Cici Liu

BA Dartmouth College in Anthropology | Master's in Accounting

Exploring finance and AI through lenses of sociology, anthropology and accounting.

Listen to Podcast Get in Touch

Featured Podcast

🎙️

Financial News & Analysis Podcast

Diving deep into global trade, AI, and cultural factors shaping financial markets. Weekly episodes exploring the intersection of these critical topics.

Listen Now → 🔒 Become a Premium Listener →

Articles

Quantum Computing & AI

A deep-dive into current AI technology and its impacts on computing, optimization, and real-world applications across industries.

Read Article →

Quantum Computing & AI

Published May 26, 2026 | A deep dive into Chris Bernhardt's book "Quantum Computing"

Right now, classical AI is hitting a wall of scale. Training massive models with billions of parameters requires an immense amount of energy, millions of data processors (GPUs), and massive time loops for adjusting mathematical weights.

Quantum computing doesn't just make this classical process "faster"; it fundamentally changes how information is represented and optimized. By swapping binary bits for qubits, quantum computing enhances AI models through three core mechanisms:

1. High-Dimensional Vector Spaces (Superposition)

In modern deep learning, AI understands data by mapping concepts into high-dimensional space. For instance, Large Language Models represent the semantic meanings of words as long lists of numbers called vectors. To process these, a classical computer must compute massive matrix multiplications line by line.

The Quantum Upgrade: Because a qubit can exist in a superposition (representing multiple states simultaneously), a quantum computer can map and hold highly complex, high-dimensional data landscapes natively (Zeguendry et al., 2023).

The Benefit: Instead of checking data parameters one at a time, a quantum processor can evaluate these massive vector spaces simultaneously. This allows AI models to process highly unstructured or complex datasets using significantly fewer steps.

2. Native Correlation Tracking (Entanglement)

When training deep neural networks (like the transformer architectures behind modern AI), the hardest part is tracking the relationships between distant data points—such as how a word at the beginning of a paragraph alters the meaning of a word at the end. Classical systems use complex, memory-heavy attention mechanisms to map these relationships.

The Quantum Upgrade: Quantum computing uses entanglement, where the mathematical states of multiple qubits are locked together into a single, unified system (Shi et al., 2024).

The Benefit: Quantum Neural Networks (QNNs) can use entangled qubits to natively mimic the deeply interconnected layers of a neural network, capturing deep correlations that classical models struggle to efficiently process (Abbas et al., 2021). This allows the model to register non-linear relationships and hidden structural patterns across a dataset without requiring massive classical pointer arrays or explosive parameter counts.

3. Elegant Optimization (Interference)

Training an AI model is essentially a massive game of hot-and-cold. The model looks at its errors and uses a process called gradient descent to slowly twist and adjust millions of mathematical weights until it finds the configuration with the lowest possible error rate. This brute-force optimization is what keeps modern data centers running at maximum capacity.

The Quantum Upgrade: Quantum algorithms use constructive and destructive interference. By applying quantum gates (which act as geometric rotations to the qubit states), an algorithm can manipulate probabilities.

The Benefit: It forces the mathematical paths representing the wrong answers to cancel each other out (like noise-canceling headphones) while causing the path to the correct or optimal answer to amplify. This turns AI optimization from a slow, step-by-step climb down an error hill into an instantaneous, physics-driven calculation.

The Core Synergy: Why AI Needs Quantum

Classical computers process information in bits (0s and 1s). AI models, especially Deep Learning networks, require massive matrix multiplications and optimizations to find patterns in data. As AI models grow to trillions of parameters, classical hardware (GPUs and TPUs) hits physical and energetic limits.

Quantum computers use qubits, which leverage two unique principles of quantum mechanics:

The result: Quantum computers can navigate massive, complex data spaces in seconds—tasks that would take a classical supercomputer thousands of years.

How Quantum Enhances AI

Quantum computing doesn't just make AI "faster"; it fundamentally changes the algorithms AI can use.

A. Accelerated Training of Deep Learning Models

Training a large language model or a complex neural network requires massive computational power and time. Quantum algorithms, such as the Quantum Approximate Optimization Algorithm (QAOA), can optimize the weights and biases of a neural network by navigating the optimization landscape significantly faster than classical gradient descent (Farhi et al., 2014).

B. Handling High-Dimensional Data

True world data is complex and multi-layered. Classical computers struggle with "the curse of dimensionality" when processing thousands of variables simultaneously. Quantum computers naturally excel at mapping data into high-dimensional "quantum feature spaces," making it easier for AI to find hidden patterns in messy data like genomic sequences or global financial markets (Ordóñez et al., 2025).

C. True Quantum Generative AI

Current generative AI (like image and text generators) uses probabilistic models to "guess" the next pixel or word. Quantum systems can handle complex probability distributions far more naturally, potentially leading to AI that generates highly accurate molecular structures, advanced materials, or flawless synthetic data.

Real-World Applications of Quantum AI

Industry Current AI Limitation The Quantum AI Solution
Pharmaceuticals Simulating molecular interactions takes years of trial and error. QAI can simulate molecular physics precisely, reducing drug discovery from a decade to days.
Finance & Logistics Portfolio optimization and supply chain routing involve too many variables. Quantum algorithms can calculate the optimal route or risk profile near-instantly.
Cybersecurity AI-driven threat detection can be bypassed by complex, novel attacks. Quantum-encrypted AI networks will offer virtually unhackable data protection.
Climate Modeling Simulating global weather systems requires omitting many variables. QAI can model environmental systems at a molecular and global scale simultaneously.

References

Abbas, A., Sutter, D., Zoufal, C., Lucchi, A., Figalli, A., & Woerner, S. (2021). The power of quantum neural networks. Nature Computational Science, 1, 403–409. https://doi.org/10.1038/s43588-021-00084-1
Farhi, E., Goldstone, J., & Gutmann, S. (2014). A Quantum Approximate Optimization Algorithm. arXiv. https://doi.org/10.48550/arxiv.1411.4028
Ordóñez, S. A. C., Torres, L. F., Bifulco, M., Durán, C. A., Bosch, C., & Carbajo, R. S. (2025). Embedding-Aware Quantum-Classical SVMs for Scalable Quantum Machine Learning. arXiv. https://doi.org/10.48550/arxiv.2508.00024
Shi, J., Xiao, Z., Shi, H., Jiang, Y., & Li, X. (2024). QuanTest: Entanglement-Guided Testing of Quantum Neural Network Systems. ACM Transactions on Software Engineering and Methodology, 2024. https://doi.org/10.1145/3688840
Zeguendry, A., Jarir, Z., & Quafafou, M. (2023). Quantum Machine Learning: A Review and Case Studies. Entropy, 25, 287. https://doi.org/10.3390/e25020287
← Back to Articles

About Me

Background

I hold a BA from Dartmouth College in Anthropology and a Master's degree in Accounting. This unique combination has shaped my approach to understanding global markets and financial systems through both cultural and quantitative lenses.

Areas of Focus

Finance & Markets
AI & Technology
Accounting
Cultural Analysis
Society & Trends
Market Insights

Let's Connect

Interested in collaboration, speaking engagements, or just want to chat about finance and AI?

cici.liu19@gmail.com