Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Identifying promising synthesis targets and designing routes to their synthesis is a grand challenge in chemistry and materials science. Recent work employing machine learning in combination with traditional approaches is opening new ways to address this truly Herculean task.
Identifying pleiotropic associations for rare variants in multi-ethnic biobank-scale whole-genome sequencing data poses considerable challenges. This study introduced MultiSTAAR as a scalable and robust multi-trait rare variant analysis framework designed for both coding and noncoding regions by integrating multiple variant functional annotations and leveraging multivariate modeling across diverse phenotypes.
We present Spatial Modeling Algorithms for Reactions and Transport (SMART), a software package that simulates spatiotemporally detailed biochemical reaction networks within realistic cellular and subcellular geometries. This paper highlights the use of SMART in several biological test cases including cellular mechanotransduction, calcium signaling in neurons and cardiomyocytes, and adenosine triphosphate synthesis.
A recent study demonstrates through numerical simulations that implementing large language models based on sparse mixture-of-experts architectures on 3D in-memory computing technologies can substantially reduce energy consumption.
By combining several probabilistic AI algorithms, a recent study demonstrates experimentally that the inherent noise and variation in memristor nanodevices can be exploited as features for energy-efficient on-chip learning.
To achieve an advanced neuromorphic computing system with brain-like energy efficiency and generalization capabilities, we propose a hardwareâsoftware co-design of in-memory reservoir computing. This co-design integrates a liquid state machine-based encoder with artificial neural network projections on a hybrid analogâdigital system, demonstrating zero-shot learning for multimodal event data.
We present Morpho, an extensible programmable environment that uses finite elements for shape optimization in soft matter. Given an energy functional that incorporates physical boundaries and effects such as elasticity and electromagnetism, together with additional constraints to be satisfied, Morpho predicts the optimized shape and structure adopted by the material.
An extensive audit of large language models reveals that numerous models mirror the âus versus themâ thinking seen in human behavior. These social prejudices are likely captured from the biased contents of the training data.
Todayâs high-performance computing systems are nearing an ability to simulate the human brain at scale. This presents a new challenge: going forward, will the bigger challenge be the brainâs size or its complexity?
We created an open-source model that simulates Caenorhabditis elegans in a closed-loop system, by integrating simulations of its brain, its physical body, and its environment. BAAIWorm replicated C. elegans locomotive behaviors, and synthetic perturbations of synaptic connections impacted neural control of movement and affected the embodied motor behavior.
Inspired by recent approaches for natural language processing and computer vision, we developed Annotatability, a framework that analyzes deep neural network training dynamics to interpret pre-annotated single-cell and spatial omics data. Annotatability identified erroneous annotations and ambiguous cell states, inferred trajectories from binary labels, and revealed underlying biological signals.
By developing an efficient spin symmetry penalty, a recent study has substantially accelerated the calculation of accurate energies with correct spin states in variational Monte Carlo for both ground and excited states of quantum many-particle systems.
A recent study proposes DeepBlock, a deep learning-based approach for generating ligands with targeted properties, such as low toxicity and high affinity with the given target. This approach outperforms existing methods in the field while maintaining synthetic accessibility and drug-likeness.
A recent study has modeled and quantified the expected rise in electronic waste due to the increasing deployment of generative artificial intelligence.
A recent study introduces a series of approaches that predict protein fitness and stability after the introduction of mutations. The work focuses on combining different data and pre-training to overcome data scarcity.
As digital data expand exponentially, traditional storage media are becoming less viable, making DNA a promising solution due to its density and durability. In this Perspective, the authors discuss the critical computational challenges associated with in vitro DNA-based data storage.
A method is introduced to compute provable bounds on noise-free quantum expectation values from noisy samples, promising potential applications in quantum optimization and machine learning.
Active machine learning is employed in academia and industry to support drug discovery. A recent study unravels the factors that influence a deep learning modelsâ ability to guide iterative discovery.
The application of machine learning techniques to small-molecule drug discovery has not yet yielded a true leap forward in the field. This Perspective discusses how a renewed focus on data and validation could help unlock machine learningâs potential.
