The first non-European Pope in more than 1,000 years, the Oscar-winning star of Annie Hall and The Godfather, a soul legend and one of the world's most famous designers - here are some of the well-known faces no longer with us. Among those we remember are Hollywood stars Robert Redford, Diane Keaton and Gene Hackman, and theatrical dames Joan Plowright and Patricia Routledge. Robert Redford's acting career spanned more than 50 films and won him an Oscar as a director. For many filmgoers though, he was simply the best-looking cinema star in the world - once described as a chunk of Mount Rushmore levered into stonewashed denims. As well as leading roles in hits such as All The President's Men, Butch Cassidy and the Sundance Kid and The Way We Were, Redford also launched the Sundance Film Festival to champion independent filmmakers. Los-Angeles-born Keaton shot to fame with her role in The Godfather, but enjoyed a long creative partnership with Woody Allen. Annie Hall, a comedy based on their off-screen relationship, earned her a Best Actress Oscar and they collaborated on several other films. She was nominated for three further Oscars - all in the best actress category - for her work in Something's Gotta Give, Marvin's Room and Reds. BASIL! - the unmistakable sound of Sybil Fawlty admonishing her pompous and incompetent husband, is probably how Prunella Scales will best be remembered. Apart from starring in sitcom Fawlty Towers, she played many other roles on screen and stage, including Queen Elizabeth II in Alan Bennett's play, A Question of Attribution.

We propose a novel hierarchical reinforcement learning framework for quadruped locomotion over challenging terrain. Our approach incorporates a two-layer hierarchy in which a high-level policy (HLP) selects optimal goals for a low-level policy (LLP). The LLP is trained using an on-policy actor-critic RL algorithm and is given footstep placements as goals. We propose an HLP that does not require any additional training or environment samples and instead operates via an online optimization process over the learned value function of the LLP. We demonstrate the benefits of this framework by comparing it with an end-to-end reinforcement learning (RL) approach. We observe improvements in its ability to achieve higher rewards with fewer collisions across an array of different terrains, including terrains more difficult than any encountered during training.

The rapid spread of misinformation in the digital era poses significant challenges to public discourse, necessitating robust and scalable fact-checking solutions. Traditional human-led fact-checking methods, while credible, struggle with the volume and velocity of online content, prompting the integration of automated systems powered by Large Language Models (LLMs). However, existing automated approaches often face limitations, such as handling complex claims, ensuring source credibility, and maintaining transparency. This paper proposes a novel multi-agent system for automated fact-checking that enhances accuracy, efficiency, and explainability. The system comprises four specialized agents: an Input Ingestion Agent for claim decomposition, a Query Generation Agent for formulating targeted subqueries, an Evidence Retrieval Agent for sourcing credible evidence, and a Verdict Prediction Agent for synthesizing veracity judgments with human-interpretable explanations. Evaluated on benchmark datasets (FEVEROUS, HOVER, SciFact), the proposed system achieves a 12.3% improvement in Macro F1-score over baseline methods. The system effectively decomposes complex claims, retrieves reliable evidence from trusted sources, and generates transparent explanations for verification decisions. Our approach contributes to the growing field of automated fact-checking by providing a more accurate, efficient, and transparent verification methodology that aligns with human fact-checking practices while maintaining scalability for real-world applications. Our source code is available at https://github.com/HySonLab/FactAgent

The match offered a front-row seat to how impressively robots can move and react almost like humans. Robot combat just got a lot more interesting in Hangzhou, China. Four Unitree G1 robots, each steered by a human operator, went head-to-head in a tournament called Unitree Iron Fist King: Awakening! The event took place right next to Unitree's massive new factory and drew a lively mix of tech fans and people just curious to see what all the buzz was about. This wasn't only about showing off robotic strength; it gave everyone a front-row seat to how impressively robots can now move and react almost like humans.

In a world where human boxers are at risk of dangerous injuries, we may have a glimpse of what the fight of the future could look like. New footage shows the world's first humanoid robot boxing tournament, which took place over the weekend in Hangzhou, east China. In the bizarre clip, two lifesize robots wearing gloves and protective headgear fight each other in a ring as a human officiator looks on. Each fighter robot weighs about 35kg and is 4.3ft (132cm) tall – roughly the height of the average eight-year-old child. Both the bots initially have trouble seeing exactly where their opponent is before successfully trading punches and kicks, to the delight of a baying crowd.

