Brief

Lawmakers are increasingly focused on regulating the development of AI systems that can create other AI systems, a process known as AI-builds-AI. This concern stems from the potential for such recursive self-improvement to lead to AI that is beyond human control, raising fears of existential risk. While some advocate for a global pause on this type of AI advancement, policymakers are exploring new legislation to govern its development, though the effectiveness and potential overreach of such laws remain subjects of debate. AI

Adam Foroughi, CEO of AI advertising company AppLovin, has led the company to a valuation near $200 billion despite a personal style that defies typical Silicon Valley founder archetypes. After facing initial rejection from top VCs and a severe stock price drop in 2022, Foroughi implemented unconventional strategies, including cutting investor relations, aggressively repurchasing stock with debt, and rebuilding the core advertising recommendation system (Axon 2.0). These moves, coupled with a focus on efficiency and hiring top-tier talent, have resulted in a remarkable market rebound. AI

The US Department of Defense has expanded its blacklist to include 188 Chinese technology firms, citing their alleged ties to the Chinese military. This move, which adds companies like Alibaba, Baidu, and BYD, raises concerns about future investment restrictions and reputational damage for these firms. The expanded list spans various sectors, including AI, electric vehicles, and biotechnology, and has already impacted investor sentiment and stock prices in Hong Kong. AI

Yunzhongma, a company focused on specific object issuance, has received approval from the Shanghai Stock Exchange for its stock issuance application. This approval signifies that the company meets the necessary conditions for issuance, listing, and information disclosure. The application will now proceed to the China Securities Regulatory Commission for registration. AI

Meta's Ray-Ban smart glasses reportedly included a facial recognition feature that the company planned to launch during a period of political instability. Internal documents suggest Meta aimed to release this feature when civil society groups would be too preoccupied to mount a strong opposition. The memo also indicated a desire to avoid scrutiny by launching when "resources focused on other concerns." AI

HelioLink has launched as the first solar-powered orbital AI infrastructure, aiming to create a space-based computing ecosystem. This initiative emphasizes collaboration through open standards and interoperable systems to build scalable, autonomous infrastructure beyond Earth. The project seeks to accelerate the development of computing capabilities in space. AI

China is planning a massive $295 billion investment in a national data center network, aiming to significantly reduce reliance on American chips. This initiative is part of a broader strategy to bolster its domestic AI capabilities. Concurrently, Taiwan is implementing strict regulations that will criminalize the smuggling of AI technology into China. AI

OpenAI and Anthropic have jointly called for the establishment of an international body to oversee AI development and deployment. This proposed watchdog would aim to ensure safety and responsible practices across the global AI landscape. The initiative reflects a growing consensus among leading AI labs about the need for external governance. AI

Researchers have published a paper detailing a generalized universal approximation theorem for neural networks. This new theorem extends previous work by enabling the approximation of not only functions but also their derivatives. The findings are applicable to differentiable maps on infinite-dimensional manifolds and have implications for approximating non-anticipative functionals and path space functionals. AI

The UK government plans to proceed with its crackdown on tech platforms, including a proposed ban on social media for individuals under 16. This initiative will move forward despite opposition from the US embassy, which expressed concerns that the ban could impact American companies. The UK government has indicated that US displeasure will not deter their regulatory actions. AI

A new bill proposed by Representatives Obernolte and Trahan marks a significant step in US federal AI policy. This legislation acknowledges the need for a unified national framework, moving beyond fragmented state-level regulations and temporary executive orders. The bill's introduction signifies a growing recognition within Congress of AI's importance and the necessity for comprehensive federal oversight. AI

Researchers have introduced Topological Neural Operators (TNOs), a new framework for learning operators on cell complexes. TNOs extend existing neural operators by modeling interactions through Discrete Exterior Calculus, allowing for explicit cross-dimensional coupling. This approach respects the geometric properties of physical quantities and can improve accuracy on Partial Differential Equation benchmarks, especially for complex flow problems. AI

Researchers have developed a new technique to enhance reinforcement learning (RL) policies by leveraging existing suboptimal baseline policies. This method gradually transfers control from the baseline to a trainable learning policy, improving training efficiency and ultimately producing a standalone policy that outperforms the original baseline. The approach is formalized with theoretical analysis and demonstrated through empirical results on continuous-control benchmarks, showing high goal-reaching rates throughout the training process. AI

Researchers have introduced iMaC (Image as Action Control), a new paradigm for embodied world models in robotics. This approach uses raw visual images as action representations, moving away from traditional low-dimensional vectors. iMaC aims to improve generalization, dynamic modeling, and control for diverse robotic agents by treating visual manipulation as image-based action tokens. AI

