Japan Tobacco and D-Wave Announce Quantum Proof-of-Concept Outperforms Classical Results for LLM Training in Drug Discovery
D-Wave Quantum (NYSE: QBTS) and Japan Tobacco's pharmaceutical division have successfully completed a groundbreaking proof-of-concept project combining quantum computing with AI for drug discovery. The project demonstrated that Large Language Models (LLMs) enhanced with D-Wave's quantum-hybrid workflow outperformed classical computing methods in generating novel drug-like molecular structures.
Key findings show that the quantum-hybrid AI system produced more valid molecules with higher drug-likeness compared to traditional approaches. The collaboration utilized D-Wave's annealing quantum computing technology within JT's AI framework to train transformer architecture models, similar to ChatGPT, for exploring chemical space.
The successful proof-of-concept validates quantum computing's potential to enhance AI capabilities while offering more energy-efficient solutions. Japan Tobacco plans to further advance this Quantum AI-driven drug discovery technology for molecular design.
D-Wave Quantum (NYSE: QBTS) e la divisione farmaceutica di Japan Tobacco hanno completato con successo un progetto innovativo di prova di concetto che combina il calcolo quantistico con l'IA per la scoperta di farmaci. Il progetto ha dimostrato che i modelli linguistici di grandi dimensioni (LLM) potenziati dal flusso di lavoro ibrido quantistico di D-Wave hanno superato i metodi di calcolo classico nella generazione di nuove strutture molecolari simili a farmaci.
I risultati chiave mostrano che il sistema IA ibrido quantistico ha prodotto più molecole valide con una maggiore somiglianza ai farmaci rispetto agli approcci tradizionali. La collaborazione ha utilizzato la tecnologia di calcolo quantistico ad annealing di D-Wave all'interno del framework IA di JT per addestrare modelli di architettura transformer, simili a ChatGPT, per esplorare lo spazio chimico.
La prova di concetto di successo convalida il potenziale del calcolo quantistico di migliorare le capacità dell'IA, offrendo al contempo soluzioni più energeticamente efficienti. Japan Tobacco prevede di avanzare ulteriormente questa tecnologia di scoperta di farmaci guidata dall'IA quantistica per il design molecolare.
D-Wave Quantum (NYSE: QBTS) y la división farmacéutica de Japan Tobacco han completado con éxito un innovador proyecto de prueba de concepto que combina la computación cuántica con IA para el descubrimiento de fármacos. El proyecto demostró que los Modelos de Lenguaje de Gran Tamaño (LLM) mejorados con el flujo de trabajo híbrido cuántico de D-Wave superaron a los métodos de computación clásica en la generación de nuevas estructuras moleculares similares a fármacos.
Los hallazgos clave muestran que el sistema de IA híbrido cuántico produjo más moléculas válidas con una mayor similitud a los fármacos en comparación con los enfoques tradicionales. La colaboración utilizó la tecnología de computación cuántica por recocido de D-Wave dentro del marco de IA de JT para entrenar modelos de arquitectura transformer, similares a ChatGPT, para explorar el espacio químico.
La exitosa prueba de concepto valida el potencial de la computación cuántica para mejorar las capacidades de la IA, al tiempo que ofrece soluciones más energéticamente eficientes. Japan Tobacco planea avanzar aún más en esta tecnología de descubrimiento de fármacos impulsada por IA cuántica para el diseño molecular.
D-Wave Quantum (NYSE: QBTS)와 일본 담배의 제약 부문이 약물 발견을 위해 양자 컴퓨팅과 AI를 결합한 혁신적인 개념 증명 프로젝트를 성공적으로 완료했습니다. 이 프로젝트는 D-Wave의 양자-하이브리드 워크플로로 강화된 대형 언어 모델(LLM)이 새로운 약물 유사 분자 구조를 생성하는 데 있어 기존 컴퓨팅 방법보다 우수하다는 것을 입증했습니다.
주요 발견은 양자-하이브리드 AI 시스템이 전통적인 접근 방식에 비해 더 높은 약물 유사성을 가진 유효한 분자를 더 많이 생성했다는 것입니다. 이 협업은 JT의 AI 프레임워크 내에서 D-Wave의 어닐링 양자 컴퓨팅 기술을 활용하여 ChatGPT와 유사한 트랜스포머 아키텍처 모델을 훈련시켜 화학 공간을 탐색했습니다.
성공적인 개념 증명은 양자 컴퓨팅이 AI 능력을 향상시킬 수 있는 잠재력을 검증하며, 더 에너지 효율적인 솔루션을 제공합니다. 일본 담배는 분자 설계를 위한 이 양자 AI 기반 약물 발견 기술을 더욱 발전시킬 계획입니다.
