A 32x32 array of hafnium diselenide memristors paired with silicon selectors, published in Nature Communications 2025, stores and processes data at the same location — eliminating the need for constant data transfers between memory and computing units and cutting AI power use by more than half. Targeted at AI-based edge computing and autonomous systems.
A ferroelectric tunnel-junction memory chip defies the decades-old assumption that chips degrade as they shrink — measured performance actually improves at extreme miniaturization.
Axiom Math's formal verification system, which writes its proofs in the Lean open-source formal programming language, flagged a gap in the foundations of Aumann's 1976 Common Knowledge Theorem — a result foundational to information economics. The system surfaced an assumption Aumann stated but never actually proved. Scott Kominers, who is co-leading the EconLib effort with Axiom's team, examined the finding. The implications reach into the foundations of models used to derive metrics written into US antitrust merger guidelines.
A preregistered behavioral implementation of Newcomb's paradox with 1,305 participants found that framing a predictor as AI increased the odds of forgoing the guaranteed reward by a factor of 3.39 (95% CI: 2.45–4.70) compared with random framing, reducing earnings by 10.7–42.9%. Over 40% of participants treated AI as a predictive authority, and the effect persisted even when predictions failed.
McMaster University's SyntheMol-RL used reinforcement learning to explore a space of 46 billion possible compounds built from roughly 150,000 molecular building blocks, inventing rather than screening molecules. From a batch of 79 model-proposed antibacterials, the system produced synthecin — a water-soluble candidate effective against drug-resistant S. aureus in mouse wound models.
Columbia/MIT/Harvard team used AI protein-language models to redesign all 52 essential ribosomal proteins in E. coli without isoleucine, creating strain Ec19 that stayed above 90% wild-type fitness for 450 generations.
University of Hong Kong engineers demonstrated a world-first silicon carbide transistor that mimics neuron spiking at 10 millikelvin, the temperature range of quantum computers.
Princeton/Flatiron study shows transfer learning cuts cosmological AI simulation costs 10x but risks negative transfer that masks genuinely new physics.
Sandia computational neuroscientists Brad Theilman and Brad Aimone developed a new algorithm enabling neuromorphic hardware to tackle partial differential equations — the mathematical foundation of fluid dynamics, electromagnetic fields, and structural mechanics.
UPenn physicists created hybrid light-matter particles (exciton-polaritons) that perform all-optical switching for AI chips using only about 4 quadrillionths of a joule per operation — published in Physical Review Letters, May 2026.
Astronomers at the University of Warwick have confirmed more than 100 exoplanets, including 31 newly identified worlds, using a new artificial intelligence system applied to NASA's Transiting Exoplanet Survey Satellite (TESS) data — among them rare ultra-short-period giants in the so-called “Neptunian desert,” a region where few planets are expected to exist.
Tufts University researchers combined neural networks with human-like symbolic reasoning to create a hybrid AI that achieved 95% success on a Tower of Hanoi robotic task vs. 34% for standard models, while using only 1% of the training energy required by a standard VLA model.