At Google I/O 2025, the tech company announced a ton of new AI features, and one of the most interesting is a virtual clothing try-on tool. The Google Shopping "Try it on" feature lets users upload a photo of themselves and then virtually try on clothes, basically the IRL version of the Clueless closet millennials have been dreaming about since 1995. Or, as Mashable Shopping Reporter Haley Henschel put it, "Google's latest shopping feature makes Cher Horowitz's computerized closet a reality." Almost as soon as the feature was released, users started trying to "jailbreak" the tool, which is becoming a fun little tradition for tech writers every time a new AI model or tool is released. On Friday, The Atlantic reported that "Google's new AI shopping tool appears eager to give J.D. Vance breasts."

On May 19, President Donald Trump and First Lady Melania Trump beamed to press and allies as they signed the administration's first major piece of tech regulation, the bipartisan Take It Down Act. It was seen as a win for those who have long been calling on the criminalization of NDII, or the nonconsensual distribution of intimate images, and a federal pathway of redress for victims. Cliff Steinhauer, director of information security and engagement at the National Cybersecurity Alliance, explained it may be a needed kick in the pants to a lethargic legislative arena. "I think it's good that they're going to force social media companies to have a process in place to remove content that people ask to be removed," he said. "This is kind of a start; to build the infrastructure to be able to respond to this type of request, and it's a really thin slice of what the issues with AI are going to be."

What started with Microsoft Build, continued with Google I/O, and ended with Anthropic Code with Claude, plus a big hardware interruption from OpenAI, the week has finally come to a close. AI announcements from the developer conferences jockeyed for news dominance this week, but OpenAI managed to make headlines without an event by announcing that it's going to start making AI devices with iPhone designer Jony Ives We'll get to that, plus all the major AI features from Google and Microsoft and details about Anthropic's new models. Take a look at the AI news of the week, then enjoy a well-deserved weekend. On Thursday, Anthropic introduced the next generation of its Claude models: Opus 4 and Sonnet 4. Claude Opus 4 is the bigger, more powerful model, while Sonnet 4 is smaller and nimbler. Anthropic said both models scored higher than their rivals on agentic AI benchmarks and said they're particularly good for coding and reasoning tasks.

The inversion of diffusion model sampling, which aims to find the corresponding initial noise of a sample, plays a critical role in various tasks. Recently, several heuristic exact inversion samplers have been proposed to address the inexact inversion issue in a training-free manner. However, the theoretical properties of these heuristic samplers remain unknown and they often exhibit mediocre sampling quality. In this paper, we introduce a generic formulation, Bidirectional Explicit Linear Multi-step (BELM) samplers, of the exact inversion samplers, which includes all previously proposed heuristic exact inversion samplers as special cases.

Ukrainian officials have renewed their calls for more sanctions on Russia after Russian forces launched dozens of attack drones and ballistic missiles at Kyiv overnight ahead of a second exchange of soldiers and civilians. Ukraine's military on Saturday said overnight attacks launched from multiple Russian regions used 250 drones and 14 ballistic missiles to hit Kyiv, damaging several apartment buildings and a shopping mall, and injuring at least 15 people. Sites in the Ukrainian regions of Dnipropetrovsk, Odesa and Zaporizhia were also hit, with Ukrainian forces saying six of the ballistic missiles were shot down by its air defences, along with 245 drones, many of which were said to be Iranian-designed. Oleh Syniehubov, head of Kharkiv's regional state administration, said on Saturday morning that four Ukrainians were killed and several others injured over the past 24 hours in the region as a result of multiple Russian attacks. Meanwhile, Russia's Ministry of Defence on Saturday said at least 100 Ukrainian drones attempted to strike Russian targets overnight.

In the literature of studying training dynamics of transformers, several simplifications are commonly adopted such as weight reparameter-ization, attention linearization, special initialization, and lazy regime.