D-Wave Quantum (NYSE: QBTS) et la division pharmaceutique de Japan Tobacco ont réussi à achever un projet de preuve de concept révolutionnaire combinant l'informatique quantique et l'IA pour la découverte de médicaments. Le projet a démontré que les grands modèles de langage (LLM) améliorés par le flux de travail hybride quantique de D-Wave ont surpassé les méthodes de calcul classiques dans la génération de nouvelles structures moléculaires semblables à des médicaments.
Les résultats clés montrent que le système IA hybride quantique a produit plus de molécules valides avec une plus grande similarité aux médicaments par rapport aux approches traditionnelles. La collaboration a utilisé la technologie de calcul quantique par recuit de D-Wave dans le cadre de l'IA de JT pour former des modèles d'architecture transformer, similaires à ChatGPT, afin d'explorer l'espace chimique.
La preuve de concept réussie valide le potentiel de l'informatique quantique pour améliorer les capacités de l'IA tout en offrant des solutions plus économes en énergie. Japan Tobacco prévoit de faire progresser encore cette technologie de découverte de médicaments pilotée par l'IA quantique pour la conception moléculaire.
D-Wave Quantum (NYSE: QBTS) und die Pharmaabteilung von Japan Tobacco haben erfolgreich ein bahnbrechendes Konzeptvalidierungsprojekt abgeschlossen, das Quantencomputing mit KI für die Arzneimittelentdeckung kombiniert. Das Projekt hat gezeigt, dass große Sprachmodelle (LLMs), die mit D-Waves quanten-hybriden Workflow verbessert wurden, in der Generierung neuartiger, arzneimittelähnlicher molekularer Strukturen besser abschnitten als klassische Rechenmethoden.
Wichtige Ergebnisse zeigen, dass das quanten-hybride KI-System mehr gültige Moleküle mit höherer Arzneimittelähnlichkeit im Vergleich zu traditionellen Ansätzen erzeugte. Die Zusammenarbeit nutzte die annealing Quantencomputing-Technologie von D-Wave innerhalb des KI-Rahmenwerks von JT, um Transformer-Architekturmodelle, ähnlich wie ChatGPT, zu trainieren und den chemischen Raum zu erkunden.
Der erfolgreiche Konzeptnachweis validiert das Potenzial von Quantencomputing zur Verbesserung der KI-Fähigkeiten und bietet gleichzeitig energieeffizientere Lösungen. Japan Tobacco plant, diese auf Quanten-KI basierende Technologie zur Arzneimittelentdeckung für das molekulare Design weiter voranzutreiben.
- Quantum-hybrid workflow demonstrated superior performance over classical methods in drug discovery
- Successfully generated more valid molecules with higher drug-likeness than traditional approaches
- Validated quantum computing's potential for enhancing AI capabilities while reducing energy consumption
- Project still in proof-of-concept phase, not yet implemented in actual drug development
- Further development and validation needed before commercial application
Insights
The successful quantum-hybrid proof-of-concept between D-Wave and Japan Tobacco represents a significant technological validation for D-Wave's quantum annealing technology in a commercial application. The results demonstrate measurable superiority over classical computing methods for LLM training in drug discovery - an achievement that moves beyond theoretical benefits to practical application.
What's particularly notable is that the quantum-hybrid workflow produced molecules with higher drug-likeness scores compared to both the training data and classical methods, indicating the quantum processing unit provided higher quality, lower energy samples. This validates the core value proposition of D-Wave's annealing approach - finding optimal solutions in complex search spaces.
This project addresses one of quantum computing's most promising near-term commercial applications: enhancing AI capabilities. While full-scale quantum advantage remains elusive in many domains, hybrid approaches like this represent the most viable path to commercial value in the NISQ (Noisy Intermediate-Scale Quantum) era.
For D-Wave, pharmaceutical partnerships offer a lucrative commercial pathway, as drug discovery costs average
This proof-of-concept represents a meaningful advancement in applying quantum computing to pharmaceutical R&D workflows. Traditional drug discovery faces diminishing returns using conventional computational methods, with hit rates typically below
The quantum-hybrid approach demonstrated two critical advantages: generating more valid molecular structures and producing compounds with enhanced drug-likeness profiles without explicit property optimization. This suggests quantum-assisted AI could potentially expand accessible chemical space beyond what's typically explored with classical methods.
For context, pharmaceutical companies typically screen millions of compounds to identify viable drug candidates. Enhanced generative models could dramatically reduce this search space, focusing resources on compounds with higher success probability. Japan Tobacco's commitment to further develop quantum AI-driven drug discovery indicates confidence in the approach's commercial viability.
What's particularly noteworthy is the use of quantum annealing for LLM training rather than just optimization problems, expanding quantum computing's potential application in pharmaceutical workflows. While still early-stage, this approach could eventually impact multiple phases of drug development including lead generation, optimization, and even ADME-Tox prediction - all significant bottlenecks in current development pipelines.
The pharmaceutical industry remains cautious about emerging technologies, making this published validation especially significant as potential evidence of quantum computing's practical utility in life sciences R&D.
