AI is all the rage, but is quantum even better? | AI: Promise or Peril

Key Concepts

Quantum Computing: A new paradigm of computing that leverages quantum mechanical phenomena like superposition and entanglement to solve certain problems more efficiently than classical computers.
Classical Computing: Traditional computing based on bits that can represent either 0 or 1.
AI (Artificial Intelligence): A field of computer science focused on creating systems that can perform tasks typically requiring human intelligence, such as learning, problem-solving, and decision-making.
Qubit (Quantum Bit): The quantum equivalent of a classical bit, capable of representing 0, 1, or a superposition of both.
Superposition: A quantum mechanical principle where a quantum system can exist in multiple states simultaneously.
Entanglement: A quantum mechanical phenomenon where two or more quantum particles become linked in such a way that they share the same fate, regardless of the distance separating them.
Quantum Advantage: The point at which quantum computers can solve problems cheaper, faster, or more accurately than any classical means.
Absolute Zero: The theoretical lowest possible temperature, 0 Kelvin or -273.15 degrees Celsius, where all molecular motion ceases.
Portfolio Construction: The process of selecting and combining various financial assets (stocks, bonds, etc.) to create an investment portfolio.
Computational Intensity: The degree to which a computational task requires significant processing power and resources.
Biomedical Research: The scientific study of biological processes and diseases to improve human health.

Quantum Computing vs. AI: A New Era of Computation

This video explores the burgeoning field of quantum computing, its potential to revolutionize problem-solving, and its relationship with the dominant force in technology today: Artificial Intelligence (AI). It questions whether quantum computing will eventually replace AI and delves into the fundamental nature of this advanced technology.

The Dawn of Quantum Advantage

For years, quantum computing was considered a theoretical breakthrough with the potential to tackle humanity's most complex challenges. However, classical computing, particularly AI, has seen rapid advancements, leading to its current dominance. This dynamic shifted with IBM's launch of Nighthawk, their latest quantum processor. IBM claims this processor will usher in the era of "quantum advantage," where quantum computers outperform classical computers in solving specific problems, offering improvements in cost, speed, or accuracy. The video investigates this claim by visiting IBM's quantum lab.

Understanding Quantum Computing

Jerry Chow, an IBM Fellow and Director of Quantum Systems, explains that quantum computing operates on a fundamentally different mathematical and rule-based system compared to classical computing. This difference allows it to handle certain problems with significantly greater efficiency. The core of these quantum systems involves superconducting quantum chips cooled to extremely low temperatures, around 15 millikelvin (0.015°C above absolute zero), which is colder than outer space. This extreme cold is necessary to maintain the delicate quantum states of the qubits.

The Need for Quantum Computing Beyond AI

The necessity of quantum computing arises from the existence of problems that are intractable for classical computers, even advanced AI. These problems often involve exponentially large solution spaces or an overwhelming number of possibilities. Quantum computing offers a new perspective and methodology to approach these challenges. While it may not solve all problems, it provides a novel tool to explore previously insurmountable computational hurdles. The current phase of quantum computing development focuses on identifying its practical applications and leveraging its unique capabilities through an ecosystem of partners developing algorithms and applications.

Problems Quantum Computing Can Solve

Quantum computing is particularly suited for problems characterized by high complexity, a large number of variables, and the need for extreme accuracy and speed. AI excels at processing vast datasets and making predictions, but it struggles with scenarios involving limited data, high accuracy requirements, and an increasing number of variables, which can slow down and reduce the reliability of AI models.

Examples of problems ideal for quantum computing include:

Energy Grid Optimization: Managing complex energy distribution networks with numerous variables.
Logistics: Optimizing intricate supply chains and transportation routes.

These problems involve a rapid accumulation of variables, making them prime candidates for quantum computation.

The Economics of Quantum Computing Access

Access to these powerful quantum processors is not inexpensive. While a free tier offers 10 minutes of usage per month, additional access comes at a premium. A pay-as-you-go plan can cost $96 per minute, potentially reaching $4,800 for an hour, including technical support. However, the actual time needed for an experiment can vary significantly, with some requiring only minutes or even seconds.

Currently, nearly 300 Fortune 500 companies, academic institutions, national laboratories, and startups are utilizing quantum computers, investing in the hope of future returns.

Case Study: Finance and Portfolio Construction

The financial industry, where time and accuracy translate to substantial financial gains, is a key area exploring quantum computing. Portfolio construction, the intricate process of selecting and combining securities for investment, is identified as a highly complex challenge.

Vanguard, a major financial institution, is investing in quantum computing, believing it can revolutionize portfolio construction. Michael Carr, Vanguard's CTO, and Paul Malloy, Head of Municipals, explain their rationale:

Complexity of Securities: Mutual funds and ETFs can comprise thousands of securities.
Exponential Combinations: The number of ways these securities can be combined is astronomically large, making exhaustive computation by classical computers infeasible, potentially taking years, decades, or even millennia.
Quantum's Advantage: Quantum computers are well-suited to handle this exponential component of combinations, enabling the exploration of a vast number of scenarios rapidly.

Vanguard's collaboration with IBM for portfolio construction highlights the learning curve involved. While Vanguard knew the problem, IBM had to develop the quantum algorithms to solve it, as quantum computers operate differently from classical ones. The ability of quantum computers to run numerous simulations very quickly is a significant advantage over existing technologies. Vanguard views their current implementation as an ongoing experiment with promising results, having successfully scaled from a 30-bond portfolio to a 109-bond portfolio, with confidence in further algorithmic improvements.

Quantum Computing in Conjunction with AI: Biomedical Research

The video also explores the synergy between quantum computing and AI, particularly in biomedical research at the Cleveland Clinic, which houses its own IBM Quantum System 1. Dr. Lara Ji, Chief Research Information Officer, discusses their work:

AI's Role in Biomedical Research: AI is typically used for classification and prediction, such as forecasting disease progression or diagnosing conditions earlier. This involves ingesting and analyzing large patient datasets.
Limitations of AI:
- Accuracy: Even with extensive data, AI models may not achieve the required accuracy for clinical deployment.
- Computational Intensity: Deploying computationally intensive AI models for routine clinical care is often impossible due to limitations in processing power.
Quantum's Potential Contribution: Quantum computing can assist by:
- Simplifying Features: Prioritizing the most relevant data features to make problems more manageable.
- Improving Prediction: Analyzing data in novel ways that AI cannot, leading to more accurate predictions.

The Cleveland Clinic is using AI and quantum computing together to solve patient-related problems, aiming to accelerate research and development for faster treatment delivery. Dr. Ji anticipates a future where, for certain problems, researchers might bypass AI and go directly to quantum computing.

The Future of Computing: Abstraction and Integration

Jerry Chow envisions a future where quantum and classical computing operate seamlessly behind the scenes. Users will not need to understand the underlying technology; the system will automatically select the appropriate computing resource (quantum, GPU, or high-performance cluster) based on the problem. The primary impact for the user will be the ability to solve a far greater range of problems and derive more value from computation. Ultimately, society will benefit from this more powerful and capable computing resource.

Conclusion

Quantum computing, once a theoretical concept, is now accessible via the cloud and holds the potential to revolutionize fields like finance, medicine, and transportation. While quantum computer chips are unlikely to be integrated into personal devices like phones or laptops, their impact on how we perceive and solve complex problems could be transformative. The ongoing development and integration of quantum computing with AI promise a future where computational limitations are significantly reduced, unlocking unprecedented possibilities.