AI has risen to the forefront of the business agenda as a result of the global pandemic. According to IDC, the AI market is expected to be worth $300 billion by 2024 in 2020. The demand is forecast to surpass $500 billion in 2024 as of February 2021.
FREMONT, CA: DefinedCrowd has announced the expansion of its online data marketplace, DefinedData, to third-party vendors to sell or exchange AI datasets, as well as a partnership with NVIDIA to provide dataset samples via the NVIDIA NGC catalogue, in response to the rapid increase in demand for high-quality, bias-aware AI training data. Furthermore, the platform also offers AI engineers unparalleled levels of training data accountability, as well as a variety of subscription options, including exclusive discounts for academic institutions.
AI is Top of the Corporate Agenda
AI has risen to the forefront of the business agenda as a result of the global pandemic. According to IDC, the AI market is expected to be worth $300 billion by 2024 in 2020. The demand is forecast to surpass $500 billion in 2024 as of February 2021. Responses to the crisis sped up the adoption of emerging technology by several years, according to a McKinsey study, with 61 percent of high-performance businesses growing their investment in AI. The demand for high-quality datasets has risen dramatically as AI growth has accelerated.
Avoiding AI Bias by Enabling Data Transparency
When more and faster AI systems are introduced, the consequences of inherent bias become more apparent. DefinedData's catalogue also includes comprehensive details on the gender, age, dialect, and phonetic distribution of datasets, as well as meta-data on the recordings and audio samples, to address this issue.
Democratizing Data Access Through NVIDIA NGC
DefinedCrowd will provide dataset samples via the NVIDIA NGC catalogue, a GPU-optimized hub for AI and HPC containers, pre-trained models, as well as SDKs that simplify and accelerate end-to-end workflows, as a key step in democratising access to data. Datasets can be utilized to train models using libraries within the NVIDIA Jarvis application framework; NVIDIA Transfer Learning Toolkit, which empowers developers to build production-quality models faster with no coding required; as well as the NVIDIA NeMo platform, a Python toolkit for training, building, and fine-tuning unmatched GPU-accelerated conversational AI models. This collaboration enables researchers and developers to build high-quality, state-of-the-art conversational AI models.