Despite rapid advances in vision-language models (VLMs), current benchmarks for multimodal reasoning fall short in three key dimensions. First, they overwhelmingly rely on static images, failing to capture the temporal complexity of real-world environments. Second, they narrowly focus on mathematical problem-solving, neglecting the broader spectrum of reasoning skills -- including abstract, physical, planning, spatial, and temporal capabilities -- required for robust multimodal intelligence. Third, many benchmarks quickly saturate, offering limited headroom for diagnosing failure modes or measuring continued progress. We introduce MORSE-500 (Multimodal Reasoning Stress-test Environment), a video benchmark composed of 500 fully scripted clips with embedded questions spanning six complementary reasoning categories. Each instance is programmatically generated using deterministic Python scripts (via Manim, Matplotlib, MoviePy), generative video models, and curated real footage. This script-driven design allows fine-grained control over visual complexity, distractor density, and temporal dynamics -- enabling difficulty to be scaled systematically as models improve. Unlike static benchmarks that become obsolete once saturated, MORSE-500 is built to evolve: its controllable generation pipeline supports the creation of arbitrarily challenging new instances, making it ideally suited for stress-testing next-generation models. Initial experiments with state-of-the-art systems -- including various Gemini 2.5 Pro and OpenAI o3 which represent the strongest available at the time, alongside strong open-source models -- reveal substantial performance gaps across all categories, with particularly large deficits in abstract and planning tasks. We release the full dataset, generation scripts, and evaluation harness to support transparent, reproducible, and forward-looking multimodal reasoning research.

Binary code similarity detection (BCSD) serves as a fundamental technique for various software engineering tasks, e.g., vulnerability detection and classification. Attacks against such models have therefore drawn extensive attention, aiming at misleading the models to generate erroneous predictions. Prior works have explored various approaches to generating semantic-preserving variants, i.e., adversarial samples, to evaluate the robustness of the models against adversarial attacks. However, they have mainly relied on heuristic criteria or iterative greedy algorithms to locate salient code influencing the model output, failing to operate on a solid theoretical basis. Moreover, when processing programs with high complexities, such attacks tend to be time-consuming. In this work, we propose a novel optimization for adversarial attacks against BCSD models. In particular, we aim to improve the attacks in a challenging scenario, where the attack goal is to limit the model predictions to a specific range, i.e., the targeted attacks. Our attack leverages the superior capability of black-box, model-agnostic explainers in interpreting the model decision boundaries, thereby pinpointing the critical code snippet to apply semantic-preserving perturbations. The evaluation results demonstrate that compared with the state-of-the-art attacks, the proposed attacks achieve higher attack success rate in almost all scenarios, while also improving the efficiency and transferability. Our real-world case studies on vulnerability detection and classification further demonstrate the security implications of our attacks, highlighting the urgent need to further enhance the robustness of existing BCSD models.

Vision-language models, which integrate computer vision and natural language processing capabilities, have demonstrated significant advancements in tasks such as image captioning and visual question and answering. However, similar to traditional models, they are susceptible to small perturbations, posing a challenge to their robustness, particularly in deployment scenarios. Evaluating the robustness of these models requires perturbations in both the vision and language modalities to learn their inter-modal dependencies. In this work, we train a generic surrogate model that can take both image and text as input and generate joint representation which is further used to generate adversarial perturbations for both the text and image modalities. This coordinated attack strategy is evaluated on the visual question and answering and visual reasoning datasets using various state-of-the-art vision-language models. Our results indicate that the proposed strategy outperforms other multi-modal attacks and single-modality attacks from the recent literature. Our results demonstrate their effectiveness in compromising the robustness of several state-of-the-art pre-trained multi-modal models such as instruct-BLIP, ViLT and others.

Deep learning systems, critical in domains like autonomous vehicles, are vulnerable to adversarial examples (crafted inputs designed to mislead classifiers). This study investigates black-box adversarial attacks in computer vision. This is a realistic scenario, where attackers have query-only access to the target model. Three properties are introduced to evaluate attack feasibility: robustness to compression, stealthiness to automatic detection, and stealthiness to human inspection. State-of-the-Art methods tend to prioritize one criterion at the expense of others. We propose ECLIPSE, a novel attack method employing Gaussian blurring on sampled gradients and a local surrogate model.

