DeliveryBench: Can Agents Earn Profit in Real World?

Lingjun Mao, Jiawei Ren, Kun Zhou, Jixuan Chen, Ziqiao Ma, Lianhui Qin†

COLM 2026

• We present DeliveryBench, a realistic embodied benchmark for food delivery that evaluates long-horizon, constraint-rich decision-making to maximize net profit over hours.

SimWorld: An Open-ended Realistic Simulator for Autonomous Agents in Physical and Social Worlds

Jiawei Ren, Yan Zhuang, Xiaokang Ye*, Lingjun Mao, Xuhong He, Jianzhi Shen, …, Tianmin Shu†, Zhiting Hu†, Lianhui Qin†

Technical Report

• We propose SimWorld Simulator, featuring three key designs: (1) realistic, open-ended world simulation, (2) rich interface for LLM/VLM agents, and (3) diverse physical and social reasoning scenarios

Evaluating Model Perception of Color Illusions in Photorealistic Scenes

Lingjun Mao, Zineng Tang, Alane Suhr

CVPR 2025

• We propose an automated framework for generating realistic color illusion images, build a large-scale dataset (RCID), and systematically investigate the underlying mechanisms by which VLMs are misled by color illusions.

Grounding Language in Multi-Perspective Referential Communication

Zineng Tang, Lingjun Mao, Alane Suhr

EMNLP main 2024

• We introduce a task and dataset for referring expression generation and comprehension in multi-agent embodied environments.

Biomedical Visual Instruction Tuning with Clinician Preference Alignment

Hejie Cui*, Lingjun Mao*, Xin Liang, Jieyu Zhang, Hui Ren, Quanzheng Li, Xiang Li, Carl Yang

NeurIPS 2024

• we propose a data-centric framework (Biomed-VITAl) that incorporates clinician preferences into both stages of generating and selecting instruction data for tuning biomedical multimodal foundation models.

Building a Real-world Autonomous Robotic System with Causality-driven Agent and World Model

Lingjun Mao†, Lukun He†, Jinglin Cao†, Wenpeng Xu†, …, Kun Zhou*, Biwei Huang

Aether AI, 2026

• CRIS-0 is a causality-driven robotic intelligence system that integrates a Causal Agent with a Causal World Model to construct structured causal representations of the physical world, enabling robots to actively explore, intervene in, reason about, and adapt to their environments.