Brief

Researchers have introduced DriveMA, a new approach for driving vision-language-action models that replaces complex natural language reasoning with simpler, one-step meta-actions. This method addresses bottlenecks in annotation, model complexity, and inference latency associated with traditional reasoning-centric interfaces. DriveMA achieves new state-of-the-art results on the Waymo End-to-End Driving Challenge, demonstrating the effectiveness of its action-centric supervised training and reinforcement learning framework. AI

Researchers have developed G2D, a novel three-stage pipeline that combines a short online reinforcement learning (RL) warm-up with offline fine-tuning for language models. This approach aims to mitigate the computational expense of continuous online rollouts required by methods like GRPO. By constructing a static preference dataset after a brief GRPO phase and then using DPO for offline training, G2D has shown to match or exceed the performance of GRPO at a significantly reduced compute cost. AI

Researchers have introduced Graph Navier Stokes Networks (GNSN), a new architecture designed to address the oversmoothing problem in Graph Neural Networks. Unlike traditional diffusion-based methods, GNSN incorporates convection to create a dynamic velocity field for more efficient message propagation. This approach allows GNSN to better handle datasets with varying homophily and has demonstrated superior performance on multiple real-world classification tasks. AI

Researchers have introduced Dynamic TMoE, a novel framework designed to improve time series forecasting for non-stationary data. This approach addresses limitations in existing Mixture-of-Experts models by dynamically creating and removing experts based on detected distribution shifts. A temporal memory router further enhances stability by using recurrent states and an anomaly repository for context-aware expert selection, leading to significant performance gains. AI

Researchers have developed a new method called Predicate Action Skills (PACTS) that allows robots to learn and compose skills without retraining. PACTS models both the physical actions and the symbolic outcomes of these actions, enabling better generalization. This approach facilitates zero-shot skill composition through planning by using predicted outcomes to sequence and monitor task execution. AI

Researchers have introduced RankE, a novel end-to-end post-training framework designed to improve discrete text-to-image generation models. Unlike previous methods that kept the VQ decoder frozen, RankE co-evolves both the policy and the decoder through alternating optimization. This approach addresses latent covariate shift, where policy improvements lead to degraded image quality. Experiments on LlamaGen-XL and Janus-Pro models demonstrate that RankE simultaneously enhances both alignment (CLIP score) and image fidelity (FID score), breaking the trade-off seen in earlier techniques. AI

Researchers have developed a new method to improve text-to-image diffusion models for generating human portraits, addressing the common trade-off between text alignment, realism, and aesthetics. Their approach uses a feature supervision paradigm with a lightweight cross-modal alignment mechanism that extracts vision-aligned text representations from SigLIP 2. This method injects guidance into the image generation process without degrading the model's original capabilities or requiring extra inference time, while also optimizing for human-perceived aesthetics. AI

Researchers have developed HRM-Text, a novel Hierarchical Recurrent Model that significantly reduces the computational resources and training data required for pretraining large language models. By decoupling computation into strategic and execution layers and training exclusively on instruction-response pairs, a 1B-parameter model achieved competitive performance on several benchmarks with a fraction of the tokens and compute used by standard models. This approach makes foundational LLM research more accessible by lowering the barrier to entry for pretraining from scratch. AI

Researchers have developed new methods to understand the internal workings of Mixture-of-Experts (MoE) models in computer vision. By analyzing how different visual categories are routed to specific experts and examining the tuning of these experts to various inputs, they found that an animate-inanimate distinction is a dominant factor in expert partitioning. The study reveals that experts tune to broader, continuous visual and semantic dimensions beyond simple category boundaries, highlighting the benefits of moving beyond basic routing analyses for a deeper understanding of MoE specialization. AI

A new research paper proposes the Structural Depth Hypothesis (SDH) to explain how self-training restructures language models. The study found that while surface-level linguistic features like discourse markers increase, deeper syntactic structures such as questions and passives decline. This effect was observed across multiple models and architectures, suggesting it's a specific outcome of self-training rather than a general language model behavior. AI

Researchers have developed DITTO, a new model that learns to simulate human behavior by incorporating verbal feedback as a primary signal in reinforcement learning. This approach, detailed in a new paper, treats subjective and multi-faceted guidance as a first-class input, optimizing for improved rollouts based on this feedback. DITTO demonstrated a 36% improvement over its base model and outperformed GPT-5.4 on six benchmarks within the newly introduced SOUL suite, which comprises ten tasks across various human-like behavior simulations. AI

Researchers have developed a new framework called In-context Training (ICT) to evaluate how language agents can improve their performance on future tasks by learning from past experiences. This approach trains a 'reflector' model to generate system prompts that guide an 'actor' model, enabling cross-task self-improvement without human examples. Experiments in ALFWorld and MiniHack demonstrated that agents trained with ICT outperformed baselines and even generalized to new environments, suggesting that the ability to learn from experience can itself be learned. AI

Researchers have developed a new dataset containing over 260,000 long-form stories, each annotated with creativity scores and review comments based on the Torrance Test of Creative Writing (TTCW). They fine-tuned Qwen3 models on this data to generate literary reviews, finding that models trained without explicit reasoning supervision performed better. The study suggests that for structured, rubric-based review generation, reasoning supervision may not be beneficial and can even lead to irrelevant or repetitive outputs. AI

Researchers have developed CoMET, a novel method for multimodal classification that leverages frozen pre-trained backbones and Tabular Foundation Models (TFMs). This approach uses Principal Component Analysis (PCA) to compress modality embeddings before feeding them into a TFM, eliminating the need for fine-tuning. For improved representation quality, especially when CLS tokens are misaligned, they propose PALPooling, an adaptive token pooler. CoMET achieves state-of-the-art results on various multimodal benchmarks and can handle large-scale datasets with over 500,000 samples and 2,000 classes without any training. AI

MiniMax AI has released over 600 new voices through its Speech 2.8 Turbo model. These voices are now accessible on the Together AI platform. This expansion aims to provide a wider range of synthetic speech options. AI

Users of the Cursor IDE are reporting that the new Composer 2.5 model significantly outperforms previous versions and even larger models like GPT-4.5. Many are finding Composer 2.5 to be faster, more accurate, and notably cheaper, leading them to adopt it as their default for most coding tasks. This shift is reducing their reliance on more expensive, high-end models for everyday development work. AI

Adobe has integrated over 50 of its Creative Cloud tools, including Photoshop and Premiere, directly into the Claude AI chat interface. This allows users to generate and edit complex graphics and videos using natural language prompts, effectively turning Adobe's software into plugins for AI. The integration aims to lower the barrier to entry for professional design, with a free tier offering access to around 40 tools without an Adobe account. AI

OpenAI is highlighting how various companies are integrating its Codex and GPT-5.5 models into their software development workflows. These case studies demonstrate accelerated code review, faster development cycles, and improved code quality across different industries. The company also notes the expansion of its GPT-5.5-Cyber model for vulnerability research and the introduction of a new safety feature, Trusted Contact, within ChatGPT. AI