Chinese Company Leads ECCV Workshop! Multimodal AI Luminaries Take the Stage as 64 Global Teams Tackle Challenges
During ECCV 2026, Tec-Do took the lead in organizing the MARS2 Workshop, which focused on multimodal reasoning and slow thinking in the era of large models. It attracted 64 teams worldwide and was the only Agentic Commerce workshop led by a Chinese company.
Key Takeaways
- Key Highlight：During ECCV 2026, Tec-Do took the lead in organizing the MARS2 Workshop, which focused on multimodal reasoning and slow thinking in the era of large models. It attracted 64 teams worldwide and was the only Agentic Commerce workshop led by a Chinese company.
- Innovation & Tech：Highlights advancements in Agent, Chinese, Company, demonstrating rapid progress in model capabilities.
- Industry Impact：Reported via 雷峰网 (CN), offering actionable signals for developers and technology leaders.
ECCV 2026, one of the world's top computer vision conferences, will be held in Malmö, Sweden, from September 8 to 12, with technology giants Google, Meta, Apple, and Amazon as main sponsors. Among the many workshops at the conference, the MARS2 Workshop was initiated by the Chinese technology company Tec-Do. It focuses on the application of multimodal reasoning and slow thinking in the era of large models, making it the only academic event in the Agentic Commerce direction led by a Chinese company.
The MARS2 Workshop, whose full name is the Second "Multimodal Reasoning and Slow Thinking in the Large Model Era" workshop, has a theme closely aligned with the current trend of large models evolving from quick answers to deep reasoning. The so-called "slow thinking" refers to a model actively performing multi-step reasoning, self-checking, and reflection when handling complex tasks. This is consistent with the technical approach of models such as OpenAI o1 and DeepSeek-R1, and is also a key capability for multimodal agents to make reliable business decisions.
The workshop attracted 64 teams from around the world to join in solving the challenges. The organizing committee invited a number of well-known researchers in the multimodal AI field to take the stage and share their work. From an organizational perspective, Tec-Do's leading role means that Chinese companies' role at international academic conferences is shifting from attendees to organizers and topic setters. Especially in Agentic Commerce, which combines agents with business scenarios, Chinese companies already have a certain degree of academic and engineering appeal.
This event affects the AI industry in two ways. First, academic attention on multimodal reasoning and slow thinking continues to heat up; related benchmarks and competitions will push the capability boundaries of models on complex vision-language tasks. Second, now that the Agentic Commerce direction has received the endorsement of a top conference, the exploration of agent applications in business scenarios such as e-commerce, marketing, and customer service is expected to accelerate, and Chinese AI companies' practical experience in this field may also feed back into the global research community through academic exchanges.
However, details of the MARS2 Workshop's specific evaluation tasks, participating teams' technical solutions, and final results have yet to be disclosed. Whether the workshop can attract more international teams in its next edition, and whether slow-thinking technology can truly improve the commercial conversion efficiency of multimodal agents, still require more public data and industrial validation.
This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.
Industry Insights & Analysis
As artificial intelligence rapidly evolves, breakthroughs surrounding Agent, Chinese, Company, Leads are shifting toward scalable, robust real-world implementations.
Driven by both open-source ecosystems and proprietary model architectures, the integration between compute optimization, data engineering, and agentic workflows is accelerating. This development provides a strategic benchmark for upcoming AI tooling and developer workflows.