Body height and weight estimation from a single non-frontal face image suffers from poor performance due to large face pose variance and lack of labeled data. To solve the problems, a research team led by Shiguang SHAN published their new research in Frontiers of Computer Science co-published by Higher Education Press and Springer Nature.

Using liquid-phase transmission electron microscopy, this study observed intrinsically disordered proteins/regions (IDPs/IDRs) transition from single molecules to oligomers, networked clusters, and a dense phase above a concentration threshold. Driven by conformational changes, this dynamic dense phase can disassemble into oligomers/monomers and reform. This provides direct experimental evidence for a non-classical, multi-step nucleation mechanism in early liquid-liquid phase separation.

Deep neural networks (DNNs) are demonstrated to be vulnerable to adversarial examples. Adversarial training is mainstrem method to improve adversarial robustness of DNNs, which augments the training set with adversarial examples and adopts adversarial regularization loss to improve the robustness of DNNs. Existing adversarial training methods are facing the challenge to balance the accuracy and robustness.

Chiral objects are important in fields like drug development because their interactions with biological systems depend on handedness. Researchers from Tokyo University of Science, Institute for Molecular Science, and Seoul National University used light to selectively manipulate nanosized chiral objects. By confining circularly polarized light in an ultra-thin optical fiber, they transported metallic chiral nanoparticles based on handedness, thereby enabling chirality control at near-molecular scales, opening possibilities for applications involving chiral systems, including drug development.

Biodegradable plastics hold potential for reducing marine plastic pollution, but degrade too quickly, limiting their practical use. Researchers from Gunma University now show that crab shell by-products can reduce the breakdown rate of biodegradable plastics in seawater by altering the microbial communities that colonize their surfaces, known as the plastisphere. These findings could help design plastics that stay durable during use and then degrade at an appropriate time once in the ocean.

Automatic link establishment (ALE) was developed to establish a simplified connection in high-frequency (HF) radio system. As a family of tweakable block ciphers, HALFLOOP was standardized by the U.S. Department of Defense to protect the transmitted messages during ALE. According to existing attacks of HALFLOOP, there is a pending search for stronger non-generic attacks against the structure of ciphers itself. To solve the problems, a research team led by Shaozhen CHEN published their new progress on 15 April 2026 in Frontiers of Computer Science co-published by Higher Education Press and Springer Nature.

The escalating volume of transactions on blockchains has made signature verification a critical performance bottleneck. Existing approaches of packing signatures inside SNARK often incur significant computational overhead, demanding substantial time and memory resources for proof generation. This not only hinders the participation of resource-constrained devices but also limits the scalability and accessibility of blockchain systems.

Understanding the dynamics of cyber threats is crucial for today's digital defenses. Researchers from Nanjing University, Jiangsu University of Science and Technology, and Southeast University have discovered that optimizing the execution time of malware in sandbox environments can significantly enhance the completeness and quality of cyber threat intelligence (CTI) data.