Volume 07

Navigating the sea of data, deciphering intricate clinical studies, and uncovering valuable insights—healthcare decision-making has reached a remarkable milestone. It's not merely about processing information; AI brings a human touch, enhancing the quality of customer care. It acts as a proactive guardian, detecting genetic indicators and alerting you to potential health challenges before they surface. Yet, the core challenge rests in ethically harnessing AI's power in healthcare—a moral puzzle demanding our responsible navigation.


ai:sight by Fractal Volume 7

and legal guidelines that affect the industry. Their key challenge is ensuring they apply the technology legally and competently. Pharma companies are starting to explore multiple use cases for GenAI, but they want to fully understand the risks before they deploy it.” Conversations with companies across the industry are helping Fractal to point the way. “In recent years, we have evolved our capabilities across the pharma value chain, including marketing and sales, supply chain analytics, data engineering, and AI,” said Sagar Shah, head of pharma practice at Fractal. “As part of that work, we are developing key GenAI use cases that can bring immediate gains for pharma companies. Some of these focus on low-risk areas like boosting internal employee productivity. Others bring an ethical and responsible approach to include GenAI in the healthcare professional (HCP) and patient experience.” Let’s look at some of those use cases.

from cutting-edge technologies is undeniable in diverse organizations. However, when it comes to the highly regulated pharmaceutical industry, the stakes are exceptionally high. A misstep in this arena can result in excessive costs and profound consequences. Imagine, for instance, a chatbot providing inaccurate information to patients or a clinical study marred by bias favoring a particular group of participants. These scenarios encompass a broad spectrum of potential pitfalls, including to secure product approval, and tarnishing a meticulously cultivated reputation. Given these formidable approaching the realm of GenAI with a blend of eager anticipation and measured caution. “Since ChatGPT took over the world, company leaders have been pushing them,” said Ashok Vardhan, client partner for pharma practice at Fractal. “However, pharma companies are generally more focused on data security and other regulatory


Volume 7 ai:sight by Fractal


One colossal challenge looms in the pharmaceutical industry’s relentless data landscape: the sheer volume of information. Amid the chaos of documents and decisions, busy teams often get bogged down. Enter GenAI, the solution they’ve been waiting for. “GenAI can accelerate our time to delivery in our core business processes,” shares Jeevaka Kiriella, director of global data science at Merck. “One fundamental way to do that is by leveraging its generative aspect as a starting point or foundation. For instance, in the realm of clinical trial operations, one use case involves the creation of protocols and related study documents. In this upstream activity, input to a large language model (LLM), the LLM can generate initial protocols This should decrease the time needed to create and revise such documents, ultimately expediting the initiation of a trial.” and increase product yield. “GenAI allows us to access the entire pharmaceutical supply chain, including research, trials, manufacturing, and Kiriella also sees GenAI helping to improve product availability commercialization. This enables faster decision-making, ultimately leading to improved product availability. The pharmaceutical industry is actively investing in computational drug discovery powered by GenAI, in research.”

we can further optimize manufacturing processes using interactive LLMs to identify bottlenecks and enhance product yield. Additionally, in the commercial sphere, foundational models can be utilized for tasks such as target materials, and gaining assisted access to primary market research insights.” As individual organizations identify the use cases that best suit their business, GenAI’s transformative effects are already extending across the industry. leap forward in applying AI/ML with the adoption of foundational models,” Kiriella reveals. “This shift allows for a drastic reduction in the time needed to develop AI applications, leading to release of LLMs, AI is now accessible to

users throughout the company. In the pharmaceutical sector, we have a unique opportunity to harness this potential due to our investments in data and analytics.” considerations that will help pharma companies ensure long-term success with the technology. “Organizations should promptly assess and address the potential impact of GenAI. In addition, they must responsively meet the this may involve securely integrating commercial foundational models to mitigate compliance risks or effectively processing large amounts of internal data to drive foundational to remain open to different models, leveraging the most suitable one for each use case.”

large foundational models is clinical operations. By optimizing this process, we can swiftly evaluate viable


ai:sight by Fractal Volume 7


Unraveling the mysteries of disease

The quantum Fractal’s research illustrates that, even with current quantum computing capabilities, it is: Faster than classical computing. As accurate as or even more accurate than classical computing. Capable of tackling complex, molecular-level tasks that classical computers deem impossible. However, this is merely the dawn of the quantum computing era. Nurturing talent and establishing standards are crucial to driving this technology forward for the greater good. Fractal’s commitment to mentoring and collaborating with external organizations, including the IIT Mumbai, Quantum Open Source Foundation, and various quantum initiatives, underscores its dedication to building a global pool of quantum computing expertise.


