Large recommender models have extended LLMs as powerful recommenders via encoding or item generation, and recent breakthroughs in LLM reasoning synchronously motivate the exploration of reasoning in recommendation. In this work, we propose R$^2$ec, a unified large recommender model with intrinsic reasoning capability. R$^2$ec introduces a dual-head architecture that supports both reasoning chain generation and efficient item prediction in a single model, significantly reducing inference latency. To overcome the lack of annotated reasoning data, we design RecPO, a reinforcement learning framework that optimizes reasoning and recommendation jointly with a novel fused reward mechanism. Extensive experiments on three datasets demonstrate that R$^2$ec outperforms traditional, LLM-based, and reasoning-augmented recommender baselines, while further analyses validate its competitive efficiency among conventional LLM-based recommender baselines and strong adaptability to diverse recommendation scenarios. Code and checkpoints available at https://github.com/YRYangang/RRec.

Vision-Language Models (VLMs) often struggle to balance visual and textual information when summarizing complex multimodal inputs, such as entire TV show episodes. In this paper, we propose a zero-shot video-to-text summarization approach that builds its own screenplay representation of an episode, effectively integrating key video moments, dialogue, and character information into a unified document. Unlike previous approaches, we simultaneously generate screenplays and name the characters in zero-shot, using only the audio, video, and transcripts as input. Additionally, we highlight that existing summarization metrics can fail to assess the multimodal content in summaries. To address this, we introduce MFactSum, a multimodal metric that evaluates summaries with respect to both vision and text modalities. Using MFactSum, we evaluate our screenplay summaries on the SummScreen3D dataset, demonstrating superiority against state-of-the-art VLMs such as Gemini 1.5 by generating summaries containing 20% more relevant visual information while requiring 75% less of the video as input.

This material may not be published, broadcast, rewritten, or redistributed. Quotes displayed in real-time or delayed by at least 15 minutes. Market data provided by Factset . Powered and implemented by FactSet Digital Solutions . Mutual Fund and ETF data provided by Refinitiv Lipper .

What do you see? 12 extreme close-ups bring'hidden science' to life New photography book encourages us to look for science everywhere. Over many years of geological time, opal is slowly formed as many small spheres of silica (what glass is made of) self-assemble into perfectly ordered layers. Breakthroughs, discoveries, and DIY tips sent every weekday. Organized into five thematic sections, the book turns learning about science into a guessing game. Detailed photographs like the ones below taken by MIT researcher and science photographer Felice Frankel challenge readers to deduce the underlying chemical, natural, or physical processes at play.

Taiwan University researchers developed a serum that regrew hair in mice within 20 days by using fatty acids to activate hair stem cells after skin injury.

Security researchers discovered SessionReaper, a serious vulnerability in Magento and Adobe Commerce that allows hackers to hijack shopping sessions and steal data.

Parents can better manage their children's digital lives by understanding five key tech terms including screen time limits, parental controls, and digital footprints.

Terror cops probe knife attack on train as nine fight for their lives and armed police arrest two amid'horrifying' scenes Furious leaders question why they weren't warned over dangerous levels of radiation detected at former San Francisco naval shipyard I descended to Hell for 8 hours after a suicide attempt. It's nothing like the movies... my mother prayed to every God - but only one came to save me Andrew Mountbatten Windsor'refused to sign off royal tributes to Jeffrey Epstein victims' I can't disclose my medical history to my partner. If I do... he'll find me so unsexy that he'll leave: DEAR JANE How Andrew's'rude' comment about Kate sparked bitter feud between ex-prince and William - who'couldn't wait for the day' when Charles finally threw him out Inside humiliated Andrew's new life in exile: From butlers and Downton-style splendour to a pokey cottage with a latch key, friends tell RICHARD KAY how disgraced royal will now live... and reveal who is'propping him up' For six years, I woke at 7.30am, had a shot of vodka, a line of cocaine... and Viagra before sex with the receptionist at work. Bill Maher, 69, and Al Pacino's baby mama Noor Alfallah, 31, reignite romance rumors at star-studded Halloween bash Anthony Hopkins, 87, 'puts his California estate on the market for £5.1 million' after devastating wildfires destroyed his home Pennsylvania diocese apologizes after Catholic school's Halloween float features replica of Auschwitz gate Nicki Minaj draws liberal fury by praising Donald Trump's latest move in emotional post SNL pokes fun at Trump's White House renovation with HGTV-style makeover as Miles Teller portrays Property Brothers in chaotic comedy skit Trump labels Seth Meyers a'deranged lunatic' and blasts his late-night rhetoric as'illegal' A terrifying simulation has revealed how people might really behave as the end of the world approaches. And it suggests that humanity's darkest instincts might reign supreme at the very end.

Meta's new display-less smart glasses are quite good, but the vibes are off. All products featured on WIRED are independently selected by our editors. However, when you buy something through our retail links, we may earn an affiliate commission. Upgraded camera shoots 3K photos and slow-motion video. Ray-Bans sure do look slick.

Adults over 70 who regularly listen to music show 39% lower dementia risk while instrument players have 35% reduced risk, according to new Monash University research.