British Prime Minister Keir Starmer has discussed the war in Ukraine with leaders of the United States, Italy, France and Germany, a 10 Downing Street spokesperson has said, in advance of US President Donald Trump's planned call with his Russian counterpart, Vladimir Putin, on Monday. The flurry of diplomacy comes shortly after inconclusive direct Russia-Ukraine talks in Istanbul, Turkiye on Friday. The leaders discussed the need for an unconditional ceasefire and for Putin to take peace talks seriously, the spokesperson said late on Sunday, adding that they also raised the use of sanctions if Russia failed to engage seriously in a ceasefire and concerted peace talks. In remarks to reporters earlier on Sunday, German Chancellor Friedrich Merz said he discussed the issue with US Secretary of State Marco Rubio while the two men were attending the inaugural mass of Pope Leo XIV at the Vatican. Merz said he also spoke at length at the Vatican with Ukraine's President Volodymyr Zelenskyy.

As Brian Chesky tells it, the reinvention of Airbnb started with the coup at OpenAI. On November 17, 2023, the board of OpenAI fired company CEO Sam Altman. His friend Chesky leapt into action--publicly defending his pal on X, getting on the phone with Microsoft's CEO, and throwing himself into the thick of Altman's battle to retake OpenAI. Five days later Altman prevailed, and Chesky--"I was so jacked up," he says--turned his buzzing mind to his own company, Airbnb. The Chesky extended family had already held their turkey get-together a week earlier, and the Airbnb CEO had no holiday plan.

In this interview series, we're meeting some of the AAAI/SIGAI Doctoral Consortium participants to find out more about their research. Onur Boyar is a PhD student at Nagoya university, working on generative models and Bayesian methods for materials and drug design. We met Onur to find out more about his research projects, methodology, and collaborations with chemists. I'm from Turkey, and I came to Japan three years ago to pursue my PhD. Before coming here, I was already interested in generative models, Bayesian methods, and Markov chain Monte Carlo techniques.

Earthquakes cause harm not only through direct ground shaking but also by triggering secondary ground failures such as landslides and liquefaction. These combined effects lead to devastating consequences, including structural damage and human casualties. A striking illustration is the 2021 Haiti earthquake, which initiated over 7,000 landslides covering more than 80 square kilometers. This catastrophic event resulted in damage or destruction to over 130,000 buildings, claimed 2,248 lives, and left more than 12,200 people injured [1]. Rapidly identifying where and how severely ground failures and structural damage have occurred following an earthquake is essential for effective victim rescue operations within the crucial "Golden 72 Hour" window, and plays a vital role in developing effective post-disaster recovery plans [2, 3]. Over the years, researchers have developed various approaches for estimating the location and intensity of earthquake-induced ground failures and building damage.

With anti-government protests sweeping across Turkey, the authorities have used all technological means to try to curb them, from restricting internet access to using facial recognition to identify protesters, who have been forced to adapt. Amid a ban on protests, nearly 2,000 people have been arrested in connection with the demonstrations that erupted on March 19 following the detention of Istanbul's mayor Ekrem Imamoglu on graft charges. As well as those apprehended in the streets, many others have been arrested in predawn raids at their homes after being identified from footage or photos taken by the police during the demonstrations.

Muslims around the world have begun celebrating Eid al-Fitr, one of the biggest celebrations in the Islamic calendar. Eid al-Fitr - which means "festival of the breaking of the fast" - is celebrated at the end of Ramadan, a month of fasting for many adults, as well as spiritual reflection and prayer.ReutersHere in Moscow, worshippers are seen preparing for prayer.ReutersHundreds took part in prayers at Tononoka grounds, in Mombasa, KenyaGetty ImagesPrayers were also observed at a stadium in Port Sudan in the east of the countryGetty ImagesLittle children joined adults at the Moskee Essalam in Rotterdam, NetherlandsGetty ImagesGifts are handed out to Muslim children in Lviv, Ukraine, as Russia's war on the country continuesReuters Palestinians in Jabaliya in the northern Gaza Strip pray amidst the rubble of a mosque destroyed in the current war between Israel and HamasGetty ImagesFamilies gather at al-Aqsa mosque in Jerusalem - the third holiest site in IslamReutersA boy yawns during prayers at a stadium in QatarEPAMuslims greet each-other at Martim Moniz Square in Lisbon, PortugalGetty ImagesWomen worshippers gather in Burgess Park, London, for an outdoor prayerEPAThere were also worshippers gathered outside Plebiscito Square in Naples, ItalyReutersSome women took pictures after attending prayers at the Hagia Sophia Grand Mosque in Istanbul, TurkeyGetty ImagesAfghan refugees pray at a mosque on the outskirts of Peshawar, PakistanMiddle EastEuropeEid al-FitrReligionIslamRelated'I was afraid for my life': At the scene of the attack on Palestinian Oscar winner 5 days agoMiddle EastMore8 hrs ago'In Bradford, families spend thousands on new clothes for Eid' Muslims spend large amounts in Bradford's supermarkets, clothes shops and other services before Eid.8 hrs agoEngland1 day ago The tourist has received an award from the city's mayor after restraining a man during a stabbing.1 day agoEurope1 day ago Another 21 people are injured, as a restaurant and several buildings are set ablaze in the city, local officials say.1 day agoWorld1 day ago Town's successful Ramadan lights project expanded A Scunthorpe community group says it has seen an "amazing" response to its lights display.1 day agoLincolnshire1 day ago Bishop says school that changed Easter events'valued' The BBC is not responsible for the content of external sites.