Quantum computing project aims to enhance the speed and quality of drug development processes to create first-in-class small molecule pharmaceuticals
The work demonstrated that LLM hybrid models that used classical computation together with D-Wave’s quantum processing unit (QPU) resulted in more valid generated molecules when compared to classical methods alone. In addition, the molecules generated by QPU-assisted LLM training showed a higher quantitative estimate of drug-likeness compared to the training dataset and the models trained with classical computation-driven LLM training methods. This indicates that the QPU provided the teams with higher quality, lower energy samples, highlighting the potential benefits of quantum computing in generative AI for drug discovery.
The goal of this project is to accelerate the discovery of first-in-class small-molecule compounds while improving quality and speed in various processes. In this proof-of-concept, D-Wave’s annealing quantum computing technology was used in JT’s AI technology framework to train LLMs such as a transformer architecture — the same engine behind ChatGPT — for the exploration of chemical space. This enabled the teams to evaluate the feasibility of building a machine learning framework capable of handling a broader range of molecular properties and activities of compounds, which, in our view, initiates a new stage in the use of Quantum AI technologies for drug discovery. By combining AI and quantum computing technologies, the project further confirmed the potential for facilitating the small-molecule compound discovery process in both quality and speed of drug development.
“We are excited by the results we are seeing from our proof-of-concept project with D-Wave. In the experiment, with support from D-Wave’s professional services and product R&D teams, we utilized D-Wave’s annealing quantum computer to train JT’s AI model,” said Dr. Masaru Tateno, Chief Scientific Officer of Central Pharma Research Institute. “Our quantum-hybrid AI system shifted generated compounds to a more ‘drug-like’ molecular ensemble than the training dataset, without imposing any driving factors of molecular properties in our AI model. To the best of our knowledge, this is the first work for annealing quantum computation to outperform classical results concerning LLM training in drug discovery. This validation has also revealed that annealing quantum computing systems can deliver high quality, low energy samples that could drive enhanced performance in generative AI architectures. So, moving forward, with D-Wave’s quantum annealing machines, we aim to maximize the use of quantum computing hardware characteristics and accelerate our efforts in achieving Quantum AI-driven drug discovery.”
“AI has made impressive advancements but faces a computational challenge due to escalating power needs and costs,” said Dr. Alan Baratz, CEO of D-Wave. “Quantum computing’s integration with AI and machine learning could offer scalable, energy-efficient solutions to address these issues and potentially offer enhanced AI capabilities. We believe that our work with JT is an important demonstration and validation of quantum’s integration with AI. When used together, these powerful technologies can help customers build more efficient, rapid, and energy-saving AI and machine learning workloads. While we are just at the beginning of exploring Quantum AI’s potential impact, in our view, this work is a resounding step forward.”
Following the proof-of-concept project, the pharmaceutical division of JT plans to further advance the development of Quantum AI-driven drug discovery technology and then use quantum computing technology for molecular design.
About JT Pharmaceutical division
JT commenced its pharmaceutical business in 1987 with its mission to create original and innovative drugs for patients suffering from diseases around the world. To learn more, visit https://www.jt.com/about/division/pharma/index.html.
About D-Wave Quantum Inc.
D-Wave is a leader in the development and delivery of quantum computing systems, software, and services. We are the world’s first commercial supplier of quantum computers, and the only company building both annealing and gate-model quantum computers. Our mission is to help customers realize the value of quantum, today. Our 5,000+ qubit Advantage™ quantum computers, the world’s largest, are available on-premises or via the cloud, supported by
Forward-Looking Statements
Certain statements in this press release are forward-looking, as defined in the Private Securities Litigation Reform Act of 1995. These statements involve risks, uncertainties, and other factors that may cause actual results to differ materially from the information expressed or implied by these forward-looking statements and may not be indicative of future results. These forward-looking statements are subject to a number of risks and uncertainties, including, among others, various factors beyond management’s control, including the risks set forth under the heading “Risk Factors” discussed under the caption “Item 1A. Risk Factors” in Part I of our most recent Annual Report on Form 10-K or any updates discussed under the caption “Item 1A. Risk Factors” in Part II of our Quarterly Reports on Form 10-Q and in our other filings with the SEC. Undue reliance should not be placed on the forward-looking statements in this press release in making an investment decision, which are based on information available to us on the date hereof. We undertake no duty to update this information unless required by law.
View source version on businesswire.com: https://www.businesswire.com/news/home/20250331962466/en/
Media Contacts:
D-Wave
Alex Daigle
media@dwavesys.com
Source: D-Wave Quantum Inc.
FAQ
What results did D-Wave (QBTS) achieve in their quantum computing drug discovery project?
How does D-Wave's quantum technology improve drug discovery processes?
What technology did QBTS use in their collaboration with Japan Tobacco?
What are the next steps for Japan Tobacco following the QBTS quantum computing project?
