# 量子计算：摆脱数据桎梏，创新药研发充满想象力的未来 | 李 翛然 | TEDxNanjing Source: https://www.youtube.com/watch?v=m8nH5HG5MhI Recap page: https://rapidrecap.app/video/m8nH5HG5MhI Generated: 2025-12-16T21:03:15.679+00:00 --- ## Quick Overview The speaker details how quantum computing combined with AI (Quantum AIDD) is revolutionizing drug discovery by drastically accelerating the process of finding the right molecular **Key Points:** - The core challenge in drug discovery is finding the correct molecule to target a specific biological "lock" (target protein), which involves searching through an enormous chemical space (estimated at 1.11 x 10^21 possibilities). - Traditional methods for finding the correct drug candidate are extremely slow and expensive, taking 10-15 years and costing over $2 billion, with a 95% failure rate. - The speaker highlights the success of Imatinib (Gleevec) for CML, which took 41 years from discovery of the target ("lock") to market approval in 2001. - Quantum computing, specifically using 105-qubit processors like the 2024 'Willow' chip, can solve problems like Random Circuit Sampling in 5 minutes that would take a classical supercomputer 10^25 years. - The presentation introduces Quantum AIDD (Quantum + AI Drug Design) as a method to overcome the limitations of traditional computational methods, which struggle with the triple interaction of Drug-Target-Body. - The speaker emphasizes that nature is fundamentally quantum mechanical, necessitating quantum computation for accurate simulation and design of molecules. - The ultimate goal is to use Quantum AIDD to design and validate previously undiscovered drug structures, as demonstrated by calculating the activation energy barrier for a carbon-carbon bond cleavage reaction. ![Screenshot at 00:07: The speaker stands on stage in front of a large screen displaying the title slide for TEDxNanjing 2025, which reads "Quantum Computing: Breaking the Shackles of Data, Creating a Future Full of Imagination for New Drug R&D," setting the stage for a discussion on accelerating pharmaceutical innovation.](https://ss.rapidrecap.app/screens/m8nH5HG5MhI/00-00-07.png) **Context:** This TEDxNanjing 2025 talk by Li Xiaoran focuses on the intersection of quantum computing, artificial intelligence (AI), and drug discovery, specifically addressing the immense computational challenges in identifying effective drug molecules. The presentation contrasts the decades-long timeline of traditional drug development, exemplified by the 41-year journey of Imatinib, with the potential speed and accuracy offered by next-generation computational tools like quantum computers. ## Detailed Analysis The presentation argues that the search for new drugs is fundamentally a search problem within an impossibly large chemical space, specifically highlighting the challenge of finding the right molecule to fit a biological target (the "lock"). The speaker quantifies this search space at $1.11 \times 10^{21}$ possible molecules, noting that current drug discovery efforts, even with AI, often narrow this down to about 5,000 possibilities, with the final selection (the "last mile") remaining difficult due to incomplete data and model accuracy. The speaker then contrasts this with the traditional drug development timeline, citing Imatinib (Gleevec) for CML, which took 41 years from target discovery (1983) to market (2001). The presentation pivots to the power of quantum computation, noting that a 105-qubit quantum chip achieved a calculation in 5 minutes that would take a classical computer $10^{25}$ years. The core solution proposed is Quantum AIDD (Quantum + AI Drug Design), a hybrid pipeline designed to accurately simulate nature, which the speaker asserts is fundamentally quantum mechanical. This approach aims to solve the triple interaction problem between the drug, the target, and the body, which classical methods cannot precisely analyze. A key result shown involves using quantum computation to design and validate a previously undiscovered drug structure by accurately calculating the activation energy barrier for a chemical reaction, demonstrating the practical application of these advanced tools in rational drug design. ### Drug Discovery Challenge - Identifying the Lock and the Search Space: The problem involves finding the one correct cause among $1.11 \times 10^{21}$ possibilities (molecules) for a disease target ("lock"). - Traditional drug research takes 10-15 years, costs over $200 million, and has a 95% failure rate, primarily due to lack of efficacy (40-50%), poor pharmacokinetics (30%), and toxicity (10-15%). ### Quantum Computing Advantage - Speed and Scale: The 2024 'Willow' 105-qubit quantum chip completed a task in 5 minutes that a classical computer would need $10^{25}$ years to solve. - Quantum computers can perform tasks like "Guessing" or "Opening a Box" simultaneously, unlike classical computers. ### Quantum AIDD Framework - The proposed solution involves a hybrid approach where quantum algorithms (like VQE) work with classical optimizers to refine molecular structures, moving from a massive search space to a single, confirmed result. ### Advantage 1 - Simulating the Microscopic World: Nature isn't classical; simulating it accurately requires quantum mechanics. Quantum algorithms provide a more 'native' simulation method. ### Advantage 2 - Industry-Specific Tailoring: Quantum AIDD is presented as an industry-driving engine for the current Noisy Intermediate-Scale Quantum (NISQ) era. ### Validation Example - Designing Novel Structures: Quantum calculation successfully designed and validated a drug structure involving a C-C bridge cleavage activation, providing results comparable to or better than classical methods (CASCI vs VQE/DFT comparison table). ![Screenshot at 00:07: Introduction slide for TEDxNanjing 2025 emphasizing the theme of quantum computing for drug discovery.](https://ss.rapidrecap.app/screens/m8nH5HG5MhI/00-00-07.png) ![Screenshot at 00:33: Slide illustrating the 41-year timeline from discovering the BCR-ABL fusion protein \(the "lock"\) to the market launch of Imatinib \(the "key"\).](https://ss.rapidrecap.app/screens/m8nH5HG5MhI/00-00-33.png) ![Screenshot at 01:57: Slide quantifying the drug discovery challenge, showing $1.11 \\times 10^{21}$ possibilities requiring search, and illustrating the funnel of drug development success rates.](https://ss.rapidrecap.app/screens/m8nH5HG5MhI/00-01-57.png) ![Screenshot at 04:40: Slide highlighting the limitations of current methods in precisely analyzing the triple interaction between drug, target, and body.](https://ss.rapidrecap.app/screens/m8nH5HG5MhI/00-04-40.png) ![Screenshot at 11:15: Slide detailing the "Hybrid Quantum Computing Pipeline for Real World Drug Discovery," showing the iterative optimization loop involving quantum and classical computation.](https://ss.rapidrecap.app/screens/m8nH5HG5MhI/00-11-15.png)