Research articles

Predicting the structure of proteins from amino acid sequences is a hard problem. Convolutional neural networks can learn to predict a map of distances between amino acid residues that can be turned into a three-dimensional structure. With a combination of approaches, including an evolutionary technique to find the best neural network architecture and a tool to find the atom coordinates in the folded structure, a pipeline for rapid prediction of three-dimensional protein structures is demonstrated.

Number processing is linked to bodily systems, especially finger movements. The authors apply convolutional neural network models in the context of cognitive developmental robotics. They show that proprioceptive information in the child-like robot iCub improves accuracy and recognition of spoken digits.

Haptic interfaces are important for the development of immersive human–machine interactions. To create a compact design with rich touch-sensitive functions, a robotic device called Foldaway, which folds flat, has been designed that can render three-degrees-of-freedom force feedback.

Metals can bind to proteins to fulfil important biological functions. Predicting the features of mutated binding sites can thus help us understand the connection between specific mutations and their role in diseases.

Identifying abnormalities in medical images across different viewing angles and body parts is a time-consuming task. Deep learning techniques hold great promise for supporting radiologists and improving patient triage decisions. A new study tests the viability of such approaches in resource-limited settings, exploring the effect of pretraining, dataset size and choice of deep learning model in the task of abnormality detection in lower-limb radiographs.

Drug combinations are often an effective means of managing complex diseases, but understanding the synergies of drug combinations requires extensive resources. The authors developed an efficient machine learning model that requires only a limited set of pairwise dose–response measurements for the accurate prediction of synergistic and antagonistic drug combinations.

To better extract meaning from natural language, some less informative words can be removed before a model is trained, which is usually done by using manually curated lists of stopwords. A new information theoretic approach can identify uninformative words automatically and more accurately.

Algorithms and bots are capable of performing some behaviours at human or super-human levels. Humans, however, tend to trust algorithms less than they trust other humans. The authors find that bots do better than humans at inducing cooperation in certain human–machine interactions, but only if the bots do not disclose their true nature as artificial.

Human face recognition is robust to changes in viewpoint, illumination, facial expression and appearance. The authors investigated face recognition in deep convolutional neural networks by manipulating the strength of identity information in a face by caricaturing. They found that networks create a highly organized face similarity structure in which identities and images coexist.

Photonic computing devices have been proposed as a high-speed and energy-efficient approach to implementing neural networks. Using off-the-shelf components, Antonik et al. demonstrate a reservoir computer that recognizes different forms of human action from video streams using photonic neural networks.

Deep learning is currently transforming digital pathology, helping to make more reliable and faster clinical diagnoses. A promising application is in the recognition of malignant white blood cells—an essential step for detecting acute myeloid leukaemia that is challenging even for trained human examiners. An annotated image dataset of over 18,000 white blood cells is compiled and used to train a convolutional neural network for leukocyte classification. The network classifies the most important cell types with high accuracy and can answer clinically relevant binary questions with human-level performance.

Deep neural networks can be led to misclassify an image when minute changes that are imperceptible to humans are introduced. While for some networks this ability can cast doubt on the reliability of the model, it also offers explainability for networks that use more robust regularization.

To keep radiation therapy from damaging healthy tissue, expert radiologists have to segment CT scans into individual organs. A new deep learning-based method for delineating organs in the area of head and neck performs faster and more accurately than human experts.

Neural network force fields promise to bypass the computationally expensive quantum mechanical calculations typically required to investigate complex materials, such as lithium-ion batteries. Mailoa et al. accelerate these approaches with an architecture that exploits both rotation-invariant and -covariant features separately.

Optoacoustic imaging can achieve high spatial and temporal resolution but image quality is often compromised by suboptimal data acquisition. A new method employing deep learning to recover high-quality images from sparse or limited-view optoacoustic scans has been developed and demonstrated for whole-body mouse imaging in vivo.

Labelling training data to train machine learning models is very time intense. A new method shows that content transformation can be effectively learned from generated data, avoiding the need for any manual labelling in segmentation and classification tasks.

Brain–machine interfaces using steady-state visually evoked potentials (SSVEPs) show promise in therapeutic applications. With a combination of innovations in flexible and soft electronics and in deep learning approaches to classify potentials from two channels and from any subject, a compact, wireless and universal SSVEP interface is designed. Subjects can operate a wheelchair in real time with eye movements while wearing the new brain–machine interface.

A combination of engineering advances shows promise for myoelectric prosthetic hands that are controlled by a user’s remaining muscle activity. Fine finger movements are decoded from surface electromyograms with machine learning algorithms and this is combined with a robotic controller that is active only during object grasping to assist in maximizing contact. This shared control scheme allows user-controlled movements when high dexterity is desired, but also assisted grasping when robustness is required.

Memristive devices can provide energy-efficient neural network implementations, but they must be tailored to suit different network architectures. Wang et al. develop a trainable weight-sharing mechanism for memristor-based CNNs and ConvLSTMs, achieving a 75% reduction in weights without compromising accuracy.

Controlling the flow and representation of information in deep neural networks is fundamental to making networks intelligible. Bergomi et al introduce a mathematical framework in which the space of possible operators representing the data is constrained by using symmetries. This constrained space is still suitable for machine learning: operators can be efficiently computed, approximated and parameterized for optimization.