Compared totheexisting sampling-based approaches, which are inapplicable to the infinite time horizon setting, wetrain aseparate deterministic neural networkthatservesasaninfinite timehorizon safety certificate.

Results ofexperiments showthatRMT performs on par with the Transformer-XL on language modeling for smaller memory sizes and outperforms it for tasks that require longer sequence processing. We show that adding memory tokens to Tr-XL is able to improve its performance.

In Section 4, we argue that there exists a discrepancy between over-smoothing based theoretical results and the practical capabilities of deep GCN models, demonstrating that over-smoothing is not the key factor that leads to the performance degradation in deeper GCNs.

In particular, we apply this approach to build CUBE (CUltural BEnchmark forText-to-Image models), afirst-of-its-kind benchmark toevaluate cultural competence of T2I models.2 CUBE covers cultural artifacts associated with 8 countries across different geo-cultural regions and along 3 concepts: cuisine, landmarks, and art. CUBE consists of 1) CUBE-1K, a set of high-quality prompts thatenable theevaluation ofcultural awareness, and2)CUBE-CSpace, a larger dataset of cultural artifacts that serves as grounding to evaluate cultural diversity.