HiDream.ai Clinches Top Honor at ACM MM
Recently, HiDream.ai received the prestigious Best Demo award at the ACM International Conference on Multimedia 2025, marking a significant achievement for a Chinese startup in the realm of multimodal generative AI. This recognition underscores their innovative capabilities and remarkable research prowess in transforming how visual content is created. At the heart of this achievement is HiDream-Agent, a groundbreaking multimodal agent that simplifies the complex process of creating visual content through an intuitive conversational interface.
About ACM MM Conference
The ACM MM Conference is renowned for being one of the most prominent academic gatherings focused on multimedia research. Sponsored by the Special Interest Group on Multimedia (SIGMM), it attracts top scholars and industry giants dedicated to fostering innovation and application in multimedia technologies. Awards like the Best Demo signify the international acknowledgment of research excellence, showcasing the abilities of teams pushing boundaries in multimedia technology.
The Innovative HiDream-Agent
HiDream-Agent is celebrated for its ability to overcome the challenges posed by fragmented multimodal tools. It enables seamless integration of features such as text-to-image generation, image editing based on instructions, and converting text or images into videos—all within a single, user-friendly platform. This innovation effectively addresses the pressing need for cross-modal semantic alignment, a common obstacle in the industry. Built on the 17-billion-parameter HiDream-I1 model, it incorporates a sparse Diffusion Transformer structure paired with a dynamic Mixture-of-Experts design, leading to outstanding performance across international benchmarks, including HPS and GenEval.
Transforming Visual Storytelling
The introduction of HiDream-Agent signals a pivotal shift in how accessible and interactive visual storytelling can be. By merging generation with editing into a dialogue-driven experience, it minimizes the barriers to producing high-quality visual content. This innovative approach significantly reduces iteration cycles, enabling creators to enjoy a streamlined creative process that carries ideas from concept to polished output. Currently, this technology has been integrated into HiDream.ai's flagship product, vivago, and is valued for enhancing natural and personalized interactions.
Engaging the Community at ACM MM 2025
During ACM MM 2025, HiDream.ai also organized an exciting Identity-Preserving Video Generation Challenge. This competitive event offered two tracks focused on maintaining facial and full-body identity during video generation. By presenting a new dataset for identity-preserving text-to-video generation, the challenge attracted several top research teams aiming to showcase their capabilities in this specialized field.
About the Visionary Behind HiDream.ai
Founded in 2023 by Dr. Mei Tao, a distinguished Academician within the Canadian Academy of Engineering and a prominent figure in generative AI research, HiDream.ai stands at the forefront of innovation. Dr. Tao's leadership is sustained by a team with over ten years of expertise in exploring and commercializing advanced generative AI technologies. Concentrating on visual multimodal foundation models, the company aims to revolutionize the creative industry through intelligent applications of AI technology. Following the launch of the HiDream-I1 model, which swiftly rose to acclaim within 24 hours, HiDream.ai cemented its position as a leading developer in the global landscape of generative AI.
Approaching the Future
As HiDream.ai moves ahead, its focus remains on deepening innovations in multimodal technology and accelerating industrial applications. Plans are in place to expand into various core application scenarios such as digital creation and film/television post-production. With a commitment to fostering global collaboration and facilitating academic exchanges, HiDream.ai seeks to craft more intelligent and efficient creative solutions for creators across the globe.
Frequently Asked Questions
What is HiDream.ai recognized for?
HiDream.ai is honored for winning the Best Demo at ACM MM 2025 for their innovative multimodal generative AI technology.
What is the core technology behind HiDream-Agent?
HiDream-Agent utilizes the 17-billion-parameter HiDream-I1 model, combining text-to-image generation, image editing, and video generation seamlessly.
How does HiDream.ai contribute to multimedia innovation?
HiDream.ai contributes by developing advanced tools and platforms that enhance visual content creation and storytelling through AI-driven solutions.
What was the Identity-Preserving Video Generation Challenge?
It was a competition aimed at maintaining identity consistency in video generation, attracting numerous top-tier research teams in the field.
Who founded HiDream.ai?
HiDream.ai was founded by Dr. Mei Tao, a notable figure in the field of engineering and AI innovation.
