Can emergent language models faithfully model the intelligence of decision-making agents?
Though modern language models exhibit already some reasoning ability, and theoretically …

The in-context learning (ICL) capability of pre-trained models based on the transformer
architecture has received growing interest in recent years. While theoretical understanding …

Temporal difference (TD) learning with linear function approximation, abbreviated as linear
TD, is a classic and powerful prediction algorithm in reinforcement learning. While it is well …

Pretrained foundation models have exhibited extraordinary in-context learning performance,
allowing zero-shot generalization to new tasks not encountered during pretraining. In the …

Can we establish provable performance guarantees for transformers? Establishing such
theoretical guarantees is a milestone in developing trustworthy generative AI. In this paper …

Modern reinforcement learning (RL) systems often struggle to balance rapid decision-
making with adaptive, reflective learning—a duality that mirrors the interplay of System 1 …