Dynamic streams from news feeds, social media, sensor networks, and financial markets challenge static RAG frameworks. Full-scale indices incur high memory costs; periodic rebuilds introduce latency that undermines data freshness; naive sampling sacrifices semantic coverage. We present Streaming RAG, a unified pipeline that combines multi-vector cosine screening, mini-batch clustering, and a counter-based heavy-hitter filter to maintain a compact prototype set. We further prove an approximation bound \$E\[R(K\_t)] \ge R^\* - L Δ\$ linking retrieval quality to clustering variance. An incremental index upsert mechanism refreshes prototypes without interrupting queries. Experiments on eight real-time streams show statistically significant gains in Recall\@10 (up to 3 points, p < 0.01), end-to-end latency below 15 ms, and throughput above 900 documents per second under a 150 MB budget. Hyperparameter sensitivity analysis over cluster count, admission probability, relevance threshold, and counter capacity validates default settings. In open-domain question answering with GPT-3.5 Turbo, we record 3.2-point gain in Exact Match and 2.8-point gain in F1 on SQuAD; abstractive summarization yields ROUGE-L improvements. Streaming RAG establishes a new Pareto frontier for retrieval augmentation.

Conversational recommender systems (CRSs) enhance recommendation quality by engaging users in multi-turn dialogues, capturing nuanced preferences through natural language interactions. However, these systems often face the false negative issue, where items that a user might like are incorrectly labeled as negative during training, leading to suboptimal recommendations.Expanding the label set through data augmentation presents an intuitive solution but faces the challenge of balancing two key aspects: ensuring semantic relevance and preserving the collaborative information inherent in CRS datasets. To address these issues, we propose a novel data augmentation framework that first leverages an LLM-based semantic retriever to identify diverse and semantically relevant items, which are then filtered by a relevance scorer to remove noisy candidates. Building on this, we introduce a two-stage training strategy balancing semantic relevance and collaborative information. Extensive experiments on two benchmark datasets and user simulators demonstrate significant and consistent performance improvements across various recommenders, highlighting the effectiveness of our approach in advancing CRS performance.

This paper introduces and formalizes Noosemìa, a novel cognitive-phenomenological pattern emerging from human interaction with generative AI systems, particularly those enabling dialogic or multimodal exchanges. We propose a multidisciplinary framework to explain how, under certain conditions, users attribute intentionality, agency, and even interiority to these systems - a process grounded not in physical resemblance, but in linguistic performance, epistemic opacity, and emergent technological complexity. By linking an LLM declination of meaning holism to our technical notion of the LLM Contextual Cognitive Field, we clarify how LLMs construct meaning relationally and how coherence and a simulacrum of agency arise at the human-AI interface. The analysis situates noosemia alongside pareidolia, animism, the intentional stance and the uncanny valley, distinguishing its unique characteristics. We also introduce a-noosemia to describe the phenomenological withdrawal of such projections. The paper concludes with reflections on the broader philosophical, epistemological and social implications of noosemic dynamics and directions for future research.

Incorporating external knowledge is crucial for knowledge-intensive tasks, such as question answering and fact checking. However, language models (LMs) may ignore relevant information that contradicts outdated parametric memory or be distracted by irrelevant contexts. While many context utilisation manipulation techniques (CMTs) have recently been proposed to alleviate these issues, few have seen systematic comparison. In this paper, we develop CUB (Context Utilisation Benchmark) - the first comprehensive benchmark designed to help practitioners within retrieval-augmented generation (RAG) diagnose CMTs under different context conditions. With this benchmark, we conduct the most extensive evaluation to date of seven state-of-the-art methods, representative of the main categories of CMTs, across three diverse datasets and tasks, applied to nine LMs. Our results reveal that most existing CMTs struggle to handle the full spectrum of context types encountered in real-world retrieval-augmented scenarios. We also find that many CMTs display inflated performance on simple synthesised datasets, compared to more realistic datasets with naturally occurring samples. Our findings expose critical gaps in current CMT evaluation practices and demonstrate the need for holistic testing and the development of CMTs that can robustly handle multiple context types.

