Quanming Yao

Associate Professor@Tsinghua University·中国·OpenReview

研究方向

few-shot learning · neural architecture search · knowledge graph learning · graph neural network · automated machine learning · low-rank · network compression

Quanming Yao

Why Low-Precision Transformer Training Fails: An Analysis on Flash Attention

DGNet: Discrete Green Networks for Data-Efficient Learning of Spatiotemporal PDEs

Searching Meta Reasoning Skeleton to Guide LLM Reasoning

Case-Based Reasoning Enhances the Predictive Power of LLMs in Drug-Drug Interaction

Attending on Multilevel Structure of Proteins enables Accurate Prediction of Cold-Start Drug-Target Interactions

Degree-Conscious Spiking Graph for Cross-Domain Adaptation

Learning Communication between Language Models through Dense Vectors

Coordinated Search for Symbolic Formulas of Complex Network Dynamics

RefP2C: Reflective Paper-to-Code Development Enabled by Fine-Grained Verification

Learning to Learn with Contrastive Meta-Objective

Curriculum-aware Training for Discriminating Molecular Property Prediction Models

Erasing Concept Combination from Text-to-Image Diffusion Model

Why In-Context Learning Models are Good Few-Shot Learners?

Neural Symbolic Regression of Complex Network Dynamics

Hierarchical Graph Tokenization for Molecule-Language Alignment

GUMP: Alleviating Oversquashing with Unitary Message Passing

Adaptive Preference Arithmetic: A Personalized Agent with Adaptive Preference Arithmetic for Dynamic Preference Modeling

UniMoT: Unified Molecule-Text Language Model with Discrete Token Representation

Improving Molecule-Language Alignment with Hierarchical Graph Tokenization

ConML: A Universal Meta-Learning Framework with Task-Level Contrastive Learning

CMIMP: Effortlessly Achieving Diverse Population Training for Zero-Shot Coordination