Quantinuum researchers tackle AI’s ‘interpretability problem’, helping us build safer systems
The Artificial Intelligence (AI) systems that have recently permeated our lives have a serious problem: they are built in a way that makes them very hard - and sometimes impossible - to understand or interpret. Luckily, our team is tackling this problem, and we’ve just published a new paper that covers the issue in detail.
It turns out that the lack of explainability in machine learning (ML) models, such as ChatGPT or Claude, comes from the way that the systems are built. Their underlying architecture (a neural network) lacks coherent structure. While neural networks can be trained to effectively solve certain tasks, the way they do it is largely (or, from a practical standpoint, almost wholly) inaccessible. This absence of interpretability in modern ML is increasingly a major concern in sensitive areas where accountability is required, such as in finance and the healthcare and pharmaceutical sectors. The “interpretability problem in AI” is therefore a topic of grave worry for large swathes of the corporate and enterprise sector, regulators, lawmakers, and the general public.
These concerns have given birth to the field of eXplainable AI, or XAI, which attempts to solve the interpretability problem through so-called ‘post-hoc’ techniques (where one takes a trained AI model and aims to give explanations for either its overall behavior or individual outputs). This approach, while still evolving, has its own issues due to the approximate nature and fundamental limitations of post-hoc techniques.
The second approach to the interpretability problem is to employ new ML models that are, by design, inherently interpretable from the start. Such an interpretable AI model comes with explicit structure which is meaningful to us “from the outside”. Realizing this in the tech we use every day means completely redesigning how machines learn - creating a new paradigm in AI. As Sean Tull, one of the authors of the paper, stated: “In the best case, such intrinsically interpretable models would no longer even require XAI methods, serving instead as their own explanation, and one of a deeper kind.”
At Quantinuum, we’re continuing work to develop new paradigms in AI while also working to sharpen theoretical and foundational tools that allow us all to assess the interpretability of a given model. In our recent paper, we present a new theoretical framework for both defining AI models and analyzing their interpretability. With this framework, we show how advantageous it is for an AI model to have explicit and meaningful compositional structure.
The idea of composition is explored in a rigorous way using a mathematical approach called “category theory”, which is a language that describes processes and their composition. The category theory approach to interpretability can be accomplished via a graphical calculus which was also developed in part by Quantinuum scientists, and which is finding use cases in everything from gravity to quantum computing.
A fundamental problem in the field of XAI has been that many terms have not been rigorously defined, making it difficult to study - let alone discuss - interpretability in AI. Our paper presents the first known theoretical framework for assessing the compositional interpretability of AI models. With our team’s work, we now have a precise and mathematically defined definition of interpretability that allows us to have these critical conversations.
After developing the framework, our team used it to analyze the full spectrum of ML approaches. We started with Transformers (the “T” in ChatGPT), which are not interpretable – pointing to a serious issue in some of the world’s most widely used ML tools. This is in contrast with (sparse) linear models and decision trees, which we found are indeed inherently interpretable, as they are usually described.
Our team was also able to make precise how other ML models were what they call 'compositionally interpretable'. These include models already studied by our own scientists including DisCo NLP models, causal models, and conceptual space models.
Many of the models discussed in this paper are classical, but more broadly the use of category theory and string diagrams makes these tools very well suited to analyzing quantum models for machine learning. In addition to helping the broader field accurately assess the interpretability of various ML models, the seminal work in this paper will help us to develop systems that are interpretable by design.
This work is part of our broader AI strategy, which includes using AI to improve quantum computing, using quantum computers to improve AI, and – in this case - using the tools of category theory and compositionality to help us better understand AI.
About Quantinuum
Quantinuum, the world’s largest integrated quantum company, pioneers powerful quantum computers and advanced software solutions. Quantinuum’s technology drives breakthroughs in materials discovery, cybersecurity, and next-gen quantum AI. With over 500 employees, including 370+ scientists and engineers, Quantinuum leads the quantum computing revolution across continents.
Progressing your organization along the five levels of quantum maturity
Quantum computing has moved from a bet on the future to a race already underway. Early adopters are locking in strategic partnerships, building proprietary IP, and positioning themselves years ahead of competitors who are still watching from the sidelines. For executives, the question isn't whether to engage with quantum computing. It's how far along that journey your organization actually is, and what it takes to move forward.
That's a harder question to answer than it sounds. Quantum maturity isn't a single milestone you either hit or miss. It's a progression, built across talent, technology access, workflow integration, partnerships, and value realization, and most organizations aren't entirely sure where they currently stand and what to do next.
Our new paper, A Roadmap for Quantum Maturity, is built to answer exactly that.
A Framework for Where You Stand, and Where to Go Next
Drawing on extensive client experience, the paper lays out five distinct levels of quantum maturity, from early awareness through full transformation, along with the leadership actions that move an organization from one level to the next.