Researchers have developed a new method for representing complex appearance effects in 3D scene reconstruction, moving beyond traditional Spherical Harmonics (SH). Their work introduces the Normalized Anisotropic Spherical Gabor function, which efficiently models high-frequency details like specular reflections and glints. This new formulation offers improved reconstruction quality while being significantly more memory-efficient and faster to evaluate than existing approaches. AI

Researchers have developed a new automated architecture for time-series prediction in volatile cloud-edge environments. This system addresses the "cold start" problem for newly discovered nodes by merging sparse local telemetry data with a high-resolution public dataset called TimeTrack. A Neural Architecture Search engine then generates accurate baseline models, significantly improving forecasting accuracy and convergence speed. AI

Researchers have developed POTATR, a new lightweight image-to-graph model for extracting tables from documents. This 29 million parameter model significantly outperforms existing methods on the PubTables-v2 benchmark, achieving a GriTS_Con score of 0.964. POTATR is also considerably faster and more cost-effective than current large language models, with its output being spatially grounded for verification and further integration. AI

Researchers have developed a new method called Intervention-Aware Variational Quantum Differentiable Predictive Control (IA-VQC-DPC) to better measure the safety contributions of AI policies versus their protective layers. This approach trains quantum circuit policies with a budget that penalizes over-reliance on safety filters. Evaluations on building control emulators demonstrated that IA-VQC-DPC significantly reduces pre-filter violations and reliance on safety layers, indicating improved policy-level safety. AI

Researchers have developed a novel data synthesis method to create neural machine translation (NMT) models for low-resource Indigenous languages, specifically Q'eqchi' Mayan. By transforming dictionaries into a synthetic corpus and using Parameter-Efficient Fine-Tuning (PEFT) with LoRA adapters on an mT5-base model, they achieved strong structural acquisition. However, the resulting model showed a significant gap in lexical grounding compared to organic language, indicating that while synthetic data is effective for learning grammar, authentic data is crucial for semantic refinement. AI

Researchers have developed SemDINO, a new network designed for semantic change detection in remote sensing imagery. This model integrates a dual-branch encoder using CNNs and frozen DINOv3 features, along with a multi-scale temporal interaction module. SemDINO also incorporates modules for semantic purification and change enhancement to improve accuracy and robustness against pseudo-changes. AI

Researchers have developed Topo-Omni, a novel topographic multimodal model that integrates visual, auditory, and language processing onto a single in-silico sheet. This model, fine-tuned from a pretrained foundation model, demonstrates that a unified spatial principle can organize representations across different modalities and processing stages. The model's clusters align with human neuroimaging findings, and manipulating these clusters selectively impacts perception, offering testable hypotheses about cortical organization. AI

Researchers have developed a theoretical framework to understand when optimizing optical front-ends with neural network back-ends improves imaging classification performance. The study found that these gains are most significant under constrained detector readout, such as limited measurements or coarse sampling, by enhancing class separability. However, under full detector readout, conventional lenses perform comparably, and joint optimization offers no empirical advantage. The research also highlights that these optical-neural network co-designs are most effective with low detector noise and when discriminative content is concentrated at lower spatial frequencies. AI

Apple is reportedly pushing back against European regulators regarding the Digital Markets Act (DMA). The company is seeking to avoid making significant changes to its core AI features, such as Siri, to comply with the DMA's interoperability requirements. Apple argues that altering these features could compromise user privacy and security, and potentially impact the user experience. AI

Researchers have introduced iOSWorld, a new benchmark designed to evaluate the personalization capabilities of AI agents on mobile devices. This benchmark features a simulated iOS environment with 26 interconnected apps that store user-specific data like messages and financial records. It includes 133 tasks, ranging from single-app operations to complex multi-app scenarios requiring memory and personalization inference. Initial evaluations show that even advanced models struggle with these tasks, with the best configuration achieving only 52% overall accuracy. AI

Researchers have developed a new imitation learning method called Difference-Aware Retrieval Policies (DARP). This approach improves generalization by using training data during inference, predicting actions based on k-nearest neighbors and their relative distances to query states. DARP achieves significant performance gains over standard behavior cloning in various domains, including robotics and continuous control. AI

Researchers have introduced Perturbative Contrastive Physical Learning (PCPL), a new framework where learning arises from contrasting how physical systems respond to slight variations. This approach unifies and extends existing methods like Equilibrium Propagation and Frequency Propagation. PCPL allows learning without centralized gradient computation, as the learning geometry emerges implicitly from the system's physical response. AI

