At the International Conference on Machine Learning (ICML 2026),, and presented their work Vision Language Models Cannot Reason About Physical Transformation . In this interview, William Yijiang Li tells us more about the research, the method the team used, and the controlled experiments they carried out. What is the topic of the research in your paper and why is it an interesting area for study? Our research asks a fairly fundamental question: do vision-language models actually understand how the physical world changes over time? Modern vision-language models are increasingly being used in areas such as robotics, embodied agents, and video understanding.

Robots walking down the street, surrounded by astounded onlookers, is an increasingly common sight. But these machines aren't yet the do-it-all assistants you'd want working in a kitchen or factory, and a major bottleneck is data. Much like humans, robots learn best by experience. The challenge is that it's labor-intensive and time-consuming to physically teach these machines so many actions across different settings. "One natural idea is to use simulation as a training ground. While there has been significant progress over the last few years in the physics engines that power robotics simulators, one of the remaining challenges has been creating sufficiently rich and diverse simulation content to capture the complexity of the real world," says Russ Tedrake, the Toyota Professor of Electrical Engineering and Computer Science (EECS), Aeronautics and Astronautics, and Mechanical Engineering at MIT, and a principal investigator at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL).

A large number of datacentre projects around the world are being challenged or cancelled. A large number of datacentre projects around the world are being challenged or cancelled. 'These are some of the most complex structures ever created': how tech reporting moved into the physical world The Guardian's global tech reporting team are investigating the impact of the vast datacentres being built to power the AI revolution. Sun 12 Jul 2026 10.00 EDTLast modified on Sun 12 Jul 2026 10.01 EDT Journalists often use the term "shoe-leather reporting" to refer to the on-the-ground legwork that goes into covering certain stories. As the tech industry's focus has shifted from screen-based realities to the physical world of colossal AI datacentres and social media harms, comfortable footwear has become more essential to a tech reporter's job. Earlier this week, we published the Guardian's latest investigation into the datacentres and energy infrastructures that underpin AI - revealing that an £8.2bn AI complex in rural Scotland has misrepresented its plans to be powered entirely by on-site renewables.

Vision-Language Models (VLMs), with their strong reasoning and planning capabilities, are widely used in embodied decision-making (EDM) tasks in embodied agents, such as autonomous driving and robotic manipulation. Recent research has increasingly explored adversarial attacks on VLMs to reveal their vulnerabilities. However, these attacks either rely on overly strong assumptions, requiring full knowledge of the victim VLM, which is impractical for attacking VLM-based agents, or exhibit limited effectiveness. The latter stems from disrupting most semantic information in the image, which leads to a misalignment between the perception and the task context defined by system prompts. This inconsistency interrupts the VLM's reasoning process, resulting in invalid outputs that fail to affect interactions in the physical world. To this end, we propose a fine-grained adversarial attack framework, AdvEDM, which modifies the VLM's perception of only a few key objects while preserving the semantics of the remaining regions. This attack effectively reduces conflicts with the task context, making VLMs output valid but incorrect decisions and affecting the actions of agents, thus posing a more substantial safety threat in the physical world. We design two variants of based on this framework, AdvEDM-R and AdvEDM-A, which respectively remove the semantics of a specific object from the image and add the semantics of a new object into the image. The experimental results in both general scenarios and EDM tasks demonstrate fine-grained control and excellent attack performance.

Cutting-edge works have demonstrated that text-to-image (T2I) diffusion models can generate adversarial patches that mislead state-of-the-art object detectors in the physical world, revealing detectors' vulnerabilities and risks. However, these methods neglect the T2I patches' attack effectiveness when observed from different views in the physical world (i.e., angle robustness of the T2I adversarial patches). In this paper, we study the angle robustness of T2I adversarial patches comprehensively, revealing their angle-robust issues, demonstrating that texts affect the angle robustness of generated patches significantly, and task-specific linguistic instructions fail to enhance the angle robustness. Motivated by the studies, we introduce Angle-Robust Concept Learning (AngleRoCL), a simple and flexible approach that learns a generalizable concept (i.e., text embeddings in implementation) representing the capability of generating angle-robust patches. The learned concept can be incorporated into textual prompts and guides T2I models to generate patches with their attack effectiveness inherently resistant to viewpoint variations. Through extensive simulation and physical-world experiments on five SOTA detectors across multiple views, we demonstrate that AngleRoCL significantly enhances the angle robustness of T2I adversarial patches compared to baseline methods. Our patches maintain high attack success rates even under challenging viewing conditions, with over 50% average relative improvement in attack effectiveness across multiple angles. This research advances the understanding of physically angle-robust patches and provides insights into the relationship between textual concepts and physical properties in T2I-generated contents.

Watch a subscriber-only discussion exploring how AI might enter the physical world. AI companies want to build systems that understand the external world and overcome the limitations of LLMs. Recent developments have brought world models to the forefront of the AI discussion. Watch a conversation with editor in chief Mat Honan, senior AI editor Will Douglas Heaven, and AI reporter Grace Huckins exploring how AI might enter the physical world. A woman's uterus has been kept alive outside the body for the first time Jessica Hamzelou Want to understand the current state of AI? Check out these charts. Want to understand the current state of AI? Check out these charts.

Many studies have proposed attack methods to generate adversarial patterns for evading pedestrian detection, alarming the computer vision community about the need for more attention to the robustness of detectors. However, adversarial patterns optimized by these methods commonly have limited performance at medium to long distances in the physical world. To overcome this limitation, we identify two main challenges. First, in existing methods, there is commonly an appearance gap between simulated distant adversarial patterns and their physical world counterparts, leading to incorrect optimization. Second, there exists a conflict between adversarial losses at different distances, which causes difficulties in optimization. To overcome these challenges, we introduce a Full Distance Attack (FDA) method. Our physical world experiments demonstrate the effectiveness of our FDA patterns across various detection models like YOLOv5, Deformable-DETR, and Mask RCNN.

Diffusion models have demonstrated remarkable capabilities in image generation tasks, including image editing and video creation, representing a good understanding of the physical world. On the other line, diffusion models have also shown promise in robotic control tasks by denoising actions, known as diffusion policy. Although the diffusion generative model and diffusion policy exhibit distinct capabilities--image prediction and robotic action, respectively--they technically follow similar denoising process. In robotic tasks, the ability to predict future images and generate actions is highly correlated since they share the same underlying dynamics of the physical world. Building on this insight, we introduce \textbf{PAD}, a novel visual policy learning framework that unifies image \textbf{P}rediction and robot \textbf{A}ction within a joint \textbf{D}enoising process. Specifically, PAD utilizes Diffusion Transformers (DiT) to seamlessly integrate images and robot states, enabling the simultaneous prediction of future images and robot actions.

Each year the AAAI recognizes a group of individuals who have made significant, sustained contributions to the field of artificial intelligence by appointing them as Fellows. Over the course of the next few months, we'll be talking to some of the 2026 AAAI Fellows . In this interview, we met with Yan Liu, University of Southern California, who was elected as a Fellow . We found out about how time series research has progressed, the vast range of applications, and what the future holds for this field. Could you start with a quick introduction to your area of research?

The Good Robot podcast: what makes a drone "good"? Hosted by Eleanor Drage and Kerry McInerney, The Good Robot is a podcast which explores the many complex intersections between gender, feminism and technology. What makes a drone "good"? In this episode, we talk to Beryl Pong, UKRI Future Leaders Fellow at the University of Cambridge, where she leads the Centre for Drones and Culture. Beryl reflects on what it means to think about drones as "good" or "ethical" technologies and how it can be assessed through its socio-political context.