|
Assel Kembay I am a PhD candidate in the Neuromorphic Computing Group at the University of California, Santa Cruz, advised by Prof. Jason Eshraghian. Before joining UCSC, I received my M.S. degree from the Korea Institute of Science and Technology (KIST). My research focuses on lifelong adaptability of AI systems in dynamic real-world environments: knowledge distillation through space and time, continual learning, and efficient sequence modeling with neuromorphic architectures. |
|
|
Research experience at  Conscium
 UC Santa Cruz
|
News |
| Aug 2026 | Cross-User Adaptation for sEMG Gesture Recognition via Continual Learning accepted at IEEE BioCAS 2026 (Incheon, Korea). |
| Aug 2026 | Awarded a 2026 IEEE Circuits and Systems Society (CASS) Student Travel Grant for APCCAS 2026 (Fukuoka, Japan). |
| Jul 2026 | Officially advanced to PhD Candidacy! |
| Jul 2026 | Three papers accepted: Complex-Pole Dynamics in Transformer Value Paths at APCCAS 2026 (Oral), Sensing with Spikes at IROS 2026 (Oral), and Training SNNs Using Lessons from SSMs at WCCI IJCNN 2026. |
| Feb 2026 | Awarded the UCSC Baskin Engineering Dean's Travel Grant. |
| Jan 2026 | Efficient Knowledge Distillation via Salient Feature Masking accepted at APL Machine Learning. |
| Sep 2025 | Future-Guided Learning, a predictive approach to enhance time-series forecasting, published in Nature Communications. |
| May 2025 | Awarded the 2025 IEEE Women in Engineering (WIE) International Leadership Conference Student Scholarship. |
| Apr 2025 | Selected as a Young Fellow at the Design Automation Conference (DAC) 2025. |
| Jan 2025 | One paper accepted at ISCAS 2025. |
| Oct 2024 | New preprint released: Future-Guided Learning; one paper accepted at BayLearn 2024. |
| Joined the Neuromorphic Computing Group at UC Santa Cruz, advised by Prof. Jason Eshraghian! | |
| Jan 2023 | Won the 2023 POSCO Asia Fellowship (POSCO TJ Park Foundation, South Korea). |
PublicationsI work on brain-inspired machine learning: efficient sequence modeling and neuromorphic architectures, knowledge distillation, and continual learning. Representative papers are highlighted. (*: equal contribution) |
|
Efficient Knowledge Distillation via Salient Feature Masking
Assel Kembay, Skye Gunasekaran, Rui-Jie Zhu, Yu Zhang, Jason K. Eshraghian APL Machine Learning, 2026 paper / code / bibtex @article{kembay2026efficient,
title={Efficient Knowledge Distillation via Salient Feature Masking},
author={Kembay, Assel and Gunasekaran, Skye and Zhu, Rui-Jie and Zhang, Yu and Eshraghian, Jason K.},
journal={APL Machine Learning},
year={2026},
}
Masking distillation to the teacher's salient features makes knowledge transfer more efficient and improves student accuracy. |
|
Future-Guided Learning: A Predictive Approach to Enhance Time-Series Forecasting
Skye Gunasekaran, Assel Kembay, Hugo Ladret, Rui-Jie Zhu, Laurent Perrinet, Omid Kavehei, Jason Eshraghian Nature Communications, 2025 paper / arXiv / code A "future-guided" model that teaches a present-time forecaster with predictive feedback improves time-series forecasting, inspired by predictive coding in the brain. |
|
A Survey on Latent Reasoning
Rui-Jie Zhu, Tianhao Peng, Tianhao Cheng, Xingwei Qu, Jinfa Huang, Dawei Zhu, ..., Assel Kembay, ..., Jason Eshraghian arXiv, 2025 arXiv / code A survey of reasoning performed in models' latent space rather than in explicit chain-of-thought tokens. |
|
Learning with Spike Synchrony in Spiking Neural Networks
Yuchen Tian, Assel Kembay, Nhan Duy Truong, Jason K. Eshraghian, Omid Kavehei arXiv, 2025 arXiv / code A spike-synchrony-dependent plasticity rule that updates weights based on group-level neuron coordination rather than pairwise timing. |
|
A Quantitative Analysis of Catastrophic Forgetting in Quantized Spiking Neural Networks
