Search results for Kumo Tabular
AI & Enterprise
From TFM to CEM, AI models different from LLMs are arriving
AI companies are increasingly betting on models that operate differently from large language models, including tabular foundation models for structured data prediction, decision-focused models that output choices and probabilities, and code execution models that optimise running programs autonomously. Gartner flagged large quantitative models as a key 2027 keyword. SAP and Nvidia have moved into TFM. Decision model competition has expanded to major firms. Nvidia-backed DoubleAI is also drawing attention with CEM technology.
AI & Enterprise
K-AI strategy under scrutiny as battle heats up over decision AI models
Government AI policy is expected to be a key issue during this year’s state audit running from Oct. 6 to Oct. 30, with attention on the independently developed AI foundation model project, Dokpamo, and the details of a separate frontier AI development initiative. The decision AI model market has quickly shifted into a multi-player competition. The article also rounds up moves by OpenAI, Google, Meta, Microsoft, Nvidia, IBM and others, along with investments and regulatory concerns.
AI & Enterprise
Nvidia unveils Kumo Tabular open-source foundation model for tabular data prediction
Nvidia has released Kumo Tabular, an open foundation model for predicting tabular data, on the open-source AI model-sharing platform Hugging Face. The model reads tables with answers and predicts values for rows with blank entries without additional training, tuning or feature engineering, supporting both classification and regression. Nvidia said three versions are offered, and that it ranked first across four benchmarks.