|
Baichuan Huang 黄百川 I am a Ph.D. student in the Department of Electrical and Information Technology, LTH, at Lund University (Sweden), working with Asst. Prof. Amir Aminifar. My research interests are bio-inspired deep learning, efficient training and inference, Internet of Things (IoT), and edge AI. Welcome to our Intelligent Systems Laboratory! Email / Github / Google Scholar / LinkedIn |
|
|
|
|
Baichuan Huang, Amir Aminifar Accepted by AAAI 2025 The 39th Annual AAAI Conference on Artificial Intelligence (Acceptance Rate: 23.4%) code/ poster We propose a memory-efficient forward-only algorithm called TinyFoA, to reduce dynamic memory overhead in the training process. |
|
Baichuan Huang, Amir Aminifar Special Issue on "tinyML – The ecosystem for next generation ML systems", 2025 IEEE Design & Test paper/ code We investigate and compare BP and forward-only algorithms in terms of binarization, finding that PEPITA and FF are more vulnerable to binary activations. |
|
Baichuan Huang (Artifact Evaluation Committee), Amir Aminifar Accepted by MLSys 2025 The Eighth Annual Conference on Machine Learning and Systems (Acceptance Rate: 22%) code We propose a biologically-plausible forward-only algorithm (Bio-FO), not only addressing the biological-implausibility issues associated with BP, but also outperforming the state-of-the-art forward-only algorithms. |
|
Baichuan Huang, Azra Abtahi, Amir Aminifar The inaugural issue of TCASAI, 2024 IEEE Transactions on Circuits and Systems for Artificial Intelligence paper We propose an energy-aware NAS framework for distributed IoT, aiming to search for distributed DNNs to maximize prediction performance subjected to Flash Memory (Flash), Random-access Memory (RAM), and energy constraints. |
|
Amin Aminifar*, Baichuan Huang*, Azra Abtahi, Amir Aminifar (*Equal contribution) Accepted by ECAI 2024 The 27th European Conference on Artificial Intelligence (Acceptance Rate: 23%) paper/ code/ slides/ poster We propose a lightweight inference scheme specifically designed for DNNs trained using the Forward-Forward algorithm ( Contributor to the Forward-Forward repository ... ⭐ ). |
|
Baichuan Huang, Renato Zanetti, Azra Abtahi Fahliani, David Atienza, Amir Aminifar Accepted by AICAS 2023 IEEE 5th International Conference on Artificial Intelligence Circuits and Systems paper/ slides We propose the first interpretable self-supervised network for seizure detection without any need for real seizure data in training, which has comparable performance with supervised methods. (Contemporaneous and similar work of peers in AAAI2023) |
|
Baichuan Huang, Azra Abtahi Fahliani, Amir Aminifar Top-5 Ranking in the Seizure Detection Challenge (Challenge Link) Accepted by ICASSP 2023 IEEE International Conference on Acoustics, Speech, and Signal Processing paper/ slides We propose a lightweight machine-learning framework for real-time epilepsy monitoring on wearable devices (SensorDot of Byteflies). |
|
Baichuan Huang, Hongwei Yi, Can Huang, Yijia He, Jingbin Liu, Xiao Liu Accepted by ICIP 2021 IEEE International Conference on Image Processing paper/ code (... ⭐) / slides/ poster/ This method establishes the state-of-the-arts unsupervised MVS method and demonstrates the powerful generalization ability. |
|
Baichuan Huang, Jun Zhao, Jingbin Liu Journal of Global Positioning Systems , 2021 paper/ project/ cite The paper makes an overview in SLAM including Lidar SLAM, visual SLAM, and their fusion. |
|
Baichuan Huang, Jingbin Liu, Wei Sun, Fan Yang Sensors, 2019 paper To improve the adaptability and robustness of the BLE positioning system, we propose making full use of the three separate channels instead of their combination. |
|
Jingbin Liu, Baichuan Huang, Bin Zhang Geomatics and Information Science of Wuhan University, 2018 paper This paper uses two virtual antennas to realize the music algorithm (Angle of Arrival) based on channel state information (CSI). |
|
|
|
|
|
|
|
|
|
Presentation: : |
|
2021.06: Outstanding Graduates, issued by Wuhan University |
|
I like swimming. |
|
|
|
Last update: Feb 2025. / Thanks for the template. |