Stanford AI researchers, along with colleagues from Harvard University, have published a new paper showing that many risk models may not be all they are cracked up to be because they have too much data. Oftentimes risk models measure things indirectly using proxies and the use of inapt proxies leads to a research phenomenon known as label bias.
A Stanford team is developing a new strategy for selectively converting and long-term storing of electrical energy in liquid fuels. "We also discovered a novel, selective catalytic system for storing electrical energy in a liquid fuel without generating gaseous hydrogen," said Robert Waymouth, professor in chemistry and senior author of this study.
The Center for AEroSpace Autonomy Research, or CAESAR, is a new center at Stanford Engineering which will leverage artificial intelligence for space science, exploration and business. CAESAR co-founders, Marco Pavone and Simone D'Amico aim to develop a foundation model for space
Researchers at SLAC developed new methods to produce intense attosecond (or billionths of a billionth of a second) pulses and pulse pairs to gain insights into the fastest motions inside atoms and molecules. Observing atoms and electrons in motion facilitates the design of new materials with tailored properties for technology, energy, and other fields.
SLAC is taking part in a multi-institutional effort to finding transformative paths to manage carbon in difficult-to-electrify industries. Bringing sustainable sources of carbon into the carbon-circle economy is essential to the future where carbon is reused and treated as a precious commodity rather than a disposable resource.
Mobile ALOHA, the robot seen at left, learns through imitation learning by running through the desired action about 50 times to help it learn that task independently. Such co-training can increase success rates by up to 90% allowing Mobile ALOHA to autonomously complete complex mobile manipulation tasks.
The NSF and the Department of Energy awarded grants to research teams as part of a National Artificial Intelligence Research pilot. Two Stanford AI projects - from the School of Engineering and School of Medicine - were selected for the pilot. The engineering project will focus on learning effective reward functions for robotics using large datasets and human feedback.
Stanford researchers have developed a prototype augmented reality headset that uses holographic imaging to overlay full-color, 3D moving images on the lenses of what would appear to be an ordinary pair of glasses. To succeed, the researchers have overcome technical barriers through a combination of AI-enhanced holographic imaging and new nanophotonic device approaches.
Researchers at Stanford University have introduced a more efficient processing technique that can print up to 1 million highly detailed and customizable microscale particles a day. These 3D-printed microscopic particles are so small that to the naked eye they look like dust, and have applications in drug and vaccine delivery, microelectronics, microfluidics, and abrasives for intricate manufacturing.
A team of scientists, seen here at left, has developed a groundbreaking method that harnesses the structure of light to twist and tweak the properties of quantum materials. This flexibility could allow the method to work for a broad range of applications, making it easier to develop new technologies. That could, for example, lead to the development of super-fast switches for quantum computers, which could drastically outperform the computers we use today.
