Scholar
Haizhong Zheng
Google Scholar ID: Zx6pKsQAAAAJ
Carnegie Mellon University
Citations & Impact
Citations
464
H-index
9
i10-index
9
Publications
17
Co-authors
9
list available
Publications
Resume (English only)
Academic Achievements
- Multiple papers accepted at NeurIPS 2025, including 'Act Only When It Pays', 'Kinetics', and 'Learn To be Efficient' (Spotlight)
- Published 'ELFS: Label-Free Coreset Selection with Proxy Training Dynamics' at ICLR 2025
- Published 'Plato: Plan to Efficiently Decode for Large Language Model Inference' at COLM 2025
- Published 'Harmful Terms and Where to Find Them' at The Web Conference (WWW) 2025 (Oral)
- Published at top venues including ECCV 2024, ICLR 2024, CVPR 2020, and NDSS
- Contributions span efficient LLM inference, data selection, adversarial training, and BEV perception for autonomous driving
Background
- Postdoctoral Researcher in Electrical and Computer Engineering at Carnegie Mellon University
- Member of the InfiniAI Lab, advised by Prof. Beidi Chen
- Also affiliated with the Catalyst Group at CMU
- Research interests: Hardware-aware efficient models, machine learning systems, and data efficiency algorithms
- Research focus: Building scalable and efficient ML models, algorithms, and systems to bridge the gap between rapid model scaling and slower hardware/data scaling
- Two main research thrusts: (1) Designing hardware/system-aware models for fast inference; (2) Developing algorithms for efficient data selection, augmentation, and human feedback