One way to address safety risks from large language models (LLMs) is to censor dangerous knowledge from their training data. While this removes the explicit information, implicit information can remain scattered across various training documents. Could an LLM infer the censored knowledge by piecing together these implicit hints? As a step towards answering this question, we study inductive out-of-context reasoning (OOCR), a type of generalization in which LLMs infer latent information from evidence distributed across training documents and apply it to downstream tasks without in-context learning. Using a suite of five tasks, we demonstrate that frontier LLMs can perform inductive OOCR. In one experiment we finetune an LLM on a corpus consisting only of distances between an unknown city and other known cities. Remarkably, without in-context examples or Chain of Thought, the LLM can verbalize that the unknown city is Paris and use this fact to answer downstream questions. Further experiments show that LLMs trained only on individual coin flip outcomes can verbalize whether the coin is biased, and those trained only on pairs (x, f(x)) can articulate a definition of f and compute inverses. While OOCR succeeds in a range of cases, we also show that it is unreliable, particularly for smaller LLMs learning complex structures. Overall, the ability of LLMs to "connect the dots" without explicit in-context learning poses a potential obstacle to monitoring and controlling the knowledge acquired by LLMs.

Diplomacy is one of the most sophisticated activities in human society, involving complex interactions among multiple parties that require skills in social reasoning, negotiation, and long-term strategic planning. Previous AI agents have demonstrated their ability to handle multi-step games and large action spaces in multi-agent tasks. However, diplomacy involves a staggering magnitude of decision spaces, especially considering the negotiation stage required. While recent agents based on large language models (LLMs) have shown potential in various applications, they still struggle with extended planning periods in complex multi-agent settings. Leveraging recent technologies for LLM-based agents, we aim to explore AI's potential to create a human-like agent capable of executing comprehensive multi-agent missions by integrating three fundamental capabilities: 1) strategic planning with memory and reflection; 2) goaloriented negotiation with social reasoning; and 3) augmenting memory through self-play games for self-evolution without human in the loop.

In this technical supplement, we provide detailed insights and additional results to support our main paper. Section A.1 outlines the generation process of the SynRS3D dataset, including the tools and plugins used. It also covers the licenses for these plugins. Section A.3 elaborates on the evaluation metrics for different tasks, including the proposed F Section A.4 describes the experimental setup and the selection of hyperparameters for the RS3DAda method. Section A.5 presents the ablation study results and analysis for the RS3DAda method. Section A.6 provides supplementary experimental results combining SynRS3D and real data scenarios, complementing Section 5.2 of the main paper. Section A.9 highlights the performance of models trained on the SynRS3D dataset using RS3DAda in the critical application of disaster mapping in remote sensing. A.1 Detailed Generation Workflow of SynRS3D The generation workflow of SynRS3D involves several key steps, from initializing sensor and sunlight parameters to generating the layout, geometry, and textures of the scene. This comprehensive process ensures that the generated SynRS3D mimics real-world remote sensing scenarios with high fidelity. The main steps of the workflow are as follows: Initialization: Set up the sensor and sunlight parameters using uniform and normal distributions to simulate various conditions. Layout Generation: Define the grid and terrain parameters to create diverse urban and natural environments. Texture Generation: Use advanced models like GPT-4 [1] and Stable Diffusion [18] to generate realistic textures for different categories of land cover.

Global semantic 3D understanding from single-view high-resolution remote sensing (RS) imagery is crucial for Earth observation (EO). However, this task faces significant challenges due to the high costs of annotations and data collection, as well as geographically restricted data availability. To address these challenges, synthetic data offer a promising solution by being unrestricted and automatically annotatable, thus enabling the provision of large and diverse datasets. We develop a specialized synthetic data generation pipeline for EO and introduce SynRS3D, the largest synthetic RS dataset. SynRS3D comprises 69,667 high-resolution optical images that cover six different city styles worldwide and feature eight land cover types, precise height information, and building change masks. To further enhance its utility, we develop a novel multi-task unsupervised domain adaptation (UDA) method, RS3DAda, coupled with our synthetic dataset, which facilitates the RS-specific transition from synthetic to real scenarios for land cover mapping and height estimation tasks, ultimately enabling global monocular 3D semantic understanding based on synthetic data. Extensive experiments on various real-world datasets demonstrate the adaptability and effectiveness of our synthetic dataset and the proposed RS3DAda method. SynRS3D and related codes are available at https://github.com/JTRNEO/SynRS3D.