HuBE: Cross-Embodiment Human-like Behavior Execution for Humanoid Robots

Achieving both behavioral similarity and appropriateness in human-like motion generation for humanoid robot remains an open challenge, further compounded by the lack of cross-embodiment adaptability. To address this problem, we propose HuBE, a bi-level closed-loop framework that integrates robot state, goal poses, and contextual situations to generate human-like behaviors, ensuring both behavioral similarity and appropriateness, and eliminating structural mismatches between motion generation and execution. To support this framework, we construct HPose, a context-enriched dataset featuring fine-grained situational annotations. Furthermore, we introduce a bone scaling-based data augmentation strategy that ensures millimeter-level compatibility across heterogeneous humanoid robots. Comprehensive evaluations on multiple commercial platforms demonstrate that HuBE significantly improves motion similarity, behavioral appropriateness, and computational efficiency over state-of-the-art baselines, establishing a solid foundation for transferable and human-like behavior execution across diverse humanoid robots.

REF-VLM: Triplet-Based Referring Paradigm for Unified Visual Decoding

Multimodal Large Language Models (MLLMs) demonstrate robust zero-shot capabilities across diverse vision-language tasks after training on mega-scale datasets. However, dense prediction tasks, such as semantic segmentation and keypoint detection, pose significant challenges for MLLMs when represented solely as text outputs. Simultaneously, current MLLMs utilizing latent embeddings for visual task decoding generally demonstrate limited adaptability to both multi-task learning and multi-granularity scenarios. In this work, we present REF-VLM, an end-to-end framework for unified training of various visual decoding tasks. To address complex visual decoding scenarios, we introduce the Triplet-Based Referring Paradigm (TRP), which explicitly decouples three critical dimensions in visual decoding tasks through a triplet structure: concepts, decoding types, and targets. TRP employs symbolic delimiters to enforce structured representation learning, enhancing the parsability and interpretability of model outputs. Additionally, we construct Visual-Task Instruction Following Dataset (VTInstruct), a large-scale multi-task dataset containing over 100 million multimodal dialogue samples across 25 task types. Beyond text inputs and outputs, VT-Instruct incorporates various visual prompts such as point, box, scribble, and mask, and generates outputs composed of text and visual units like box, keypoint, depth and mask. The combination of different visual prompts and visual units generates a wide variety of task types, expanding the applicability of REF-VLM significantly. Both qualitative and quantitative experiments demonstrate that our REF-VLM outperforms other MLLMs across a variety of standard benchmarks. The code, dataset, and demo available at https://github.com/MacavityT/REF-VLM.