Unmanned Aerial Vehicles (UAVs) are one of the most revolutionary inventions of 21st century. At the core of a UAV lies the central processing system that uses wireless signals to control their movement. The most popular UAVs are quadcopters that use a set of four motors, arranged as two on either side with opposite spin. An autonomous UAV is called a drone. Drones have been in service in the US army since the 90's for covert missions critical to national security. It would not be wrong to claim that drones make up an integral part of the national security and provide the most valuable service during surveillance operations. While UAVs are controlled using wireless signals, there reside some challenges that disrupt the operation of such vehicles such as signal quality and range, real time processing, human expertise, robust hardware and data security. These challenges can be solved by programming UAVs to be autonomous, using object detection and tracking, through Computer Vision algorithms. Computer Vision is an interdisciplinary field that seeks the use of deep learning to gain a high-level understanding of digital images and videos for the purpose of automating the task of human visual system. Using computer vision, algorithms for detecting and tracking various objects can be developed suitable to the hardware so as to allow real time processing for immediate judgement. This paper attempts to review the various approaches several authors have proposed for the purpose of autonomous navigation of UAVs by through various algorithms of object detection and tracking in real time, for the purpose of applications in various fields such as disaster management, dense area exploration, traffic vehicle surveillance etc.

This position paper states that AI Alignment in Multi-Agent Systems (MAS) should be considered a dynamic and interaction-dependent process that heavily depends on the social environment where agents are deployed, either collaborative, cooperative, or competitive. While AI alignment with human values and preferences remains a core challenge, the growing prevalence of MAS in real-world applications introduces a new dynamic that reshapes how agents pursue goals and interact to accomplish various tasks. As agents engage with one another, they must coordinate to accomplish both individual and collective goals. However, this complex social organization may unintentionally misalign some or all of these agents with human values or user preferences. Drawing on social sciences, we analyze how social structure can deter or shatter group and individual values. Based on these analyses, we call on the AI community to treat human, preferential, and objective alignment as an interdependent concept, rather than isolated problems. Finally, we emphasize the urgent need for simulation environments, benchmarks, and evaluation frameworks that allow researchers to assess alignment in these interactive multi-agent contexts before such dynamics grow too complex to control.

We propose Data Swarms, an algorithm to optimize the generation of synthetic evaluation data and advance quantitative desiderata of LLM evaluation. We first train a swarm of initial data generators using existing data, and define various evaluation objectives to reflect the desired properties of evaluation (e.g., generate more difficult problems for the evaluated models) and quantitatively evaluate data generators. We then employ particle swarm optimization to optimize the swarm of data generators, where they collaboratively search through the model parameter space to find new generators that advance these objectives. We further extend it to Adversarial Swarms, where the data generator swarm generates harder data while the test taker model swarm learns from such data, co-evolving dynamically for better data and models simultaneously. Extensive experiments demonstrate that Data Swarms outperforms eight data generation baselines across five evaluation objectives, while Adversarial Swarms produce more robust learning of synthetic data and stronger generalization. Further analysis reveals that Data Swarms successfully optimizes compositions of multiple evaluation objectives and generalizes to new off-the-shelf LLMs, unseen at optimization time.

This study presents a vision-guided robotic control system for automated fruit tree pruning applications. Traditional pruning practices are labor-intensive and limit agricultural efficiency and scalability, highlighting the need for advanced automation. A key challenge is the precise, robust positioning of the cutting tool in complex orchard environments, where dense branches and occlusions make target access difficult. To address this, an Intel RealSense D435 camera is mounted on the flange of a UR5e robotic arm and CoTracker3, a transformer-based point tracker, is utilized for visual servoing control that centers tracked points in the camera view. The system integrates proportional control with iterative inverse kinematics to achieve precise end-effector positioning. The system was validated in Gazebo simulation, achieving a 77.77% success rate within 5mm positional tolerance and 100% success rate within 10mm tolerance, with a mean end-effector error of 4.28 +/- 1.36 mm. The vision controller demonstrated robust performance across diverse target positions within the pixel workspace. The results validate the effectiveness of integrating vision-based tracking with kinematic control for precision agricultural tasks. Future work will focus on real-world implementation and the integration of force sensing for actual cutting operations.

The US administration's targeting of academic research and international students is a "great gift" to China in the race to compete on artificial intelligence, former OpenAI board member Helen Toner has said. The director of strategy at Georgetown's Center for Security and Emerging Technology (CSET) joined the board of OpenAI in 2021 after a career studying AI and the relationship between the United States and China. Toner, a 33-year-old University of Melbourne graduate, was on the board for two years until a falling out with founder Sam Altman in 2023. Altman was fired by the board over claims that he was not "consistently candid" in his communications and the board did not have confidence in Altman's ability to lead. The chaotic months that followed saw Altman fired and then re-hired with three members of the board, including Toner, ousted instead.

The'Outnumbered' panel reacts to the House Oversight Committee's move to subpoena former President Joe Biden's doctor amid concerns about his mental fitness during his time in office. Former President Barack Obama's White House physician said in a new interview that former President Joe Biden's doctor should have performed a cognitive test to evaluate his fitness to serve in office. Obama's doctor, Jeffrey Kuhlman, told The Washington Post that Biden White House physician Kevin O'Connor should have performed a cognitive test during Biden's last year as president, given his age. O'Connor, who Kuhlman first appointed as Biden's doctor in 2009 when he was vice president, declared in a 2024 report that the then-81-year-old president "continues to be fit for duty." The report did not mention any neurocognitive testing.