Information-processing systems that coordinate multiple agents and objectives face fundamental thermodynamic constraints. We show that solutions with maximum utility to act as coordination focal points have a much higher selection pressure for being findable across agents rather than accuracy. We derive that the information-theoretic minimum description length of coordination protocols to precision $\varepsilon$ scales as $L(P)\geq NK\log_2 K+N^2d^2\log (1/\varepsilon)$ for $N$ agents with $d$ potentially conflicting objectives and internal model complexity $K$. This scaling forces progressive simplification, with coordination dynamics changing the environment itself and shifting optimization across hierarchical levels. Moving from established focal points requires re-coordination, creating persistent metastable states and hysteresis until significant environmental shifts trigger phase transitions through spontaneous symmetry breaking. We operationally define coordination temperature to predict critical phenomena and estimate coordination work costs, identifying measurable signatures across systems from neural networks to restaurant bills to bureaucracies. Extending the topological version of Arrow's theorem on the impossibility of consistent preference aggregation, we find it recursively binds whenever preferences are combined. This potentially explains the indefinite cycling in multi-objective gradient descent and alignment faking in Large Language Models trained with reinforcement learning with human feedback. We term this framework Thermodynamic Coordination Theory (TCT), which demonstrates that coordination requires radical information loss.

We introduce RegionFocus, a visual test-time scaling approach for Vision Language Model Agents. Understanding webpages is challenging due to the visual complexity of GUI images and the large number of interface elements, making accurate action selection difficult. Our approach dynamically zooms in on relevant regions, reducing background clutter and improving grounding accuracy. To support this process, we propose an image-as-map mechanism that visualizes key landmarks at each step, providing a transparent action record and enables the agent to effectively choose among action candidates. Even with a simple region selection strategy, we observe significant performance gains of 28+\% on Screenspot-pro and 24+\% on WebVoyager benchmarks on top of two state-of-the-art open vision language model agents, UI-TARS and Qwen2.5-VL, highlighting the effectiveness of visual test-time scaling in interactive settings. We achieve a new state-of-the-art grounding performance of 61.6\% on the ScreenSpot-Pro benchmark by applying RegionFocus to a Qwen2.5-VL-72B model. Our code will be released publicly at https://github.com/tiangeluo/RegionFocus.

This study proposes a mathematical programming-based algorithm for the integrated selection of variable subsets and bandwidth estimation in geographically weighted regression, a local regression method that allows the kernel bandwidth and regression coefficients to vary across study areas. Unlike standard approaches in the literature, in which bandwidth and regression parameters are estimated separately for each focal point on the basis of different criteria, our model uses a single objective function for the integrated estimation of regression and bandwidth parameters across all focal points, based on the regression likelihood function and variance modeling. The proposed model further integrates a procedure to select a single subset of independent variables for all focal points, whereas existing approaches may return heterogeneous subsets across focal points. We then propose an alternative direction method to solve the nonconvex mathematical model and show that it converges to a partial minimum. The computational experiment indicates that the proposed algorithm provides competitive explanatory power with stable spatially varying patterns, with the ability to select the best subset and account for additional constraints.

Throughout nature, the intricate and disordered lattice structures that are observed in bones, plant stems, dragonfly wings, coral, radiolarians [1], amongst many other examples, demonstrate how powerful geometry is for designing structures with extreme mechanical properties from a very limited selection of base materials [2]. Metamaterials [3] are a recent example of human-engineered lattice structures that utilise the geometric design space of unit cells to change the properties of the lattice obtained by tiling this motive, often producing structures with different properties than those of the underlying lattice material - for instance, having a soft and compressible lattice made of a very brittle material such as ceramic [4]. In addition to metamaterials that follow a periodic design philosophy, there is a growing interest in (inversely) designing disordered lattice materials and structures [5-12], allowing us to fully tap into the functional design space explored by nature. Since lattices can be constructed using additive manufacturing, they combine ease of manufacturing with a highly expressive design space that only requires a small amount of building materials. It is not surprising that lattices have found applications on a variety of scales, ranging from nano-and mesoscale materials to large-scale structures such as space habitats [13-16]. The static nature of lattices also means that once they have been constructed, their properties are fixed - unless physically stimulating the lattice changes the properties of its base materials or allows switching between different shapes (e.g., magnetically [17-19]), therefore enabling a certain degree of reprogrammability of the lattice's properties; also known as active metamaterials [20, 21].

