About This Research Theme
Computers are often perceived as highly logical entities that process inputs, execute specified operations, and generate outputs. Yet, it's crucial to remember that they are still physical systems governed by the laws of physics. This might sound trivial, but it carries a significant implication: computers are subject to the statistical behaviors of particles, and hence, susceptible to "noise." The quest for building a reliable computing machine by assembling physical components has been a pivotal area of research in computer science since the field's inception.
At our lab, we focus on two critical reliability challenges in current and upcoming computing technologies: the scale and the quantum noise. In classical computing, machine learning and AI algorithms are demanding unprecedented amounts of computation, far-exceeding the pace of semiconductor technology innovations. To handle large-scale AI algorithms, our only option is to parallelize the computation over multiple computing units (e.g., thousands of GPUs). However, this approach introduces a higher number of potential failure points and increases the likelihood of system failures. Our focus is on overcoming these reliability issues in large-scale distributed and parallel computing environments. We achieve this at the software level, utilizing advanced information-theoretic methodologies.
Quantum computing is an alternative avenue to overcome the limitations inherent in classical computing technologies. Despite the relentless efforts of researchers to construct a stable quantum computer, the fragility of qubits remains a significant barrier to developing a practical quantum computer with an adequate number of qubits. Leveraging our expertise and methodologies honed in the realm of reliable classical computing, we are pioneering innovative solutions aimed at developing fault-tolerant quantum computing algorithms.
👩🏻🔬 Projects in this Theme
| Project Name | People | Last edited | Themes |
|---|---|---|---|
Q Quang Iain Weissburg | Sep 27, 2024 8:12 PM | ||
Sep 27, 2024 8:12 PM | |||
J Jesse Lee | Sep 27, 2024 8:12 PM |
📚 Selected Publications
- Nguyen, Quang Minh, Iain Weissburg, and Haewon Jeong. "Coded Computing for Fault-Tolerant Parallel QR Decomposition." arXiv preprint arXiv:2311.11943 (2023).
- Cadambe, Viveck R., Haewon Jeong, and Flavio P. Calmon. "Differentially Private Secure Multiplication: Hiding Information in the Rubble of Noise." In 2023 IEEE International Symposium on Information Theory (ISIT), pp. 2207-2212. IEEE (2023).
- Jeong, Haewon, Ateet Devulapalli, Viveck R. Cadambe, and Flavio P. Calmon. "ϵ-approximate coded matrix multiplication is nearly twice as efficient as exact multiplication." IEEE Journal on Selected Areas in Information Theory 2, no. 3 (2021).
- Dutta, Sanghamitra, Mohammad Fahim, Farzin Haddadpour, Haewon Jeong, Viveck Cadambe, and Pulkit Grover. "On the optimal recovery threshold of coded matrix multiplication." IEEE Transactions on Information Theory 66, no. 1 (2019).