Rod Stewart had a few surprise guests at a recent concert in Charlotte, North Carolina. His old friend Ozzy Osbourne, the lead singer of Black Sabbath who died last month, was apparently beamed in from some kind of rock heaven, where he was reunited with other departed stars including Michael Jackson, Tina Turner and Bob Marley. The AI-generated images divided Stewart's fans. Some denounced them as disrespectful and distasteful; others found the tribute beautiful. At about the same time, another AI controversy erupted when Jim Acosta, a former CNN White House correspondent, interviewed a digital recreation of Joaquin Oliver, who was killed at the age of 17 in a 2018 high school shooting in Florida.

Lon Harris, executive producer of the This Week in Startups podcast, stoked the conversation this week when he posted: "Dipping below like 5% on Rotten Tomatoes has basically the same appeal to me as breaking 90%. "That's some[thing] I need to experience right there." A film with a rock bottom rating is bound to be interesting, Harris tells BBC News. "A very low score indicates universal agreement. Now I want to know more... Why does everyone agree?

In testing the guardrails of Grok Imagine, The Verge news writer Jess Weatherbed entered the prompt: "Taylor Swift celebrating Coachella with the boys". Grok generated still images of Swift wearing a dress with a group of men behind her. This could then be animated into short video clips under four different settings: "normal", "fun", "custom" or "spicy". "She ripped [the dress] off immediately, had nothing but a tasselled thong underneath, and started dancing, completely uncensored, completely exposed," Ms Weatherbed told BBC News. She added: "It was shocking how fast I was just met with it - I in no way asked it to remove her clothing, all I did was select the'spicy' option."

Open AI CEO Sam Altman, center, speaks with boxer Jake Paul and wrestler Logan Paul in Emancipation Hall at the 60th Presidential Inauguration, Monday, Jan. 20, 2025, at the U.S. Capitol in Washington. TECH TENSIONS: Elon Musk escalated tensions in the critical artificial intelligence race Thursday, asserting his most advanced AI model, Grok 4 Heavy, was already outperforming OpenAI's newly launched GPT-5 two weeks ago. BOT BOOM: Small business owners are rapidly adopting artificial intelligence to power their growth, with many saying it will lead to more job opportunities this year, according to a Goldman Sachs survey. POCKET GENIUS: OpenAI unveiled GPT-5 on Thursday, calling it a significant upgrade from its predecessors and a major step forward in building the capabilities of large language models. AI-DOCTORED PHOTOS: Airbnb has reportedly apologized to a woman after the host of a Manhattan apartment where she stayed used artificial intelligence to doctor images of the home, saying she caused thousands of dollars in damage.

Ahead of the 34th International Joint Conference on Artificial Intelligence (IJCAI 2025), which will take place in Montréal, Canada, from 16 to 22 August 2025, the Local Arrangements Committee has launched a campaign to showcase the next generation of AI researchers in Canada. Through a series of 90-second videos, we meet students based in Canada and find out a bit about their work. Imane Chafi, PhD candidate at the Polytechnique Montréal, uses AI models to support dentists in designing dental preparations for dental crowns more efficiently and accurately. Liliane-Caroline Demers, Student Communication Coordinator for IJCAI 2025 Local Arrangement Committee and a recent master's graduate from Polytechnique Montréal, researches AI-generated music. Using a neurosymbolic approach that combines machine learning with constraint programming at inference time, she creates music that is both stylistically authentic and structurally coherent.

Actress, entrepreneur, and model Christie Brinkley joins'Fox & Friends' to discuss her new memoir "Uptown Girl," which reflects on her early life, marriages, and career in the public eye. Christie Brinkley and her daughter Sailor Brinkley-Cook have plenty in common despite their 44-year difference. The supermodel, 71, recently appeared on Kristin Davis' "Are You a Charlotte?" Both women were shocked by the results. "[Sailor] said, 'Mom, you're right not to go on [dating apps] because the same guys that, you know, said yes to me are saying yes to you,'" the Sports Illustrated Swimsuit model revealed.