Breakthroughs, discoveries, and DIY tips sent every weekday. After decades of being tortured, shoved, kicked, burned, and bludgeoned, robots are finally getting their chance to fight back. This weekend, Chinese robotics maker Unitree says it will livestream the world's first boxing match between two of its humanoid robots. The event, titled Unitree Iron Fist King: Awakening, will feature a face-off between two of Unitree's 4.3-foot-tall G1 robots. The robots will reportedly be remotely controlled by human engineers, though they are also expected to demonstrate some autonomous, pre-programmed actions as well.

Current evaluations of defenses against prompt attacks in large language model (LLM) applications often overlook two critical factors: the dynamic nature of adversarial behavior and the usability penalties imposed on legitimate users by restrictive defenses. We propose D-SEC (Dynamic Security Utility Threat Model), which explicitly separates attackers from legitimate users, models multi-step interactions, and rigorously expresses the security-utility in an optimizable form. We further address the shortcomings in existing evaluations by introducing Gandalf, a crowd-sourced, gamified red-teaming platform designed to generate realistic, adaptive attack datasets. Using Gandalf, we collect and release a dataset of 279k prompt attacks. Complemented by benign user data, our analysis reveals the interplay between security and utility, showing that defenses integrated in the LLM (e.g., system prompts) can degrade usability even without blocking requests. We demonstrate that restricted application domains, defense-in-depth, and adaptive defenses are effective strategies for building secure and useful LLM applications. Code is available at \href{https://github.com/lakeraai/dsec-gandalf}{\texttt{https://github.com/lakeraai/dsec-gandalf}}.

Recently, directly using large language models (LLMs) has been shown to be the most reliable method to evaluate QA models. However, it suffers from limited interpretability, high cost, and environmental harm. To address these, we propose to use soft exact match (EM) with entitydriven answer set expansion. Our approach expands the gold answer set to include diverse surface forms, based on the observation that the surface forms often follow particular patterns depending on the entity type. The experimental results show that our method outperforms traditional evaluation methods by a large margin. Moreover, the reliability of our evaluation method is comparable to that of LLM-based ones, while offering the benefits of high interpretability and reduced environmental harm.

The task of text2motion is to generate human motion sequences from given textual descriptions, where the model explores diverse mappings from natural language instructions to human body movements. While most existing works are confined to coarse-grained motion descriptions, e.g., "A man squats.", fine-grained descriptions specifying movements of relevant body parts are barely explored. Models trained with coarse-grained texts may not be able to learn mappings from fine-grained motion-related words to motion primitives, resulting in the failure to generate motions from unseen descriptions. In this paper, we build a large-scale language-motion dataset specializing in fine-grained textual descriptions, FineHumanML3D, by feeding GPT-3.5-turbo with step-by-step instructions with pseudo-code compulsory checks. Accordingly, we design a new text2motion model, FineMotionDiffuse, making full use of fine-grained textual information. Our quantitative evaluation shows that FineMotionDiffuse trained on FineHumanML3D improves FID by a large margin of 0.38, compared with competitive baselines. According to the qualitative evaluation and case study, our model outperforms MotionDiffuse in generating spatially or chronologically composite motions, by learning the implicit mappings from fine-grained descriptions to the corresponding basic motions. We release our data at https://github.com/KunhangL/finemotiondiffuse.

Headline generation aims to summarize a long document with a short, catchy title that reflects the main idea. This requires accurately capturing the core document semantics, which is challenging due to the lengthy and background information-rich na ture of the texts. In this work, We propose using a unified semantic discourse structure (S3) to represent document semantics, achieved by combining document-level rhetorical structure theory (RST) trees with sentence-level abstract meaning representation (AMR) graphs to construct S3 graphs. The hierarchical composition of sentence, clause, and word intrinsically characterizes the semantic meaning of the overall document. We then develop a headline generation framework, in which the S3 graphs are encoded as contextual features. To consolidate the efficacy of S3 graphs, we further devise a hierarchical structure pruning mechanism to dynamically screen the redundant and nonessential nodes within the graph. Experimental results on two headline generation datasets demonstrate that our method outperforms existing state-of-art methods consistently. Our work can be instructive for a broad range of document modeling tasks, more than headline or summarization generation.