Researchers have developed a new framework called Spatio-Temporal Bound Propagation (STBP) to improve the verification of neural networks used in safety-critical applications like autonomous driving and medical imaging. This method models adversarial perturbations with more realistic spatio-temporal constraints, leading to tighter approximations and better robustness guarantees than existing techniques. The framework also introduces ST-Bench, a new benchmark designed to systematically evaluate verifiable robustness in these domains. AI

Researchers have developed a new framework for understanding the training dynamics of feed-forward ReLU neural networks. Their work rewrites gradient descent not as a weight-space dynamic, but as a collective dynamic on the training-set space. For deeper networks, this reveals a hierarchical structure of weight-induced operators that manage information flow between layers. AI

Researchers have developed a new framework for estimating gradients in parameterized quantum circuits (PQCs) that significantly reduces the measurement cost associated with training. This approach, based on the forward mode of automatic differentiation, offers an unbiased gradient estimator by averaging random directional derivatives. The proposed QUIVER optimizer, derived from this framework, demonstrates orders of magnitude greater efficiency in training quantum neural networks compared to the standard parameter-shift rule, outperforming other measurement-frugal optimizers on various quantum algorithms. AI

Researchers have developed a novel unsupervised learning algorithm for neural disentanglement using holographic reduced representations (HRR). This approach treats disentangled representations as symbolic structures, moving away from continuous representations common in prior work. The HRR unbinding operation demonstrates an inductive bias for separating factors, achieving competitive results on disentanglement metrics and showing robustness to noise. AI

Researchers have precisely defined the VC dimension for depth-L Transformers with W parameters, establishing an upper bound of O(LW log(TW)) and a nearly matching lower bound. The study also characterizes the sample complexity for chain-of-thought learning with these Transformers, showing teacher forcing achieves O(LW log((T+T')W)) complexity. Any learning rule utilizing chain-of-thought data requires at least \Omega(LW log((T+T')W/L)) examples. AI

Researchers have published a paper detailing methods for converting privacy parameters between pure differential privacy ($\varepsilon$) and Gaussian differential privacy (GDP, $\mu$). The study proposes principled mappings by aligning worst-case membership inference attack success rates across three metrics. The authors recommend a general-purpose conversion of $\mu \approx \varepsilon/5$ for conservative privacy reporting in machine learning. AI

Researchers have introduced PRIME, a new capability that assesses task correctness and predicts proxy acceptance in AI models. This capability emerges before visible reward hacking occurs and can forecast the onset and severity of such issues. PRIME adapts to changing evaluators and can serve as an early warning signal for alignment risks in AI systems. AI

Researchers have developed BrainSurgery, a new tool designed to simplify the complex process of modifying large deep learning model weights. This system allows for reproducible "tensor surgery" through declarative YAML plans, abstracting away storage and memory management challenges. BrainSurgery supports various modifications, including structural changes and mathematical transformations, with built-in assertions to prevent errors and ensure reliability. AI

Researchers have developed a new diagnostic theory and benchmark to understand how well local score models can extrapolate across different system sizes. They found that architectural locality alone is insufficient for stable size extrapolation, which is instead governed by the quasi-locality of the Gaussian-smoothed score. The study introduces the Finite-Depth Local Flow (FDLF) benchmark to empirically validate these findings, demonstrating that stable extrapolation depends on the interplay between spatial mixing, score quasi-locality, and model receptive fields. AI

Researchers have developed a new method called Human-Perceptible Adversarial Attacks (HPAA) that exploits the difference between human and large language model (LLM) perception of harmful content. By using typographic manipulations like spacing and visual emphasis, these attacks can make harmful text easily recognizable to humans while remaining undetected by LLM-based moderation systems. In tests, HPAA achieved over 86% human recognition with less than 1% detection by moderation systems, revealing a significant vulnerability in current content moderation. AI

Researchers have developed AdvGRPO, a novel co-training framework designed to enhance the adaptive red teaming of language models. This method addresses the instability of GRPO in attacker-defender optimization by employing dense multi-channel rewards and decoupled advantage normalization. The training process follows a curriculum, starting with single-turn attacks and progressing to multi-turn scenarios before initiating co-training, ultimately producing more effective attacks and robust defenders. AI

A new paper introduces a comprehensive catalog of 84 numeric formats used in machine learning hardware, addressing the challenge of silent divergences when porting models across different accelerators. The catalog includes bit-exact conformance packs for various formats like FP8, BF16, and MXFP4, serving as a vendor-neutral reference. This work aims to provide a shared standard for engineers to diagnose and resolve discrepancies, ensuring greater consistency in model performance across diverse hardware. AI

