Retrieval systems rely on representations learned by increasingly powerful models. However, due to the high training cost and inconsistencies in learned representations, there is significant interest in facilitating communication between representations and ensuring compatibility across independently trained neural networks. In the literature, two primary approaches are commonly used to adapt different learned representations: affine transformations, which adapt well to specific distributions but can significantly alter the original representation, and orthogonal transformations, which preserve the original structure with strict geometric constraints but limit adaptability. A key challenge is adapting the latent spaces of updated models to align with those of previous models on downstream distributions while preserving the newly learned representation spaces. In this paper, we impose a relaxed orthogonality constraint, namely λ-Orthogonality regularization, while learning an affine transformation, to obtain distribution-specific adaptation while retaining the original learned representations. Extensive experiments across various architectures and datasets validate our approach, demonstrating that it preserves the model's zero-shot performance and ensures compatibility across model updates.

Large language model (LLM) evaluation is increasingly costly, prompting interest in methods that speed up evaluation by shrinking benchmark datasets. Benchmark prediction (also called efficient LLM evaluation) aims to select a small subset of evaluation points and predict overall benchmark performance from that subset. In this paper, we systematically assess the strengths and limitations of 11 benchmark prediction methods across 19 diverse benchmarks. First, we identify a highly competitive baseline: Take a random sample and fit a regression model on the sample to predict missing entries. Outperforming most existing methods, this baseline challenges the assumption that careful subset selection is necessary for benchmark prediction.

The rest of Dyson's promised 2026 vacuum lineup is here, from the new Dyson V16 Piston Animal to an updated version of the favored Dyson V8 Cyclone. Dyson's vacuum lineup had a new look planned for this year . Some of the vacuums have already arrived, like the Dyson PencilVac and Dyson Spot+Scrub robot vacuum, but others we've still been waiting to see. That wait is over as of this month, as Dyson has finally dropped the rest of its anticipated models. Dyson now has three new cordless vacuums you can shop, plus one with a Submarine head variant: the Dyson V16 Piston Animal ($980) and Dyson V16 Piston Animal Submarine ($1,100), the Dyson V10 Konical ($500), and the Dyson V8 Cyclone ($400) .

The ChatGPT Images 2.0 model is here. Our testing shows it's better at creating more detailed images and rendering text, but it still struggles with languages other than English. OpenAI launched a new image generation AI model on Tuesday, dubbed ChatGPT Images 2.0. This model can generate more than one image from a single prompt, like an entire study booklet, as well as output text, including in non-English languages, like Chinese and Hindi. This release is available globally for ChatGPT and Codex users, with a more powerful version available for paying subscribers.

ChatGPT Images 2.0 is better at rendering non-Latin text OpenAI describes it as a step change for image generation models. OpenAI's new ChatGPT Images 2.0 model is now available. A little more than a year after OpenAI gave ChatGPT users the option to create images and designs directly from its chatbot, it's now releasing ChatGPT Images 2.0 . OpenAI describes the new system as a "step change" for image generation models, particularly when it comes to the tool's ability to follow instructions in detail, render dense text and place and relate objects in a scene. For the first time, OpenAI has also built an image model with reasoning capabilities, giving the system the ability to do things like search the web and verify its outputs.

We investigate the space of weights spanned by a large collection of customized diffusion models. We populate this space by creating a dataset of over 60,000 models, each of which is a base model fine-tuned to insert a different person's visual identity.

With the rapid development of deep learning, the increasing complexity and scale of parameters make training a new model increasingly resource-intensive. In this paper, we start from the classic convolutional neural network (CNN) and explore a paradigm that does not require training to obtain new models. Similar to the birth of CNN inspired by receptive fields in the biological visual system, we draw inspiration from the information subsystem pathways in the biological visual system and propose Model Disassembling and Assembling (MDA). During model disassembling, we introduce the concept of relative contribution and propose a component locating technique to extract task-aware components from trained CNN classifiers. For model assembling, we present the alignment padding strategy and parameter scaling strategy to construct a new model tailored for a specific task, utilizing the disassembled task-aware components.The entire process is akin to playing with LEGO bricks, enabling arbitrary assembly of new models, and providing a novel perspective for model creation and reuse. Extensive experiments showcase that task-aware components disassembled from CNN classifiers or new models assembled using these components closely match or even surpass the performance of the baseline,demonstrating its promising results for model reuse. Furthermore, MDA exhibits diverse potential applications, with comprehensive experiments exploring model decision route analysis, model compression, knowledge distillation, and more.

Graph Neural Networks (GNNs), known for their effective graph encoding, are extensively used across various fields. Graph self-supervised pre-training, which trains GNN encoders without manual labels to generate high-quality graph representations, has garnered widespread attention. However, due to the inherent complex characteristics in graphs, GNNs encoders pre-trained on one dataset struggle to directly adapt to others that have different node feature shapes.

GPT-5.4 mini brings some of the smarts of OpenAI's latest model to ChatGPT Free and Go users The new model offers performance improvements in reasoning, multimodal understanding and more. The ChatGPT icon, as seen on iPhone 12 running iOS. When OpenAI released GPT-5.4 at the start of March, the company said the new model was designed primarily for professional work like programming and data analysis. Now OpenAI is launching GPT-5.4 mini and nano, and while it is once again highlighting the usefulness of these new systems for tasks like coding, one of the new models is available to Free and Go users . What's more, that model, GPT-5.4 mini, even offers performance that approaches GPT-5.4 in a handful of areas.