Materials Informatics– category –
-
CrystalJev: Calibrated Probabilities for Materials Discovery 2026
arXiv USA Overview CrystalJev introduces a System-1 decision layer for atomistic foundation models that uses calibrated probabilities to determine when to escalate candidates to slower, costlier computations. In a prospective test of 2,1... -
MatterGen-e⁻: AI Material Design via Electronic Structure Prediction
Bioengineer.org / npj Computational Materials USA Overview Researchers at the University of Tokyo have developed MatterGen-e⁻, a generative AI model that predicts electronic behavior alongside atomic arrangement to design new materials. ... -
Quantinuum Helios: Fault-Tolerant Performance vs Superconducting Systems
Quantinuum USA Overview Quantinuum’s Helios trapped-ion system demonstrated a hardware error rate approximately 10x lower than superconducting processors in tests of fault-tolerant operations. The system's QCCD architecture also enabled ... -
High-Entropy Alloys: Data-Driven Framework for Heat Resistance
Results in Engineering International Overview Researchers from the University of Tehran have developed a physics-informed framework to map the composition of high-entropy alloys (HEAs) against high-temperature oxidation and molten-salt h... -
Materials Informatics Weekly Report October 11, 2026
▼ ▼ ▼ If the infographic makes you want to read the full report, please click the download button below. ▼ ▼ ▼ 📄 Weekly Report October 11, 2026 (PDF) — Download Weekly Report October 11, 2026 (PDF) — DownloadDownload 🎙 Podcast October 1... -
PNNL U-Net model: Residual stress distribution for parts explained
PNNL (Pacific Northwest National Laboratory) (The International Journal of Advanced Manufacturing Technology) USA Overview Research from PNNL proposes a machine learning (ML) based Residual Stress Generator (RSG) to infer full-field resi... -
Molecular crystal prediction: 3-stage generative process specs
arXiv Unknown Overview A new three-stage generative process for molecular crystal structure prediction (MCSP) has been introduced, learning invariant lattice descriptors from molecular graphs, Hall settings, and the number of molecules i... -
Elements Claw AI: Alibaba DAMO Academy’s 4 new superconductors
Startup Fortune USA Overview Alibaba DAMO Academy, in collaboration with Renmin University and University of Chinese Academy of Sciences, announced that their AI agent 'Elements Claw' successfully screened 2.4 million candidate crystal s... -
Claude Opus 4.7: Multi-agent LLM physics model success rates
TUHH Open Research Germany Overview A novel multi-agent LLM-driven approach has been introduced for generating constitutive models that describe material deformation under load. By employing a Creator agent to propose data-matched models... -
KAIST MoS2 technology: 2D semiconductor position control specs
BigGo Finance South Korea Overview A joint research team led by KAIST in South Korea has developed a technology to precisely control the formation location of 2D semiconductor crystals (molybdenum disulfide, MoS₂) with over 99% success r...
