dutta, Adam W. Hantman, Kenneth D. Harris, Albert K. Lee, Edvard Moser, John O'Keefe, Alfonso Renart, Karel Svoboda, Michael Hausser, Sebastian Haesler, Matteo Carandini, Timothy D. Harris
Summary: The research introduces the Neuropixels 2.0 probe and newly designed analysis algorithms, enabling high-quality recordings for long time scales in small mammals. Automatic post hoc correction for brain movements allows recording from the same neurons for more than 2 months. These probes and algorithms enable stable recordings from thousands of sites during free behavior.
Editorial Material
Behavioral Sciences
Tara van Viegen, Athena Akrami, Kathryn Bonnen, Eric DeWitt, Alexandre Hyafil, Helena Ledmyr, Grace W. Lindsay, Patrick Mineault, John D. Murray, Xaq Pitkow, Aina Puce, Madineh Sedigh-Sarvestani, Carsen Stringer, Titipat Achakulvisut, Elnaz Alikarami, Melvin Selim Atay, Eleanor Batty, Jeffrey C. Erlich, Byron V. Galbraith, Yueqi Guo, Ashley L. Juavinett, Matthew R. Krause, Songting Li, Marius Pachitariu, Elizabeth Straley, Davide Valeriani, Emma Vaughan, Maryam Vaziri-Pashkam, Michael L. Waskom, Gunna Blohm, Konra Kording, Paul Schrater, Brad Wyble, Sean Escola, Megan A. K. Peters
Summary: Neuromatch Academy (NMA) successfully organized an online Computational Neuroscience Summer School with 1757 students and 191 teaching assistants, promoting inclusivity and universal accessibility through active community management and low cost.
TRENDS IN COGNITIVE SCIENCES
(2021)
Article
Multidisciplinary Sciences
Kyu Hyun Lee, Yu-Li Ni, Jennifer Colonell, Bill Karsh, Jan Putzeys, Marius Pachitariu, Timothy D. Harris, Markus Meister
Summary: State-of-the-art silicon probes for electrical recording from neurons have thousands of recording sites, but typically have fewer wires carrying signals off the probe, limiting the number of channels that can be recorded simultaneously. The authors propose electrode pooling, using a single wire to serve multiple channels through controllable switches, as a solution to this limitation.
NATURE COMMUNICATIONS
(2021)
Article
Neurosciences
Edward Zagha, Jeffrey C. Erlich, Soohyun Lee, Gyorgy Lur, Daniel H. O'Connor, Nicholas A. Steinmetz, Carsen Stringer, Hongdian Yang
Summary: Recent studies have found that movement-related activity is present throughout the mouse brain, including early sensory areas. Failing to consider movement when interpreting neuronal function may lead to misattributing activity to other processes. Functional couplings between movement and other activities make it difficult to fully isolate sensory, motor, and cognitive-related activity.
JOURNAL OF NEUROSCIENCE
(2022)
Review
Neurosciences
Lilach Avitan, Carsen Stringer
Summary: Sensory areas exhibit spontaneous activity in the absence of sensory stimuli, and recent studies have revealed high-dimensional patterns, correlation with behavior, and dissimilarity with sensory-driven activity. These findings suggest a new role for spontaneous activity in neural sensory computation.
Article
Biochemical Research Methods
Kevin J. Cutler, Carsen Stringer, Teresa W. Lo, Luca Rappez, Nicholas Stroustrup, S. Brook Peterson, Paul A. Wiggins, Joseph D. Mougous
Summary: This paper presents a deep neural network image-segmentation algorithm called Omnipose, which achieves accurate segmentation performance on various types of cells, including bacteria and non-bacterial subjects, using different imaging modalities and three-dimensional objects. Omnipose is especially useful for characterizing cells with extreme morphological phenotypes.
Article
Biochemical Research Methods
Biraj Pandey, Marius Pachitariu, Bingni W. Brunton, Kameron Decker Harris
Summary: This study models the receptive fields of sensory neurons in a way that incorporates randomness and connects to the theory of artificial neural networks. The models enhance signal and remove noise, enabling more efficient learning in artificial tasks. This research has significance for both neuroscience and machine learning communities.
PLOS COMPUTATIONAL BIOLOGY
(2022)
Article
Neurosciences
Atika Syeda, Lin Zhong, Renee Tung, Will Long, Marius Pachitariu, Carsen Stringer
Summary: Recent studies in mice have found that orofacial behaviors have a significant impact on neural activity in the brain. To better understand these signals and their functions, researchers developed Facemap, a framework consisting of a keypoint tracker and a deep neural network encoder. The algorithm for tracking mouse orofacial behaviors was more accurate and faster than existing tools, making it useful for real-time experimental interventions. Using the deep neural network, they were able to predict the activity of thousands of neurons and found that neural activity clusters related to behavior were more spread out across the cortex. They also discovered that the behavioral features had sequential dynamics that were irreversible in time. Facemap provides a stepping stone towards understanding the relationship between neural signals and behavior.
NATURE NEUROSCIENCE
(2023)