Jacob Andreas

The “self-steering” DisCIPL system directs small models to work together on tasks with constraints, like itinerary planning and budgeting.

MIT-IBM Watson AI Lab researchers have developed a universal guide for estimating how large language models will perform based on smaller models in the same family.

Language models follow changing situations using clever arithmetic, instead of sequential tracking. By controlling when these approaches are used, engineers could improve the systems’ capabilities.

Researchers developed a way to make large language models more adaptable to challenging tasks like strategic planning or process optimization.

Three neurosymbolic methods help language models find better abstractions within natural language, then use those representations to execute complex tasks.