Articles

Deep learning is a powerful method to process large datasets, and shown to be useful in many scientific fields, but models are highly parameterized and there are often challenges in interpretation and generalization. David Gleich and colleagues develop a method rooted in computational topology, starting with a graph-based topological representation of the data, to help assess and diagnose predictions from deep learning and other complex prediction methods.

Continual learning is an innate ability in biological intelligence to accommodate real-world changes, but it remains challenging for artificial intelligence. Wang, Zhang and colleagues model key mechanisms of a biological learning system, in particular active forgetting and parallel modularity, to incorporate neuro-inspired adaptability to improve continual learning in artificial intelligence systems.

scBERT, a pretrained neural network for single-cell sequencing tasks, was published last year in Nature Machine Intelligence. To test the reusability of the method, Khan et al. use the code to assess the generalizablility of transformer architectures on single-cell genomics tasks.

Prediction of high-level visual representations in the human brain may benefit from multimodal sources in network training and the incorporation of complex datasets. Wang and colleagues show that language pretraining and a large, diverse dataset together build better models of higher-level visual cortex compared to earlier models.

Graph neural networks have proved useful in modelling proteins and their ligand interactions, but it is not clear whether the patterns they identify have biological relevance or whether interactions are merely memorized. Mastropietro et al. use a Shapley value-based method to identify important edges in protein interaction graphs, enabling explanatory analysis of the model mechanisms.

Halide perovskites are promising materials for light-emitting devices, given their narrowband emission and solution processability. However, detailed information on device degradation during operation is required to improve their stability, and this is challenging to obtain. Ji et al. propose a self-supervised deep learning method to capture multi-dimensional images of such devices in their operating regime faster than allowed by conventional imaging techniques.

The reconstruction of dynamic, spatial fields from sparse sensor data is an important challenge in various fields of science and technology. Santos et al. introduce the Senseiver, a deep learning framework that reconstructs spatial fields from few observations using attention layers to encode and decode sparse data, enabling efficient inference.

Geometric deep learning has become a powerful tool in virtual drug design, but it is not always obvious when a model makes incorrect predictions. Luo and colleagues improve the accuracy of their deep learning model using uncertainty calibration and Bayesian optimization in an active learning cycle.

Human and animal motion planning works at various timescales to allow the completion of complex tasks. Inspired by this natural strategy, Yuan and colleagues present a hierarchical motion planning approach for robotics, using deep reinforcement learning and predictive proprioception.

Organisms show complex behaviour resulting from a trade-off between obtaining information (explore) and using current information (exploit). Biswas et al. observe a mode-switching strategy modulated by sensory salience in a diverse range of organisms, including electric fish and humans, and argue that the observed heuristic could inform the design of active-sensing behaviours in robotics.

Prime editors are innovative genome-editing tools, but selecting guide RNAs with high efficiency remains challenging and requires costly experimental efforts. Liu and colleagues develop a method to design prime-editing guide RNAs based on transfer learning for in silico prediction of editing efficacy.

Learning causal relationships between variables in large datasets is an outstanding challenge in various scientific applications. Lagemann et al. introduce a deep neural network approach combining convolutional and graph models intended for causal learning in high-dimensional biomedical problems.

Deep learning methods in natural language processing generally become more effective with larger datasets and bigger networks. But it is not evident whether the same is true for more specialized domains such as cheminformatics. Frey and colleagues provide empirical explorations of chemistry models and find that neural-scaling laws hold true even for the largest tested models and datasets.

The immense amount of Wikipedia articles makes it challenging for volunteers to ensure that cited sources support the claim they are attached to. Petroni et al. use an information-retrieval model to assist Wikipedia users in improving verifiability.

Identifying unknown peptides in tandem mass spectrometry is challenging as fragmentation of precursor peptides can be incomplete. Mao and colleagues present a method based on graph neural networks and a path-searching model to create more stable sequence predictions.

Computational methods for analysing single 2D tissue slices from spatial transcriptomics studies are well established, but their extension to the 3D domain is challenging. Wang et al. develop a deep learning framework that can perform 3D reconstruction of cellular structures in tissues as well as whole organisms.

Contact prediction between two proteins is still computationally challenging, but is vital for understanding multi-protein complexes. Lin et al. use a geometric deep learning approach to provide accurate predictions of inter-protein residue–residue contacts.

Deconvolution of cell types in tissue proteomic data is a challenging computational task for the bioinformatics community. A deep-learning method termed scpDeconv is introduced that makes efficient use of single-cell proteomics data to deconvolve cell types and states from bulk proteomics measurements.

AlphaFold2 has revolutionized bioinformatics, but its ability to predict protein structures with high accuracy comes at the price of a costly database search for multiple sequence alignments. Fang and colleagues pre-train a large-scale protein language model and use it in conjunction with AlphaFold2 as a fully trainable and efficient model for structure prediction.

It is widely known that AI-based recommendation systems on social media and news websites can isolate humans from diverse information, eventually trapping them in so-called information cocoons, where they are exposed to a narrow range of viewpoints. Li et al. introduce an adaptive information dynamics model to uncover the origin of information cocoons in complex human–AI interaction systems, and test their findings on two large real-world datasets.