- Awareness — early conversations, but no clear ownership or use cases yet
- Exploration — exploring partnerships and prioritizing use cases with limited budget
- Experimentation — quantum roadmap established and guiding dedicated teams, funding, and partners to execute pilot use case projects
- Integration — quantum computing applications are being embedded into business unit workflows, and the quantum roadmap is integrated with broader digital technology strategy, including AI, HPC, and data
- Transformation — quantum capabilities are embedded in core products and decision-making, with differentiated, proprietary advantage
Most industry leaders today sit at the exploration or experimentation stages, with clear ambitions to reach transformation within the next several years. The paper breaks down what separates organizations that progress from those that stall out at proof-of-concept.
The Advantage Is Built, Not Bought
One of the paper's central takeaways is one many executives underestimate: investing in quantum technology alone isn't enough. Organizations that advance fastest pair that investment with a deliberate strategy, building quantum literacy across leadership and technical teams, honestly assessing capability gaps, focusing on a small number of high-impact use cases tied to real business metrics, and defining a clear roadmap that connects research to business advantage.
How Quantinuum Can Help
Achieving quantum maturity is a journey, not a single step, and most organizations don't need to make that journey alone. Quantinuum's consulting services are built to support every stage of it, from advisory and use-case identification, to capability building, technology access, and the co-development of scalable quantum solutions.
Whether your organization is just starting to build awareness or already scaling toward transformation, our team can help you identify exactly where you stand today, and what it takes to move to the next level.
- SoftBank Corp. and Quantinuum articulated a roadmap connecting quantum algorithms, hardware evolution, and commercial applications, helping organizations understand when quantum computing can address real-world challenges.
- The analysis demonstrates that quantum value will emerge in stages, with early opportunities in data analysis and longer-term breakthroughs in scientific computing as fault-tolerant systems mature.
- The work provides a practical planning framework for enterprises preparing to integrate quantum computing alongside AI and classical high-performance computing infrastructure in quantum-AI data centers.
Quantum computing is increasingly moving from exploratory discussion to structured enterprise planning. As organizations begin to assess where and when quantum technologies may deliver real business impact, a new class of work is emerging: integrated roadmaps that connect algorithmic feasibility, hardware development, and commercial opportunity.
A recent white paper from SoftBank Corp. and Quantinuum represents one of the most comprehensive examples of this approach to date. Rather than treating quantum computing as a distant, abstract capability, the study constructs a detailed, quantitative framework for understanding how real-world use cases evolve as hardware matures—and what this means for enterprise strategy.
A Comprehensive Roadmap
The SoftBank–Quantinuum white paper is broad in scope. It attempts to answer a fundamental question:
Which real-world problems can benefit from quantum computation, at what scale, with what accuracy requirements, and under what hardware conditions?
To address this, the study adopts a structured methodology that connects:
- Algorithmic resource estimates (logical qubits, circuit depth, gate counts)
- Explicit error correction assumptions and overhead models
- Hardware roadmaps across multiple generations
- Timelines for when specific classes of problems may become feasible
Two representative domains anchor the analysis:
- Quantum chemistry, particularly excited-state dynamics and photochemical processes
- Topological data analysis (TDA), including graph-based structure and anomaly detection
These domains were selected because they combine industrial relevance with computational structures that scale poorly on classical systems but map naturally onto quantum approaches.
Quantum chemistry is closely tied to materials science, energy systems, and the development of sustainable technologies. TDA, by contrast, offers tools for understanding complex data structures in networks, finance, and large-scale systems—where identifying structure and anomalies is increasingly critical.
Together, they illustrate how quantum computing may create value across both deep scientific domains and high-impact data applications.
An important caveat is that the resulting roadmap assumes a widely-studied but inefficient error correcting code. As more error correcting codes come online, the resources required to run algorithms will shrink. That means that the timelines detailed in this work can be thought of as “worst case” scenarios, which adds to the value by setting out a clear limit.
A distinguishing feature of the work is its emphasis on implementation over abstraction. Rather than relying solely on theoretical models, the study explicitly constructs quantum circuits and executes them on Quantinuum’s Helios and H2 system.
Parallel Paths to Value
A notable insight from the study is that quantum value creation will not follow a single linear path.
Instead, two complementary regimes are expected to emerge:
- In quantum chemistry, the focus is on fault-tolerant, logical-qubit-based computation, where error correction is essential for scaling toward meaningful scientific and industrial outcomes.
- In TDA, value can emerge earlier, where partial quantum advantage may already support useful computational tasks without full error correction.
This dual-track structure is important: it shows that quantum computing is not a single “threshold technology,” but a spectrum of capabilities that unlock value at different stages of maturity.
The Vision for Quantum Industrialization
The framework ultimately supports a broader strategic vision: the evolution of Quantum AI Data Centers—hybrid infrastructures where quantum processors operate alongside AI and classical HPC systems.
For enterprises, the implication is clear: quantum computing readiness is no longer about speculation. It is about structured preparation, disciplined modeling, and early engagement with the full stack of capabilities that will define the next generation of computational infrastructure.
