PCWorld reports Microsoft has stopped automatically installing Microsoft 365 Copilot on Windows 11 following significant user backlash against forced AI integration. This decision addresses privacy concerns after a previous Copilot bug allowed unauthorized access to confidential Outlook emails. Administrators can still manually deploy the AI assistant, while existing installations remain unaffected by this policy change. Since October, the Microsoft 365 Copilot app has been automatically installed on computers running Windows 11, a move that has upset many users. Fortunately, Microsoft has taken the criticism on board and is no longer automatically installing this app--for now.

Female pastor is suspended after her shocking Epstein link is exposed... as she compares herself to JESUS while defending their relationship'Tell me to my face': Republican senator torches Noem's replacement as their vicious personal feud spills into public Outrageous full story of scandalous affair that's the talk of Manhattan's exclusive private schools: Family insiders reveal humiliating sex secrets... shock'confession' letter... and the furious relative who exposed it all Ugly new Nicole Kidman and Keith Urban divorce fight ERUPTS: Her friends share humiliating details of'midlife crisis'... and reveal brutal REAL reason daughter Sunday Rose'snubbed' him Perfect All-American family lived in stunning $1.1m Colorado mansion and bankrolled glamorous daughter's horse stables... now matriarch has sullied their good name with a HUGE scandal Meghan unveils new As Ever line with Lilibet... amid claims Netflix has been left with huge $10m surplus of her unsold products after'split' with streamer Woke Democrat, 26, who can't get out of bed in time for meetings loses primary to professor accused of inappropriate relationship by former student I watched the children's book author who poisoned her husband from 5ft away. This is the off-camera moment her mask finally slipped... it was truly chilling I ran America's only Supermax jail: What history's most notorious terrorists and serial killers told me as they waited to die Sinister truth about explosive resignation of Trump's top counter-terror chief Joe Kent... and his shock claim Israel is manipulating the president: MARK HALPERIN Hairdresser who weighs 300lbs says Southwest airport check-in worker looked him up and down and told him he'd have to buy extra seat Kim Kardashian takes a VERY dramatic tumble in towering $80 'stripper heels' and accidentally grabs an'old lady' as she falls on her way out of Vanity Fair Oscar party Everything JFK Jr told friends about his love affair with'sexual dynamo' Madonna... her unprintable pillow talk... and his perverse incest request that she couldn't go through with Saudi, UAE and Qatar energy facilities are evacuated after Iran threatens'full scale economic war' as oil price jumps 5%: Live updates New PILL for psoriasis approved... giving hope to millions suffering from debilitating skin condition How I lost 8st in my 50s and now finally have the figure of my dreams. I've been large my whole life, but I now feel happier than I ever did in my 20s. New York City's accent is dying out, study finds It's something that's fixed from roughly the age of 14. But your accent could be hindering your job prospects, according to a new study.

Watermarking is a technique that involves embedding nearly unnoticeable statistical signals within generated content to help trace its source. This work focuses on a scenario where an untrusted third-party user sends prompts to a trusted language model (LLM) provider, who then generates a text from their LLM with a watermark. This setup makes it possible for a detector to later identify the source of the text if the user publishes it. The user can modify the generated text by substitutions, insertions, or deletions. Our objective is to develop a statistical method to detect if a published text is LLM-generated from the perspective of a detector. We further propose a methodology to segment the published text into watermarked and non-watermarked sub-strings. The proposed approach is built upon randomization tests and change point detection techniques. We demonstrate that our method ensures Type I and Type II error control and can accurately identify watermarked sub-strings by finding the corresponding change point locations. To validate our technique, we apply it to texts generated by several language models with prompts extracted from Google's C4 dataset and obtain encouraging numerical results.

When applying reinforcement learning from human feedback (RLHF), the reward is learned from data and, therefore, always has some error. It is common to mitigate this by regularizing the policy with KL divergence from a base model, with the hope that balancing reward with regularization will achieve desirable outcomes despite this reward misspecification. We show that when the reward function has light-tailed error, optimal policies under less restrictive KL penalties achieve arbitrarily high utility. However, if error is heavy-tailed, some policies obtain arbitrarily high reward despite achieving no more utility than the base model--a phenomenon we call catastrophic Goodhart. We adapt a discrete optimization method to measure the tails of reward models, finding that they are consistent with light-tailed error. However, the pervasiveness of heavy-tailed distributions in many real-world applications indicates that future sources of RL reward could have heavy-tailed error, increasing the likelihood of reward hacking even with KL regularization.

