AI & ML Research 24 Hours

Last updated: July 24, 2026, 8:30 PM ET

AI & ML Research

LLM Advancements and Applications

Researchers are developing new methods for Large Language Models (LLMs) to process and understand data beyond traditional text. Tabular foundation models can now predict missing spreadsheet columns zero-shot, outperforming gradient-boosted trees on the Tab Arena benchmark. For document processing, an Intelligent Document Processing (IDP) system can be built and run in the cloud to automate the classification and extraction of PII from emails using AWS. Additionally, a loop engineering approach for RAG generation utilizes an LLM cascade, moving from local models to a hosted flagship, incorporating cost optimization and a validation loop.

Enhancing AI Memory and Data Preservation

Innovations are addressing limitations in AI memory and data retention. One approach to improve AI agent memory involves a usage-reinforced decay engine that prioritizes important information over simply the newest data, inspired by the Ebbinghaus forgetting curve. In a different domain, significant progress is being made in organ preservation, with supercooled kidneys being transplanted into pigs in a laboratory setting, aiming to address the critical shortage of donor organs by extending viability outside the body transplant organs. This research is part of a broader effort to keep organs alive outside the body for longer periods.