Conventions are crucial for strong performance in cooperative multi-agent games, because they allow players to coordinate on a shared strategy without explicit communication. Unfortunately, standard multi-agent reinforcement learning techniques, such as self-play, converge to conventions that are arbitrary and non-diverse, leading to poor generalization when interacting with new partners. In this work, we present a technique for generating diverse conventions by (1) maximizing their rewards during self-play, while (2) minimizing their rewards when playing with previously discovered conventions (cross-play), stimulating conventions to be semantically different. To ensure that learned policies act in good faith despite the adversarial optimization of cross-play, we introduce mixed-play, where an initial state is randomly generated by sampling self-play and cross-play transitions and the player learns to maximize the self-play reward from this initial state. We analyze the benefits of our technique on various multi-agent collaborative games, including Overcooked, and find that our technique can adapt to the conventions of humans, surpassing human-level performance when paired with real users.

Humans communicate with graphical sketches apart from symbolic languages. Primarily focusing on the latter, recent studies of emergent communication overlook the sketches; they do not account for the evolution process through which symbolic sign systems emerge in the trade-off between iconicity and symbolicity. In this work, we take the very first step to model and simulate this process via two neural agents playing a visual communication game; the sender communicates with the receiver by sketching on a canvas. We devise a novel reinforcement learning method such that agents are evolved jointly towards successful communication and abstract graphical conventions. To inspect the emerged conventions, we define three key properties -- iconicity, symbolicity, and semanticity -- and design evaluation methods accordingly. Our experimental results under different controls are consistent with the observation in studies of human graphical conventions. Of note, we find that evolved sketches can preserve the continuum of semantics under proper environmental pressures. More interestingly, co-evolved agents can switch between conventionalized and iconic communication based on their familiarity with referents. We hope the present research can pave the path for studying emergent communication with the modality of sketches.

AI coding is now everywhere. But not everyone is convinced. Developers are navigating confusing gaps between expectation and reality. So are the rest of us. Depending who you ask, AI-powered coding is either giving software developers an unprecedented productivity boost or churning out masses of poorly designed code that saps their attention and sets software projects up for serious long term-maintenance problems. The problem is right now, it's not easy to know which is true. As tech giants pour billions into large language models (LLMs), coding has been touted as the technology's killer app. Both Microsoft CEO Satya Nadella and Google CEO Sundar Pichai have claimed that around a quarter of their companies' code is now AI-generated. And in March, Anthropic's CEO, Dario Amodei, predicted that within six months 90% of all code would be written by AI.

Abstract--We address the correspondence-free alignment of two rotation sets on SO(3), a core task in calibration and registration that is often impeded by missing time alignment, outliers, and unknown axis conventions. T o handle axis relabels and sign flips, we introduce a Permutation-and-Sign Invariant (PASI) wrapper that enumerates the 24 proper signed permutations, scores them via summed correlations, and fuses the per-axis estimates into a single rotation by projection/Karcher mean. Experiments on EuRoC Machine Hall simulations (axis-consistent) and the ETH Hand-Eye benchmark (robot_arm_real) (axis-ambiguous) show that our methods are accurate, 6-60x faster than traditional methods, and robust under extreme outlier ratios (up to 90%), all without correspondence search. Estimating the 3D rotation that aligns one sensor or object frame to another is a fundamental problem in robotics and computer vision. Closed-form or least-squares solutions (e.g., Davenport/QUEST, SVD/Procrustes, and modern quaternion solvers) are mature [25], [26], [27], [28], [29], but they typically assume paired measurements (known correspondences) and degrade under heavy outliers or axis-convention mismatches.

AI/ML models have rapidly gained prominence as innovations for solving previously unsolved problems and their unintended consequences from amplifying human biases. Advocates for responsible AI/ML have sought ways to draw on the richer causal models of system dynamics to better inform the development of responsible AI/ML. However, a major barrier to advancing this work is the difficulty of bringing together methods rooted in different underlying assumptions (i.e., Dana Meadow's "the unavoidable a priori"). This paper brings system dynamics and structural equation modeling together into a common mathematical framework that can be used to generate systems from distributions, develop methods, and compare results to inform the underlying epistemology of system dynamics for data science and AI/ML applications.

Climate science demands automated workflows to transform comprehensive questions into data-driven statements across massive, heterogeneous datasets. However, generic LLM agents and static scripting pipelines lack climate-specific context and flexibility, thus, perform poorly in practice. We present ClimateAgent, an autonomous multi-agent framework that orchestrates end-to-end climate data analytic workflows. ClimateAgent decomposes user questions into executable sub-tasks coordinated by an Orchestrate-Agent and a Plan-Agent; acquires data via specialized Data-Agents that dynamically introspect APIs to synthesize robust download scripts; and completes analysis and reporting with a Coding-Agent that generates Python code, visualizations, and a final report with a built-in self-correction loop. To enable systematic evaluation, we introduce Climate-Agent-Bench-85, a benchmark of 85 real-world tasks spanning atmospheric rivers, drought, extreme precipitation, heat waves, sea surface temperature, and tropical cyclones. On Climate-Agent-Bench-85, ClimateAgent achieves 100% task completion and a report quality score of 8.32, outperforming GitHub-Copilot (6.27) and a GPT-5 baseline (3.26). These results demonstrate that our multi-agent orchestration with dynamic API awareness and self-correcting execution substantially advances reliable, end-to-end automation for climate science analytic tasks.

At New Zealand's Kawaiican cybersecurity convention, organizers hacked together a way for attendees to track CO levels throughout the venue--even before they arrived. Hacker conferences--like all conventions--are notorious for giving attendees a parting gift of mystery illness. To combat "con crud," New Zealand's premier hacker conference, Kawaiicon, quietly launched a real-time, room-by-room carbon dioxide monitoring system for attendees. To get the system up and running, event organizers installed DIY CO monitors throughout the Michael Fowler Centre venue before conference doors opened on November 6. Attendees were able to check a public online dashboard for clean air readings for session rooms, kids' areas, the front desk, and more, all before even showing up. It's ALMOST like we are all nerds in a risk-based industry, the organizers wrote on the convention's website.

We will clarify that'policy hierarchy' in no way refers to HRL.

We could otherwise learn this model via online or offline supervised learning. The mean and standard deviation are shown. In this section, we provide detailed descriptions of each of the experimental domains featured in this work. In addition, we describe our architecture and hyperparameter choices for each setting. The agent's observation consists of two primary elements: The nethack "bottom-line stats" of the game, such as the agent's health stats, attribute levels, armor class, and The convolutions have square kernels of size 2, 2, 2, 2, 3, 3, output channels of dimension 8, 16, 32, 64, 128, 256, and stride lengths of 2, 2, 2, 2, 1, 1.

Players have to manage the ingredients, use the stove, and deliver meals. As the team works together, they decide how tasks should be allocated among themselves so resources are used effectively. For example, player 1 could notice that player 2 tends to stay near the stove, so they instead spend more time preparing ingredients and delivering food, allowing player 2 to continue working at the stove. Through these interactions, the team creates a "convention" in the