Scholar
Sangmin Bae
Google Scholar ID: T5rHY14AAAAJ
Postdoc at KAIST AI || PhD at KAIST AI
Citations & Impact
Citations
695
H-index
14
i10-index
16
Publications
20
Co-authors
37
list available
Publications
Resume (English only)
Academic Achievements
- Publications: 'Mixture-of-Recursions' accepted at NeurIPS 2025; 'Contrastive Decoding' accepted at TMLR 2025; 'Accelerating Large Language Model Inference via Early-Exiting Algorithms' PhD Dissertation; and other conference and journal papers. Awards: $30,000 Google Grant Project.
Research Experience
- Postdoctoral Researcher, Graduate School of AI, KAIST (Oct. 2025 - Present)
Education
- Postdoc: Graduate School of AI, KAIST, advised by Se-Young Yun (Oct. 2025 - Present); Ph.D.: Graduate School of AI, KAIST, advised by Se-Young Yun (Mar. 2021 - Aug. 2025); M.S.: Industrial and Systems Engineering, KAIST, advised by Se-Young Yun (Mar. 2019 - Feb. 2021); B.S.: Industrial and Systems Engineering, KAIST (Mar. 2014 - Feb. 2019)
Background
- Research Interests: Efficiency and Multimodality; developing Scalable and Efficient Foundation Language Models; proposed novel Adaptive Computation methodologies that significantly boost Inference-Efficiency; focuses on enhancing Training- and Data-Efficiency across various modalities including Vision, Audio, and Tabular data.
Miscellany
- Personal interests not provided
