RMBench: Memory-Dependent Robotic Manipulation Benchmark with Insights into Policy Design

Tianxing Chen*, Yuran Wang*, Mingleyang Li*, Yan Qin*, Hao Shi, Zixuan Li, Yifan Hu, Yingsheng Zhang, Kaixuan Wang, Yue Chen, Hongcheng Wang, Renjing Xu, Ruihai Wu, Yao Mu, Yaodong Yang, Hao Dong✉, Ping Luo✉

Under Review 2026 | Paper | Code | Homepage | Huggingface

• RMBench is a memory-oriented benchmark built on the RoboTwin platform, and it also provides a memory-enhanced hierarchical VLA model, Mem-0.

MemoryVLA: Perceptual-Cognitive Memory in Vision-Language-Action Models for Robotic Manipulation

Hao Shi, Bin Xie, Yingfei Liu, Lin Sun, Fengrong Liu, Tiancai Wang, Erjin Zhou, Haoqiang Fan, Xiangyu Zhang, Gao Huang✉

ICLR 2026 | Paper | Code | Homepage | Huggingface

• MemoryVLA is among the early works exploring memory in Vision-Language-Action models. Inspired by human memory systems, it builds a hippocampal-like memory to capture the temporal dependencies. It has since been cited by numerous research institutions, including Physical Intelligence.

SpatialActor: Exploring Disentangled Spatial Representations for Robust Robotic Manipulation

Hao Shi, Bin Xie, Yingfei Liu, Yang Yue, Tiancai Wang, Haoqiang Fan, Xiangyu Zhang, Gao Huang✉

AAAI 2026 Oral (Accept rate≈4%) | Paper | Code | Homepage | Huggingface

• SpatialActor is a disentangled spatial representations framework for robust robotic manipulation.

Grounding Beyond Detection: Enhancing Contextual Understanding in Embodied 3D Grounding

Yani Zhang*, Dongming Wu*, Hao Shi, Yingfei Liu, Tiancai Wang, Haoqiang Fan, Xingping Dong✉

CVPR 2026 Findings | Paper | Code

• DEGround is an embodied perception framework for 3D grounding, achieving 1st place on EmbodiedScan.

Dexbotic: Open-Source Vision-Language-Action Toolbox

Dexbotic Team

Technical Report 2025 | Paper | Code | Homepage | Huggingface

• Dexbotic is an open-source VLA codebase (similar to MMDetection). It unifies multiple mainstream VLA frameworks and benchmarks, provides strong pretrained models, and has garnered nearly 1000 GitHub stars.

GeoVLA: Enpowering 3D Representations in Vision-Language-Action Models

Under Review 2025 | Paper | Code | Homepage

Lin Sun*, Bin Xie*, Yingfei Liu, Hao Shi, Tiancai Wang, Jiale Cao✉

• GeoVLA is a framework that bridges 2D semantics and 3D geometry for VLA, it achieves robustness across diverse camera views, object heights, and sizes.

DenseGrounding: Improving Dense Language-Vision Semantics for Ego-centric 3D Visual Grounding

Henry Zheng*, Hao Shi*, Qihang Peng, Yong Xien Chng, Rui Huang, Yepeng Weng, Zhongchao Shi, Gao Huang✉

*: equal contribution, ✉: corresponding author.

ICLR 2025 | Paper | Code

• DenseGrounding is an embodied perception framework for multi-view 3D visual grounding, and it achieved first place on EmbodiedScan.

DenseG: Alleviating Vision-Language Feature Sparsity in Multi-View 3D Visual Grounding

Henry Zheng*, Hao Shi*, Yong Xien Chng, Rui Huang, Zanlin Ni, Tianyi Tan, Qihang Peng, Yepeng Weng, Zhongchao Shi, Gao Huang✉

*: equal contribution, ✉: corresponding author.

CVPR 2024 Workshop (Oral) | Paper | Code

• 1st Place and Innovation Award in CVPR 2024 Autonomous Grand Challenge, Embodied 3D Grounding Track (1/64 teams, 1/154 submissions).

Open Compound Domain Adaptation with Object Style Compensation for Semantic Segmentation

Tingliang Feng*, Hao Shi*, Xueyang Liu, Wei Feng, Liang Wan, Yanlin Zhou, Di Lin✉

*: equal contribution, ✉: corresponding author.

NeurIPS 2023 | Paper | Code

• We propose a memory-bank-based object-style compensation method for open compound domain adaptation.