NVIDIA, the pioneering technology company, has recently introduced a groundbreaking innovation for the gaming industry: ACE for Games. Short for Omniverse Avatar Cloud Engine, ACE for Games is a powerful tool that aims to transform the way non-player characters (NPCs) are created, empowered, and given a voice in video games, all in real-time.
Improving NPC Interactions with ACE for Games
ACE for Games combines cutting-edge technologies, including text-to-speech, natural language understanding (NLU), and facial animation, to revolutionize the NPC experience. During the announcement, NVIDIA’s CEO Jensen Huang highlighted the comprehensive capabilities provided under the ACE umbrella.
With ACE for Games, AI-driven NPCs actively listen to players’ input, enabling immersive interactions. For instance, players can pose questions to NPCs, and the AI generates in-character responses, which are vocalized and accompanied by synchronized facial animations. This seamless integration of advanced AI technologies empowers NPCs to respond dynamically to players’ queries and engage in lifelike conversations.
Real-time Demo: Showcasing the Power of ACE for Games
To demonstrate the potential of ACE for Games, Huang presented a real-time demo crafted in collaboration with Convai, an AI startup, utilizing Unreal Engine 5. The demo’s captivating cyberpunk setting featured a player entering a ramen shop and engaging in conversation with the owner. Notably, the owner had no scripted dialogue but responded instantly to the player’s inquiries, eventually sending them on an impromptu mission.
The impressive demo serves as a glimpse into the future of gaming, showcasing the possibilities that AI brings to the table. As Huang confidently asserted, “AI will be a very big part of the future of video games.” While his optimism may be expected, it is worth noting that NVIDIA stands to benefit greatly from the growing demand for AI, bolstered by the company’s AI accelerator sales.
While we have witnessed some initial integrations of AI models like ChatGPT in games, the ACE platform undoubtedly raises the bar with its polished and truly real-time performance. Huang’s presentation left us with some unanswered questions regarding the technical requirements for running the ACE for Games demo. While he alluded to the inclusion of ray tracing and DLSS, details about the necessary hardware or potential cloud-based components remain elusive. Nevertheless, as games begin to embrace this revolutionary technology, we anticipate more information will surface.
Optimized Neural Networks for Real-time Inferencing
NVIDIA emphasizes that the neural networks powering ACE for Games are optimized for different capabilities, encompassing various size, performance, and quality trade-offs. To assist developers in tailoring models for their games, NVIDIA provides the ACE for Games foundry service, enabling fine-tuning and subsequent deployment through NVIDIA DGX Cloud, GeForce RTX PCs, or on-premises setups, facilitating real-time inferencing.
“The models are optimized for latency—a critical requirement for immersive, responsive interactions in games,” states NVIDIA, underscoring their commitment to delivering exceptional and seamless gameplay experiences.