Assel Kembay*, Karina Aguilar*, Jason Eshraghian ISCAS, 2025 paper / code / bibtex @inproceedings{kembay2025quantitative,
title={A Quantitative Analysis of Catastrophic Forgetting in Quantized Spiking Neural Networks},
author={Kembay, Assel and Aguilar, Karina and Eshraghian, Jason},
booktitle={IEEE International Symposium on Circuits and Systems (ISCAS)},
year={2025},
}
Quantifying how weight quantization interacts with catastrophic forgetting in spiking neural networks under continual learning. |
|
Leveraging Spiking Neural Networks for Solar Energy Prediction in Agriculture
Assel Kembay, Rui-Jie Zhu, Nicholas Kuipers, Jason Eshraghian, Colleen Josephson BayLearn, 2024 paper / code / bibtex @inproceedings{kembay2024agtechsnn,
title={Leveraging Spiking Neural Networks for Solar Energy Prediction in Agriculture},
author={Kembay, Assel and Zhu, Rui-Jie and Kuipers, Nicholas and Eshraghian, Jason and Josephson, Colleen},
booktitle={Bay Area Machine Learning Symposium (BayLearn)},
year={2024},
}
Energy-efficient spiking neural networks for forecasting solar energy availability on farms. |
|
Simulation Web Platform for the Electro-Chemical Oxygen Reduction Reaction
Kim Sch., Lee Ch., Lee B., Seol D., Kim D., Assel Kembay, Yun K., Jang S., Lee J. International Workshop on Computational Nanotechnology (IWCN) (Oral) paper / video |
|
Web Platforms for Conventional Simulations of Matters
Kim Sch., Kim D., Assel Kembay, Kim S., Yun K., et al. Korean Physical Society Spring Meeting (Oral) paper / video |
|
A Simulation Web Platform for Analyzing Electronic Structures of Semiconductors
Kim S., Assel Kembay, Lee J., et al. Korean Physical Society Spring Meeting paper / video |
Research Experience |
|
|
Research Associate, Conscium — London, UK (remote)
May 2026 – Present • Mentor: Dr. Max Ward Generative modeling and optimization of mRNA sequences. |
|
|
Research Scientist Intern, AI Fund — Mountain View, CA, USA
Jan 2026 – Apr 2026 • Mentor: Dr. Andrew Ng AI systems for vision-based reinforcement learning in interactive environments. |
|
|
Graduate Student Researcher, UC Santa Cruz — Santa Cruz, CA, USA
Oct 2023 – Present • Advisor: Prof. Jason Eshraghian Knowledge distillation, continual learning, and efficient language models. |
|
|
Research Scientist Intern, Korea University Medicine — Seoul, South Korea
Mentor: Dr. Il-Joo Cho Wireless brain chip optimization and data transfer algorithms. |
|
|
Research Assistant, AI Research Group, KIST — Seoul, South Korea
Mentor: Dr. Suhyun Kim Data-free knowledge transfer for neuromorphic systems. |
|
|
Research Intern, Computational Science Research Center, KIST — Seoul, South Korea
Mentor: Dr. Seungchul Kim Quantum dot simulation platform development. |
Awards & Milestones |
| 2026 | IEEE CASS Student Travel Grant (Flagship Conferences), for APCCAS 2026, Fukuoka, Japan |
| 2026 | Advanced to Ph.D. Candidacy, UC Santa Cruz |
| 2026 | Graduate Dean's Research Travel Grant, UC Santa Cruz, USA |
| 2025 | DAC 2025 Young Fellow, Design Automation Conference, San Francisco, USA (competitive international selection) |
| 2025 | IEEE WIE Student Scholarship, International Leadership Conference, San Jose, USA |
| 2025 | Graduate Studies DEI Research & Travel Award, UC Santa Cruz, USA |
| 2023 | Divisional MIP Fellowship, UC Santa Cruz, USA (merit-based, ~$20K) |
| 2023 | POSCO Asia Fellowship, South Korea (Next Generation Global Leaders program, full funding) |
Service
Conference reviewer: NeuroAI Workshop @ NeurIPS (2024); IEEE International Symposium on Circuits and Systems, ISCAS (2024–2026); IEEE BioCAS (2026).
|
Miscellaneous1 / 4
@Yntymaq, Tulkibas District, Turkistan Region, Kazakhstan
2 / 4
@Yntymaq, Tulkibas District, Turkistan Region, Kazakhstan
3 / 4
@Santa Cruz, CA, USA
4 / 4
@Santa Cruz, CA, USA
|
|
Design and source code from Jon Barron's website. |