Xingyuan Yuan
3 papers · 2018–2025 · 2 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓
๐
Conference Polyglot ยท Bronze
๐
Academic Marathon ยท Sprint
๐
Interdisciplinary Bridge ยท Bronze
๐
Conference Pioneer
Top co-authors
Research topics
Keywords
pre-training perplexity
(2)
reinforcement learning
(1)
policy gradient
(1)
program synthesis
(1)
transformer block
(1)
multi-head attention
(1)
large language model
(1)
neural programming
(1)
combinator abstraction
(1)
dcformer
(1)
dcmha
(1)
dynamically composable multi-head attention
(1)
cross-layer information flow
(1)
dynamic dense connections
(1)
transformer architecture
(1)
mudd connections
(1)
Papers
MUDDFormer: Breaking Residual Bottlenecks in Transformers via Multiway Dynamic Dense Connections
ICML 2025
Improving Transformers with Dynamically Composable Multi-Head Attention
ICML 2024
Improving the Universality and Learnability of Neural Programmer-Interpreters with Combinator Abstraction
ICLR 2018