Machine learning methods are best suited to catch liars, according to new science of deception detection
Peer-Reviewed Publication
Updates every hour. Last Updated: 6-Jul-2025 03:10 ET (6-Jul-2025 07:10 GMT/UTC)
Detecting lies and deception has always challenged scientists and authorities, but AI-based lie detection devices may have the potential to rise to the challenge.
Designed proteins are anticipated to have groundbreaking impact on a range of issues from treating disease to tackling environmental problems. With a DKK 700 million grant from the Novo Nordisk Foundation and headed by Professor Dek Woolfson, a new Center for Protein Design (CPD) at the University of Copenhagen has ambitions to match this potential. The CPD will spearhead developments in protein design and its applications through strong interdisciplinary collaborations across the university and partnerships in Denmark and internationally.
A research team led by Associate Professor Yasushi Segawa, graduate students Mai Nagase (at the time of the research) and Rui Yoshida, and technical staff member Sachiko Nakano of the Institute for Molecular Science (IMS) and SOKENDAI (The Graduate University for Advanced Studies), together with Associate Professor Takashi Hirose of Kyoto University's Institute for Chemical Research, has synthesized three-dimensionally shaped molecules containing an internal twist and shown that they possess the properties of organic semiconductors. By introducing methyl groups into a planar molecule containing several thiophene units and forcing it into a twisted conformation, the team created a solid-state structure in which electricity can flow three-dimensionally. The molecule was verified to act as an organic semiconductor in an organic field-effect transistor, paving the way for next-generation electronic devices.
These results were published online in the Royal Society of Chemistry journal Chemical Communications on 19 June 2025.
This work marks the first practical use of boson sampling, long seen as a key demonstration of quantum computing’s potential to outperform classical methods.
The researchers used computer simulations to model a quantum optical experiment that recognizes images using just three photons, successfully identifying images from several well-known datasets.
This paves the way towards future applications of quantum AI in complex image recognition, and represents a step toward low-resource, energy-efficient quantum computing.