A naukri.com initiative
Siliconangle
6d
41
Image Credit: Siliconangle
Multiverse Computing bags $215M for its quantum-inspired AI model compression tech
- Multiverse Computing S.L. secures $215 million in funding to advance its quantum computing-inspired AI model compression technology, aiming to significantly reduce AI inference costs without compromising performance.
- The Series B funding round, led by Bullhound Capital, with participation from various investors including Hewlett Packard Enterprise Co. and Toshiba Corp., signifies immense potential for Multiverse's technology.
- The company's technology can decrease the size of large language models (LLMs) by up to 95% while maintaining performance, offering substantial cost savings for AI inference in production.
- Multiverse's CompatifAI uses quantum-inspired algorithms to compress LLMs, allowing them to run on smaller hardware clusters, thus reducing the need for expensive GPUs and energy consumption.
- The technology can strip out unnecessary parts of AI models and create highly compressed versions that are faster and more efficient, enabling reductions of up to 80% in inference costs.
- Multiverse's CompactifAI models, based on open-source LLMs like Llama and Mistral, can run in various environments, including edge devices like personal computers and smartphones.
- This approach is claimed to be superior to existing model compression techniques like quantization and pruning, offering improved accuracy, performance, and cost-effectiveness in AI training as well.
- Co-founder Román Orús pioneered tensor networks to optimize AI models by eliminating unnecessary correlations, leading to more efficient AI deployment and rapid adoption of Multiverse's technology.
- Enterprises like HPE have embraced Multiverse's quantum-inspired algorithms to enable running AI locally on personal computers, offering enhanced performance, personalization, and cost efficiency.
- The technology has garnered support for its innovative solutions in AI model efficiency, aiming to bring enhanced AI benefits to companies globally by reducing hardware requirements and costs.
Read Full Article
2 Likes
For uninterrupted reading, download the app