Zhiwei Wang

Black holes, fast scrambling and the breakdown of the equivalence principle

2022 · arXiv (Cornell University)

Under reasonable assumptions, black holes have been argued to form firewalls, burning up anything crossing their horizons. This argument finds that a firewall would appear very late in a black hole's lifetime, when Hawking radiation …

Robust One-Shot Segmentation of Brain Tissues via Image-Aligned Style Transformation

2023 · Proceedings of the AAAI Conference on Artificial Intelligence

One-shot segmentation of brain tissues is typically a dual-model iterative learning: a registration model (reg-model) warps a carefully-labeled atlas onto unlabeled images to initialize their pseudo masks for training a segmentation model (seg-model); the seg-model …

Complexity Control Facilitates Reasoning-Based Compositional Generalization in Transformers

2025 · arXiv (Cornell University)

Transformers have demonstrated impressive capabilities across various tasks, yet their performance on compositional problems remains a subject of debate. In this study, we investigate the internal mechanisms underlying Transformers' behavior in compositional tasks. We find …

PanGu-$α$: Large-scale Autoregressive Pretrained Chinese Language Models with Auto-parallel Computation

2021 · arXiv (Cornell University)

Large-scale Pretrained Language Models (PLMs) have become the new paradigm for Natural Language Processing (NLP). PLMs with hundreds of billions parameters such as GPT-3 have demonstrated strong performances on natural language understanding and generation with …