With the long-term goal of reducing the image processing time on an autonomous mobile robot in mind we explore in this paper the use of log-polar like image data with gaze control. The gaze control is not done on the Cartesian image but on the log-polar like image data. For this we start out from the classic deep reinforcement learning approach for Atari games. We extend an A3C deep RL approach with an LSTM network, and we learn the policy for playing three Atari games and a policy for gaze control. While the Atari games already use low-resolution images of 80 by 80 pixels, we are able to further reduce the amount of image pixels by a factor of 5 without losing any gaming performance.

Machine learning techniques are attractive options for developing highly-accurate automated analysis tools for nanomaterials characterization, including high-resolution transmission electron microscopy (HRTEM). However, successfully implementing such machine learning tools can be difficult due to the challenges in procuring sufficiently large, high-quality training datasets from experiments. In this work, we introduce Construction Zone, a Python package for rapidly generating complex nanoscale atomic structures, and develop an end-to-end workflow for creating large simulated databases for training neural networks. Construction Zone enables fast, systematic sampling of realistic nanomaterial structures, and can be used as a random structure generator for simulated databases, which is important for generating large, diverse synthetic datasets. Using HRTEM imaging as an example, we train a series of neural networks on various subsets of our simulated databases to segment nanoparticles and holistically study the data curation process to understand how various aspects of the curated simulated data -- including simulation fidelity, the distribution of atomic structures, and the distribution of imaging conditions -- affect model performance across several experimental benchmarks. Using our results, we are able to achieve state-of-the-art segmentation performance on experimental HRTEM images of nanoparticles from several experimental benchmarks and, further, we discuss robust strategies for consistently achieving high performance with machine learning in experimental settings using purely synthetic data.

The transform is fundamental tool in science, but also is how your eyes see the world. There is a great YouTube channel called 3Blue1Brown. I would bet that you've heard of it, but in case not, it is a very popular math education channel by Grant Sanderson that heavily uses visualizations, with a particular style that has noticeably influenced many other YouTubers. I bring this up at the beginning because he recently did a set of videos on the Fourier Transform (FT), with a particular focus on explaining the Fourier-convolution theorem, which is pretty relevant to topics I work on so it caught my interest. Like all of Sanderson's content, it is wonderfully put together, and pertinently it inspired me to finally write up this article about a Fourier-related topic.

Potential advancements in artificial intelligence (AI) could have profound implications for how countries research and develop weapons systems, and how militaries deploy those systems on the battlefield. The idea of AI-enabled military systems has motivated some activists to call for restrictions or bans on some weapon systems, while others have argued that AI may be too diffuse to control. This paper argues that while a ban on all military applications of AI is likely infeasible, there may be specific cases where arms control is possible. Throughout history, the international community has attempted to ban or regulate weapons or military systems for a variety of reasons. This paper analyzes both successes and failures and offers several criteria that seem to influence why arms control works in some cases and not others. We argue that success or failure depends on the desirability (i.e., a weapon's military value versus its perceived horribleness) and feasibility (i.e., sociopolitical factors that influence its success) of arms control. Based on these criteria, and the historical record of past attempts at arms control, we analyze the potential for AI arms control in the future and offer recommendations for what policymakers can do today.

Fact-checking requires retrieving evidence related to a claim under investigation. The task can be formulated as question generation based on a claim, followed by question answering. However, recent question generation approaches assume that the answer is known and typically contained in a passage given as input, whereas such passages are what is being sought when verifying a claim. In this paper, we present {\it Varifocal}, a method that generates questions based on different focal points within a given claim, i.e.\ different spans of the claim and its metadata, such as its source and date. Our method outperforms previous work on a fact-checking question generation dataset on a wide range of automatic evaluation metrics. These results are corroborated by our manual evaluation, which indicates that our method generates more relevant and informative questions. We further demonstrate the potential of focal points in generating sets of clarification questions for product descriptions.