Large language models (LLMs) can generate credible but inaccurate responses, so researchers have developed uncertainty quantification methods to check the reliability of predictions. One popular ...

Nvidia's KV Cache Transform Coding (KVTC) compresses LLM key-value cache by 20x without model changes, cutting GPU memory costs and time-to-first-token by up to 8x for multi-turn AI applications.

In high-stakes settings like medical diagnostics, users often want to know what led a computer vision model to make a certain prediction, so they can determine whether to trust its output. Concept ...

MIT researchers introduce a technique that improves how AI systems explain their predictions, helping users assess trust in ...

To prepare for extreme heat waves around the world -- particularly in places known for cool summers -- climate-simulation models that include a new computing concept may save tens of thousands of ...

Classical computations rely on binary bits, which can be in either of the two states, 0 or 1. In contrast, quantum computing is based on qubits, which can be 0, 1, or a superposition or entanglement ...

FSU College of Engineering and Florida State University’s Resilient Infrastructure and Disaster Response Center examined several types of flood models to highlight their strengths and weaknesses and ...

MIT researchers unveil a new fine-tuning method that lets enterprises consolidate their "model zoos" into a single, continuously learning agent.

Using tumor growth modeling and informed neural networks as early predictive clinical endpoints. 2007 Continuous dispersion for invasive motility. 2009 Invasive growth with cell density and oxygen.

Researchers at the Department of Energy’s Oak Ridge National Laboratory (ORNL) have developed a dynamic modeling method that uses machine learning to provide accurate simulations of grid behavior and ...