Every year, The IEEE International Conference on Quantum Computing and Engineering – or IEEE Quantum Week – brings together engineers, scientists, researchers, students, and others to learn about advancements in quantum computing. This year’s conference, from September 13th - 18th in Toronto, Canada, will focus on translating research into real-world impact through the convergence of generative AI, distributed quantum systems, and quantum software engineering.
Throughout IEEE Quantum Week, our quantum experts will be on-site to share insights on upgrades to our hardware, enhancements to our software stack, our path to error correction, and more.
Meet our team at Booth #501 and join the below sessions to discover how Quantinuum is forging the path to fault-tolerant quantum computing with our integrated full-stack.
Keynote with Quantinuum's CEO, Dr. Rajeeb Hazra
September 14th
5:00 – 6:30pm | 800 Hall G
Quantum computing has passed the point where error correction is theoretical. What comes next depends on systems that hold logical performance steady and do real work at scale. On September 14th, join Quantinuum’s CEO Dr. Rajeeb Hazra for his keynote session on “Logically Speaking: The Next Era of Error Correction” where he will explore what the next era of quantum computing requires: shared definitions of logical performance, and benchmarks built on real workloads.
Speaking Sessions
September 13th
11:00 – 11:12am | Location: 601A/B
Workshop: QGenAI: Synergies between Quantum Computing and Generative Artificial Intelligence
Finding Compatible Datasets for Quantum Generative Modeling
Presenting Author: Chen-Yu Liu
September 14th
1:00 – 1:20pm | Location: 701B
Workshop: Q3-Control: Integrated Systems for Quantum Computing, Sensing, and Networking
Cryo-ASICs for Scalable Control
Speaker: Dr. Patty Lee
September 15th
2:30 – 3:30pm | Exhibit Hall (informal event)
Workshop: Q3-Control: Integrated Systems for Quantum Computing, Sensing, and Networking
The Quantum Spectrum
Speaker: Dr. Patty Lee
3:00 – 4:30pm | Location: 801A
Where Quantum-HPC Integration Actually Stands in 2026
Panelist: Neal Erickson
3:00 – 4:30pm | Location: 718A
From Research to Commercialization: Defining the Quantum Workforce for the Next Five Years
Panelist: Kortny Rolston-Duce
September 16th
10:00 – 11:30am | Location: 801B
Who Will Turn Quantum Computing into Value? Specialists, Domain Experts, and the Workforce Gap
Panelist: Enrico Rinaldi
September 17th
10:00 – 11:30am | Location: 701B
International Workshop on Quantum Computing for Power Systems: From Optimization Algorithms to Grid-Scale Applications | Session 1
Organizer: Kortny Rolston-Duce
10:00 – 11:30am | Location: 714A
openQSE: Co-Designing the Quantum-HPC Software Stack from Applications to Control Systems | Session 1
Speaker: Neal Erickson
10:00 – 11:30am
The Impact of Qubit Connectivity on Quantum Advantage in Noisy IQP Circuits
Presenting Author: Leonardo Placidi
10:15 – 11:00 am | Location: 601A/B
Workshop: AI for Circuit Synthesis, Optimization, and Discovery
Automated near-term quantum algorithm discovery
Speaker: Konstantinos Meichanetzidis
11:00 – 11:15am | Location: 601A/B
Workshop: AI for Circuit Synthesis, Optimization, and Discovery
Fast Stabilizer State Preparation via AI-Optimized Graph Decimation
Presenting Author: Jasmine Brewer
1:00 – 1:15pm | Location: 601A/B
Workshop: AI for Circuit Synthesis, Optimization, and Discovery
Reinforcement Learning for Adaptive Composition of Quantum Circuit Optimisation Passes
Speaker: Gabriel Matos
1:00 – 1:20am | 718B
Quantum Software 2.6: current challenges and headways in quantum software | Session 1
Quantum compilation and hybrid compilation
Speaker: Ross Duncan
1:00 – 2:30pm | Location: 714A
openQSE: Co-Designing the Quantum-HPC Software Stack from Applications to Control Systems | Session 2
Speaker: Phillipp Seitz
1:45 – 2:00pm | Location: 601A/B
Workshop: AI for Circuit Synthesis, Optimization, and Discovery
Graph-Theoretic Quantum Circuit Optimization with the ZX-Calculus and Gumbel AlphaZero
Speaker: Alexander Koziell-Pipe
2:00 – 2:15pm | Location: 601A/B
Workshop: AI for Circuit Synthesis, Optimization, and Discovery
Reusable Equivariant Neural Compilers for Matrix-Group Quantum Circuit Synthesis
Speaker: Richie Yeung
3:30 - 4:00pm | Location: 715B
Workshop: 4th Workshop on Quantum Computing for Natural Sciences: Technology and Applications
Quantum Algorithms
Speaker: Setso Metodi
September 18th
10:00 – 11:30am | Location: 801B
Scalable QPU architectures
Panelist: Setso Metodi