Designing Proteins from Scratch: A New Diffusion Approach
![SeedProteo successfully generated binders-molecules designed to bind to specific protein targets-for notoriously difficult multi-chain proteins, including H1 (as a dimer), VEGF-A (as a dimer), and TNF-[latex]αα[/latex] (as a trimer), demonstrating its capacity to address complex protein interactions and meet predefined computational success criteria.](https://arxiv.org/html/2512.24192v1/x4.png)
Researchers have developed a novel method for creating entirely new protein structures with precise binding capabilities, pushing the boundaries of protein engineering.
Science
Putting Manipulation in Human Hands
Researchers have unveiled a comprehensive ecosystem designed to teach robots how to perform everyday tasks with human-like dexterity.
Bringing AI to Life at the Edge
A new FPGA-accelerated framework unlocks the potential for real-time physical AI by drastically improving energy efficiency and performance on resource-constrained devices.
Beyond Recognition: Building Emotionally Intelligent AI
A new model, MotivNet, demonstrates robust facial expression analysis by leveraging a powerful foundation model to achieve generalization without complex training techniques.
Tiny Robots, Smart Moves: Reinforcement Learning Under Power Limits
![The development of a single-chip micro mote, designated SC[latex]\mu\mu[/latex]M, alongside silicon-on-insulator (SOI)-based microrobots, demonstrates a progression toward increasingly miniaturized and integrated systems-a natural consequence of technological decay and refinement rather than a quest for perpetual novelty.](https://arxiv.org/html/2512.24740v1/figures/scum_robot.png)
Researchers demonstrate how reinforcement learning can enable robust control of microrobots even with severely constrained on-board computing resources.
Sharper Science: Polishing Abstracts for Better Insights
A new language model aims to declutter scientific abstracts, improving the accuracy and efficiency of research analysis.
Thinking Recommenders: Bridging AI and Cognition for Explainable Choices
A new approach combines the power of large language models with a cognitive architecture to create recommendation systems that not only suggest items but also explain why.
Beyond Keywords: A New Approach to Document Understanding
Researchers have developed an algorithm that mimics human-level accuracy in classifying even highly similar documents using only a single example per category.
Twisting the Rules: How Chirality Impacts Actin Filament Mechanics
![A computational model demonstrates that helical actin filaments emerge from the interplay of local mechanical rules-specifically, torque applied to interconnected protofilaments-and are characterized by torsional rigidity [latex]K_{\tau}[/latex], bending persistence length [latex]L_P[/latex], and inter-protofilament separation rigidity [latex]K_S[/latex], as validated through Cytosim simulations.](https://arxiv.org/html/2512.24154v1/x1.png)
New computational modeling reveals the crucial role of helical structure in determining the mechanical behavior of actin filaments, key components of the cell’s internal scaffolding.
Decoding the Signals: Linking Lung Scan Data to Cancer Risk
Researchers have created a comprehensive dictionary that connects quantitative image features from lung scans with established radiological assessments, paving the way for more transparent and reliable AI-driven cancer screening.