Frontier AI systems, including large language models (LLMs), hold increasing influence over the epistemology of human users. Such influence can reinforce prevailing societal values, potentially contributing to the lock-in of misguided moral beliefs and, consequently, the perpetuation of problematic moral practices on a broad scale.

Mean-field variational inference (VI) is computationally scalable, but its highly-demanding independence requirement hinders it from being applied to wider scenarios. Although many VI methods that take correlation into account have been proposed, these methods generally are not scalable enough to capture the correlation among data instances, which often arises in applications with graph-structured data or explicit constraints. In this paper, we developed the Tree-structured Variational Inference (TreeVI), which uses a tree structure to capture the correlation of latent variables in the posterior distribution. We show that samples from the tree-structured posterior can be reparameterized efficiently and parallelly, making its training cost just 2 or 3 times that of VI under the mean-field assumption. To capture correlation with more complicated structure, the TreeVI is further extended to the multiple-tree case. Furthermore, we show that the underlying tree structure can be automatically learned from training data. With experiments on synthetic datasets, constrained clustering, user matching and link prediction, we demonstrate that the TreeVI is superior in capturing instance-level correlation in posteriors and enhancing the performance of downstream applications.

This paper presents a new gradient flow dissipation geometry over non-negative and probability measures.This is motivated by a principled construction that combines the unbalanced optimal transport and interaction forces modeled by reproducing kernels. Using a precise connection between the Hellinger geometry and the maximum mean discrepancy (MMD), we propose the interaction-force transport (IFT) gradient flows and its spherical variant via an infimal convolution of the Wasserstein and spherical MMD tensors. We then develop a particle-based optimization algorithm based on the JKO-splitting scheme of the mass-preserving spherical IFT gradient flows. Finally, we provide both theoretical global exponential convergence guarantees and improved empirical simulation results for applying the IFT gradient flows to the sampling task of MMD-minimization. Furthermore, we prove that the spherical IFT gradient flow enjoys the best of both worlds by providing the global exponential convergence guarantee for both the MMD and KL energy.

Federated Learning (FL) has recently been applied to the parameter-efficient fine-tuning of Large Language Models (LLMs). While promising, it raises significant challenges due to the heterogeneous resources and data distributions of clients.This study introduces FlexLoRA, a simple yet effective aggregation scheme for LLM fine-tuning, which mitigates the buckets effect in traditional FL that restricts the potential of clients with ample resources by tying them to the capabilities of the least-resourced participants. FlexLoRA allows for dynamic adjustment of local LoRA ranks, fostering the development of a global model imbued with broader, less task-specific knowledge.

Images produced by diffusion models are increasingly popular in digital artwork and visual marketing. However, such generated images might replicate content from existing ones and pose the challenge of content originality. Existing Image Copy Detection (ICD) models, though accurate in detecting hand-crafted replicas, overlook the challenge from diffusion models. This motivates us to introduce ICDiff, the first ICD specialized for diffusion models. To this end, we construct a Diffusion-Replication (D-Rep) dataset and correspondingly propose a novel deep embedding method.

Memory models such as Recurrent Neural Networks (RNNs) and Transformers address Partially Observable Markov Decision Processes (POMDPs) by mapping trajectories to latent Markov states. Neither model scales particularly well to long sequences, especially compared to an emerging class of memory models called Linear Recurrent Models. We discover that the recurrent update of these models resembles a monoid, leading us to reformulate existing models using a novel monoid-based framework that we call memoroids. We revisit the traditional approach to batching in recurrent reinforcement learning, highlighting theoretical and empirical deficiencies. We leverage memoroids to propose a batching method that improves sample efficiency, increases the return, and simplifies the implementation of recurrent loss functions in reinforcement learning.