Profiles2020-2024
Pengcheng He
Microsoft Azure AI
16Papers
0Last author
23Unique co-authors
2First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers16
- Learning Stackable and Skippable LEGO Bricks for Efficient, Reconfigurable, and Variable-Resolution Diffusion ModelingAuthor 5 of 8ICLR 2024 · 17 citations
- DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding SharingAuthor 1 of 4ICLR 2023 · 394 citations
- Guiding Large Language Models via Directional Stimulus PromptingAuthor 3 of 6NeurIPS 2023 · 163 citations
- Diffusion-GAN: Training GANs with DiffusionAuthor 3 of 5ICLR 2023 · 65 citations
- Truncated Diffusion Probabilistic Models and Diffusion-based Adversarial Auto-EncodersAuthor 2 of 4ICLR 2023 · 19 citations
- DIONYSUS: A Pre-trained Model for Low-Resource Dialogue SummarizationAuthor 3 of 6ACL 2023 · 4 citations
- No Parameters Left Behind: Sensitivity Guided Adaptive Learning Rate for Training Large Transformer ModelsAuthor 4 of 8ICLR 2022 · 18 citations
- CAMERO: Consistency Regularized Ensemble of Perturbed Language Models with Weight SharingAuthor 2 of 5ACL 2022 · 6 citations
- Deberta: decoding-Enhanced Bert with Disentangled AttentionAuthor 1 of 5ICLR 2021 · 3,729 citations
- Reader-Guided Passage Reranking for Open-Domain Question AnsweringAuthor 2 of 7ACL Findings 2021 · 28 citations
- Generation-Augmented Retrieval for Open-Domain Question AnsweringAuthor 2 of 9ACL 2021
- Super Tickets in Pre-Trained Language Models: From Model Compression to Improving GeneralizationAuthor 6 of 8ACL 2021
- UnitedQA: A Hybrid Approach for Open Domain Question AnsweringAuthor 4 of 7ACL 2021
- On the Variance of the Adaptive Learning Rate and BeyondAuthor 4 of 8ICLR 2020 · 2,210 citations
- SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized OptimizationAuthor 3 of 7ACL 2020 · 148 citations
- Exploiting Structured Knowledge in Text via Graph-Guided Representation LearningAuthor 3 of 6EMNLP 2020 · 60 citations