Researchers have developed Cranio-Diff, a novel diffusion-based framework for reconstructing faces from 2D X-ray skull images. This method addresses limitations in existing generative models by integrating skull-conditioned structural guidance and biometric text conditioning to ensure semantic and structural alignment between the skull and the generated face. The framework was evaluated on a unique dataset of 120 subjects, generating synthesized faces across different age groups and BMI variations, and demonstrated superior performance in image quality and retrieval tasks compared to existing approaches, suggesting its utility in forensic investigations. AI

Researchers have introduced MeCo, a novel one-step generative corrector for multi-channel speech separation. This method uses a MeanFlow-based approach to map estimated audio directly to clean speech, aiming to improve human listening quality beyond traditional discriminative models. MeCo incorporates Data-Space Optimization with an $\mathbf{x}_r$-loss and an Endpoint SI-SDR loss to enhance both signal fidelity and subjective listening experience. AI

Researchers have developed AutoMegaKernel (AMK), a system that compiles HuggingFace Llama-family models into a single, persistent CUDA kernel for efficient forward passes. AMK's static validator ensures schedule safety, preventing deadlocks and race conditions. The system supports multiple NVIDIA GPU architectures from a single codebase and has demonstrated self-improvement capabilities. AI

Researchers have developed Trellis, an autoformalization system designed to assist in creating rigorous mathematical proofs. The system utilizes LLM agents within a structured workflow to refine natural language proofs incrementally. Trellis aims for reliable formalization with generalist agents by enforcing a process semantics inspired by the notion of mathematical rigor. AI

Researchers have developed a new sampling approach called CFips for exploring large pattern spaces, specifically focusing on interval patterns with user-defined constraints. This method integrates constraints directly into the sampling procedure, decomposing them into elementary predicates on interval bounds to ensure exact sampling guarantees. Experimental results indicate that CFips can successfully complete mining tasks that might otherwise fail due to time constraints. AI

Researchers have developed FMplex, a novel system designed to optimize the serving of foundation models (FMs) by treating them as a virtualization substrate. This approach allows multiple downstream tasks to share a single physical FM instance, reducing memory waste and amortizing costs associated with batching and loading. FMplex enables task-specific extensions and isolation while improving efficiency, demonstrated by significant reductions in latency and increased task hosting capacity. AI

Researchers have developed GenEyePose, a novel pipeline for generating synthetic eye movement data to train AI models for neurophysiologic biomarker development. This approach addresses the scarcity of real-world clinical data and privacy concerns associated with eye-tracking studies. A deep learning classifier trained on this synthetic data demonstrated promising performance in distinguishing normal from abnormal saccadic eye movements, showing potential for clinical applications in screening and localization of brain abnormalities. AI

Researchers have introduced a new framework to unify and clarify concept-based representational similarity in machine learning models. The framework decomposes alignment into representation vs. concept and instance-wise vs. distributional levels, identifying four key properties. They also developed an intervention-based benchmark called \InterVenchA to measure these properties and proposed the Coupled Sparse Autoencoder (CoSAE) method, which demonstrates that strong alignment emerges when multiple objectives are jointly enforced, even with minimal paired data. AI

Researchers have developed an enhanced system for the SoccerNet 2026 Player-Centric Ball-Action Spotting Challenge. Their approach builds upon existing FOOTPASS baselines by incorporating gradient checkpointing for efficient fine-tuning, fusing graph neural network (GNN) outputs with visual features, and applying square-root frequency class weighting to balance imbalanced training data. The system achieved a Macro F1 score of 0.548 on the test set and 0.446 on the challenge set. AI

Researchers have developed a new benchmark to evaluate how well multimodal large language models (MLLMs) identify the correct visual evidence for their answers, particularly in autonomous driving scenarios. The benchmark uses synchronized multi-view driving data from NuScenes, presenting models with questions and requiring them to pinpoint the supporting camera view before answering. This approach aims to expose grounding failures that traditional answer-only evaluations might miss, by explicitly separating evidence identification from response accuracy. AI

A new research paper introduces the concept of "MetaAI Recursive Self-Design," defining it as an AI-assisted development pattern where the AI itself modifies its building and improvement mechanisms. The paper proposes a framework to evaluate such systems and highlights the Darwin Goedel Machine (DGM) as a prime example, showing significant performance gains on coding benchmarks after 80 iterations. To facilitate further research, the authors also release MetaAI-Mini, a reproducible protocol and codebase based on HumanEval. AI

Researchers have developed a new approach to anomaly detection that addresses limitations in real-world scenarios where object scale, viewpoint, and background vary. Their method incorporates a visual prompting pipeline for object isolation, a technique to unfreeze teachers in student-teacher models for better domain adaptability, and data augmentation using diffusion-generated images. This approach achieved a 3.5 percentage point improvement over the previous state-of-the-art on the AeBAD dataset. AI