Large language models have been widely adopted in natural language processing, yet they face the challenge of generating unreliable content. Recent works aim to reduce misinformation and hallucinations by resorting to attribution as a means to provide evidence (i.e., citations). However, current attribution methods usually focus on the retrieval stage and automatic evaluation that neglect mirroring the citation mechanisms in human scholarly writing to bolster credibility. In this paper, we address these challenges by modelling the attribution task as preference learning and introducing an Automatic Preference Optimization (APO) framework. First, we create a curated collection for post-training with 6,330 examples by collecting and filtering from existing datasets. Second, considering the high cost of labelling preference data, we further propose an automatic method to synthesize attribution preference data resulting in 95,263 pairs. Moreover, inspired by the human citation process, we further propose a progressive preference optimization method by leveraging fine-grained information. Extensive experiments on three datasets (i.e., ASQA, StrategyQA, and ELI5) demonstrate that APO achieves state-of-the-art citation F1 with higher answer quality.

Police in South Wales have arrested a man using automatic facial recognition software. It's the first time a person has been seized this way in the UK, according to Wales Online, following a series of trials at large-scale public events including Download music festival and Notting Hill Carnival. The most recent was the Champions League final in Cardiff, which took place last Saturday (June 3rd). The man, however, was arrested three days beforehand (May 31st). In a statement to Ars Technica UK, police confirmed he was a local resident and "unconnected" to the game in Cardiff.

When thousands of football fans pour into Cardiff's Principality Stadium on June 3 to watch the final match of the UEFA Champions League, few will be aware that their faces will have already been scanned, processed, and compared to a police database of some 500,000 "persons of interest". According to a government tender issued by South Wales Police, the system will be deployed during the day of the game in Cardiff's main train station, as well as in and around the Principality Stadium situated in the heart of Cardiff's central retail district. Cameras will potentially be scanning the faces of an estimated 170,000 visitors plus the many more thousands of people in the vicinity of the bustling Saturday evening city center on match day, June 3. Captured images will then be compared in real time to 500,000 custody images stored in the police information and records management system alerting police to any "persons of interest," according to the tender. The security operation will build on previous police use of Automated Facial Recognition, or AFR technology by London's Metropolitan Police during 2016's Notting Hill Carnival.

Police in Wales plan to use facial recognition on fans during the Champions League final in Cardiff on 3 June, according to a government contract posted online. Faces will be scanned at the Principality Stadium and Cardiff's central railway station. They can then be matched against 500,000 "custody images" stored by local police forces. South Wales Police confirmed the pilot and said it was a "unique opportunity". A report on the plan was first posted by tech news site Motherboard.

If you're headed to the UEFA Champions League final in Cardiff on June 3rd, you might just be part of a massive experiment in security -- and a privacy uproar. South Wales Police are conducting a face recognition trial that could scan every one of the 170,000 visitors expected to show up in the city for the match, whether or not they're heading to the stadium. Cameras around both the stadium and Cardiff's main train station will compare faces against a police database of 500,000 people of interest. If there's a match, police will get a heads-up that could help them stop a terrorist or frequent hooligan. The UK's surveillance camera commissioner, Tony Porter, tells Motherboard that the South Wales Police will have to honor the country's usage guidelines.