Material discovery is a critical area of research with the potential to revolutionize various fields, including carbon capture, renewable energy, and electronics. However, the immense scale of the chemical space makes it challenging to explore all possible materials experimentally. In this paper, we introduce FlowLLM, a novel generative model that combines large language models (LLMs) and Riemannian flow matching (RFM) to design novel crystalline materials. FlowLLM first finetunes an LLM to learn an effective base distribution of meta-stable crystals in a text representation. After converting to a graph representation, the RFM model takes samples from the LLM and iteratively refines the coordinates and lattice parameters. Our approach significantly outperforms state-of-the-art methods, increasing the generation rate of stable materials by over three times and increasing the rate for stable, unique, and novel crystals by 50% - a huge improvement on a difficult problem. Additionally, the crystals generated by FlowLLM are much closer to their relaxed state when compared with another leading model, significantly reducing post-hoc computational cost.

In this work, we introduce Scribbles for All, a label and training data generation algorithm for semantic segmentation trained on scribble labels. Training or fine-tuning semantic segmentation models with weak supervision has become an important topic recently and was subject to significant advances in model quality. In this setting, scribbles are a promising label type to achieve high quality segmentation results while requiring a much lower annotation effort than usual pixel-wise dense semantic segmentation annotations. The main limitation of scribbles as source for weak supervision is the lack of challenging datasets for scribble segmentation, which hinders the development of novel methods and conclusive evaluations. To overcome this limitation, Scribbles for All provides scribble labels for several popular segmentation datasets and provides an algorithm to automatically generate scribble labels for any dataset with dense annotations, paving the way for new insights and model advancements in the field of weakly supervised segmentation. In addition to providing datasets and algorithm, we evaluate state-of-the-art segmentation models on our datasets and show that models trained with our synthetic labels perform competitively with respect to models trained on manual labels. Thus, our datasets enable state-of-the-art research into methods for scribble-labeled semantic segmentation.

Existing Goal-Conditioned Reinforcement Learning (GCRL) algorithms are built upon Hindsight Experience Replay (HER), which densifies rewards through hindsight replay and leverages historical goal-achieving information to construct a learning curriculum. However, when the task is characterized by a non-Markovian reward (NMR), whose computation depends on multiple steps of states and actions, HER can no longer densify rewards by treating a single encountered state as the hindsight goal. The lack of informative rewards hinders policy learning, resulting in rolling out failed trajectories. Consequently, the replay buffer is overwhelmed with failed trajectories, impeding the establishment of an applicable curriculum. To circumvent these limitations, we deviate from existing HER-based methods and propose an on-policy GCRL framework, GCPO, which is applicable to both multi-goal Markovian reward (MR) and NMR problems. GCPO consists of (1) Pre-training from Demonstrations, which pre-trains the policy to possess an initial goal-achieving capability, thereby diminishing the difficulty of subsequent online learning.

Designing a fast and effective entropy model is challenging but essential for practical application of neural codecs. Beyond spatial autoregressive entropy models, more efficient backward adaptation-based entropy models have been recently developed. They not only reduce decoding time by using smaller number of modeling steps but also maintain or even improve rate-distortion performance by leveraging more diverse contexts for backward adaptation. Despite their significant progress, we argue that their performance has been limited by the simple adoption of the design convention for forward adaptation: using only a single type of hyper latent representation, which does not provide sufficient contextual information, especially in the first modeling step. In this paper, we propose a simple yet effective entropy modeling framework that leverages sufficient contexts for forward adaptation without compromising on bit-rate. Specifically, we introduce a strategy of diversifying hyper latent representations for forward adaptation, i.e., using two additional types of contexts along with the existing single type of context. In addition, we present a method to effectively use the diverse contexts for contextualizing the current elements to be encoded/decoded. By addressing the limitation of the previous approach, our proposed framework leads to significant performance improvements. Experimental results on popular datasets show that our proposed framework consistently improves rate-distortion performance across various bit-rate regions, e.g., 3.73% BD-rate gain over the state-of-the-art baseline on the Kodak dataset.