Cerence Revolutionizes In-Car Experiences with Automotive-Specific LLM in Collaboration with NVIDIA

TL;DR:

  • Cerence and NVIDIA collaborate to introduce CaLLM™, an automotive-specific large language model.
  • CaLLM™ forms the foundation for Cerence’s advanced in-car computing platform, integrated with NVIDIA DRIVE.
  • The partnership aims to expedite time-to-market for new user experiences, bridging embedded systems and AI-powered domains.
  • CaLLM™ offers automotive-specific intelligence, user personalization, and built-in information retrieval.
  • Cerence utilizes NVIDIA’s AI foundry service, including DGX Cloud, for training and fine-tuning CaLLM™.
  • NVIDIA AI Enterprise and NVIDIA-accelerated computing further enhance CaLLM™’s capabilities.
  • This innovation promises to revolutionize in-car interactions, providing a seamless and intuitive conversational interface.
  • It has the potential to benefit both the automotive industry and other sectors.
  • CaLLM™ and Cerence’s new platform are being tested with leading automakers and will be showcased at CES 2024.

Main AI News:

Cerence Inc., a trailblazer in AI for the automotive industry, is set to redefine the future of in-car experiences through its groundbreaking automotive-specific large language model, CaLLM™, powered by NVIDIA technology. CaLLM™, or Cerence Automotive Large Language Model, stands as the cornerstone of Cerence’s next-generation in-car computing platform, seamlessly integrated with the powerful NVIDIA DRIVE platform. In close partnership with NVIDIA, Cerence is tackling pivotal challenges faced by automakers, including expediting time-to-market by enabling the deployment of new user experiences via cloud integration, seamlessly bridging existing embedded systems and innovative generative AI-powered domains. This strategic collaboration empowers OEMs to stay ahead of the curve as drivers increasingly seek intelligent features and capabilities within their vehicles.

In an era where automakers and mobility OEMs are harnessing the potential of generative AI and large language models to enhance the user experience, CaLLM™ rises to the occasion, fulfilling critical requirements. It boasts an unparalleled level of automotive-specific intelligence, leveraging Cerence’s extensive expertise in the automotive domain and a meticulously curated automotive dataset containing billions of tokens. The result? Integrated in-car user experiences that transcend the capabilities of generic knowledge large language models. CaLLM™ is uniquely equipped to support a spectrum of automotive functions, features, and requirements, all while offering deep customization options for automakers through training, fine-tuning, and tailored applications. Its feature set includes user personalization and built-in information retrieval to power Cerence’s generative AI-driven applications like Cerence Car Knowledge.

Cerence is forging ahead with the development of CaLLM™ through the NVIDIA AI foundry service, an ecosystem comprising NVIDIA AI Foundation Models, NVIDIA AI Enterprise software, and NVIDIA accelerated computing. The process unfolds in two critical stages. First, Cerence will train CaLLM™ on its extensive dataset, leveraging the capabilities of NVIDIA DGX Cloud and NVIDIA DGX systems. Subsequently, Cerence will deploy CaLLM™ within an NVIDIA-accelerated infrastructure, seamlessly integrating it with NVIDIA AI Enterprise. This end-to-end, cloud-native, secure software platform accelerates training, fine-tuning, and inferencing of large language models, thanks to the prowess of the NVIDIA NeMo framework and the NVIDIA TensorRT-LLM optimization library.

Iqbal Arshad, Chief Technology Officer at Cerence, emphasized, “We are delivering a groundbreaking in-car computing platform that harnesses our vertical tech stack, enabling users to accomplish tasks across applications through a unified conversational interface. When combined with NVIDIA’s industry-leading AI frameworks and platforms, our exclusive dataset, robust automaker partnerships, and extensive market presence position us as leaders with our pioneering automotive-specific LLM, CaLLM™, and our proprietary in-car computing platform. Looking ahead, we envision this unique architecture benefiting not only the automotive industry but also industry leaders beyond.”

Ali Kani, Vice President of Automotive at NVIDIA, concurred, stating, “Generative AI and automotive LLMs are poised to usher in new, intuitive, and personalized user experiences, capabilities, and features, both within and beyond the vehicle. Leveraging NVIDIA’s core expertise in cloud and edge technologies, Cerence can expedite the training, scaling, and deployment of these models within their in-car user-experience platform, resulting in safer, smarter, and more enjoyable rides.”

CaLLM™ assumes a pivotal role in shaping the cars of the future as it serves as the cornerstone of Cerence’s forthcoming in-car computing platform. This platform heralds a new era of in-car interaction, offering an automotive- and mobility-specific assistant that delivers an integrated in-cabin experience. Unlike previous solutions that necessitated multiple, disjointed interactions, burdening drivers with a high cognitive load, Cerence’s innovative platform seamlessly consolidates all facets of a user’s interaction into a fluid and intuitive conversational interface. Cerence’s cutting-edge platform will facilitate a seamless transition for OEMs by reimagining their existing applications within the conversational interface while introducing new generative AI-powered services.

The future of in-car experiences is here, and it’s a testament to Cerence’s commitment to innovation and collaboration. CaLLM™ and Cerence’s groundbreaking platform are currently in proof-of-concept with leading global automakers and will be showcased at CES 2024 in Cerence’s booth, number 6627, located in the West Hall. Stay tuned for a firsthand look at the automotive revolution in action.

Conclusion:

The collaboration between Cerence and NVIDIA, resulting in the introduction of CaLLM™ and an advanced in-car computing platform, signifies a significant leap in the automotive industry. With the promise of enhanced user experiences and seamless integration of generative AI, this innovation has the potential to set new standards for in-car interactions. It not only benefits automakers but also opens doors for broader applications across industries, emphasizing the growing importance of AI-driven solutions in the market.

Source