Dexterous grasping remains a fundamental yet challenging problem in robotics. A general-purpose robot must be capable of grasping diverse objects in arbitrary scenarios. However, existing research typically relies on restrictive assumptions, such as single-object settings or limited environments, showing constrained generalization. We present DexGraspVLA, a hierarchical framework for robust generalization in language-guided general dexterous grasping and beyond. It utilizes a pre-trained Vision-Language model as the high-level planner and learns a diffusion-based low-level Action controller. The key insight to achieve generalization lies in iteratively transforming diverse language and visual inputs into domain-invariant representations via foundation models, where imitation learning can be effectively applied due to the alleviation of domain shift. Notably, our method achieves a 90+% dexterous grasping success rate under thousands of challenging unseen cluttered scenes. Empirical analysis confirms the consistency of internal model behavior across environmental variations, validating our design. DexGraspVLA also, for the first time, simultaneously demonstrates free-form long-horizon prompt execution, robustness to adversarial objects and human disturbance, and failure recovery. Extended application to nonprehensile grasping further proves its generality.

Visual Terrain Classification (VTC) plays a vital role in enabling unmanned ground vehicles to understand complex environments. Existing research relies on image-label pairs annotated by static label sets, where semantic ambiguity and high annotation costs constrain fine-grained terrain characterization. These limitations hinder the model’s adaptation to real-world terrain diversity and restrict its applicability. To address these issues, we propose TerraX, a vision-language learning framework that integrates multi-modal image-label-text data, unifying structured annotations with fine-grained natural language descriptions. The framework introduces a composite dataset TerraData, an evaluation benchmark suite TerraBench, and a CLIP-based visual terrain classification model TerraCLIP. TerraData aggregates multi-source terrain images from public and self-collected datasets, annotated through a VLM-based vision-language data annotation pipeline. TerraBench defines three evaluation benchmarks to systematically assess model robustness and adaptability in real-world terrain classification scenarios. Built on the CLIP model, TerraCLIP utilizes multi-granularity contrastive loss and LoRA fine-tuning to enhance understanding for terrain categories and attributes, and incorporates confidence-weighted inference for accurate predictions. Extensive experiments across benchmarks and real-world platforms demonstrate that our approach significantly enhances VTC performance, highlighting its potential for deployment in complex environments.

Frontier AI systems, including large language models (LLMs), hold increasing influence over the epistemology of human users. Such influence can reinforce prevailing societal values, potentially contributing to the lock-in of misguided moral beliefs and, consequently, the perpetuation of problematic moral practices on a broad scale. We introduce progress alignment as a technical solution to mitigate this imminent risk. Progress alignment algorithms learn to emulate the mechanics of human moral progress, thereby addressing the susceptibility of existing alignment methods to contemporary moral blindspots. To empower research in progress alignment, we introduce ProgressGym, an experimental framework allowing the learning of moral progress mechanics from history, in order to facilitate future progress in real-world moral decisions. Leveraging 9 centuries of historical text and 18 historical LLMs, ProgressGym enables codification of real-world progress alignment challenges into concrete benchmarks. Specifically, we introduce three core challenges: tracking evolving values (PG-Follow), preemptively anticipating moral progress (PG-Predict), and regulating the feedback loop between human and AI value shifts (PG-Coevolve). Alignment methods without a temporal dimension are inapplicable to these tasks. In response, we present lifelong and extrapolative algorithms as baseline methods of progress alignment, and build an open leaderboard soliciting novel algorithms and challenges.