Managing complex diseases like Alzheimer’s, Huntington’s, Parkinson’s, cancer, and even COVID-19 hinges on understanding the molecular structure of proteins and how they fold. Classical computing struggles to simulate these processes, leading scientists to rely on slow and costly laboratory experiments. Quantum computing offers an alternative. Using a variational quantum eigensolver (VQE) algorithm, Fractal’s team simulated a portion of the Alzheimer’s molecule folding into compared with DeepMind Alphafold, a cutting-edge classical computing AI system for protein structure prediction. Surprisingly, even with a small quantum processor and modest settings, the quantum algorithm matched Alphafold’s accuracy.

As quantum computing advances astonishingly, the future holds exciting possibilities. Research is already showing us how even the smallest quantum computer can surpass the capabilities of classical ones. Just four years ago, molecular simulation was impossible, but today, we can simulate and dissect atomic-level processes like protein folding on a computer. Breakthroughs in quantum AI could revolutionize industries from energy production to geopolitics. It’s foreseeable that quantum computing will be applied to real- world problems within the next decade, and organizations must prepare to harness its potential. In conclusion, the age of quantum computing is upon us, and the opportunities it presents are limitless. It’s time for organizations to embrace this transformative technology and chart their course toward a future powered by quantum computation. Want to learn more about the publication. Read the paper here:

Quantum-powered drug discovery

Another area of exploration involves the application of quantum generative AI to create molecules for medicinal drugs. Fractal developed a quantum version of a generative adversarial network (GAN), a machine learning algorithm where two neural networks compete to enhance their predictions. These quantum GANs (QGANs) were compared with state-of-the-art architectures and demonstrated exceptional performance, yielding the most promising drug candidates.

Quantum computing

Fractal built a hybrid quantum neural classical and quantum layers. Using housing data to predict property prices in Boston, USA, this hybrid network showcased superior accuracy and generalization capabilities compared to classical counterparts. As quantum technology matures, hybrid approaches like this could

Prateek Jain Lead Architect, AI@

Prateek combines an inquisitive mind with a lifelong passion for science and deep experience in AI and machine learning. As a founding member of Fractal’s quantum computing research team, he is focused on helping to build the knowledge and nurture the talent that will drive this new chapter in technology.

Scale, Machine Vision and Conversational AI, Fractal


ai:sight by Fractal Volume 7


Qure.ai’s innovative approach extends beyond diagnosis. Areas they’re looking at include the early diagnosis of heart failure . A recent study has shown that recommendations from the Qure.ai algorithm helped identify 50 new heart failure patients from 5,000 routine chest X-rays. They’re pioneering models that organizations can employ to create solutions, potentially giving medics access to AI-powered insights about medical images. Patients may even have the option to obtain AI- generated reports alongside those from radiographers. As regulations progress, we anticipate fully automated screening processes for X-rays and blood tests, diminishing the necessity for manual valuation or medical professional sign-offs. Presently limited to TB screening, this innovation foreshadows a future where broader areas embrace automated diagnostic protocols, reshaping healthcare procedures. Right now, though, our biggest priority is to extend our solutions to more people worldwide. We’ve already touched about 25 million patients and aim to reach a billion in the This is AI for good, transforming the path to treatment for some of our biggest public health challenges.

Stroke is another condition where every minute counts. Medics have two choices. If no bleeding is involved, they can give anticoagulant drugs in the emergency room to dissolve clots. They can coordinate the resources to send the patient for more complex surgery if bleeding is present. Ruling out bleeding is crucial before deciding which treatment to give, but this can take hours to do manually. swiftly interpret CT scans to rule out bleeding, ensuring that more stroke patients receive prompt, accurate treatment, ultimately improving their chances of full recovery.

Prashant has deep expertise in AI, data science, and optimization. He joined Fractal as chief data scientist in 2015 when it acquired its own AI company, which provided targeted advertising for e-commerce firms. He incubated Qure.ai within Fractal to apply AI healthcare, especially in interpreting images like X-rays and CT scans.'

Prashant Warier https://www.qure.ai/ Founder & CEO, Qure.ai


ai:sight by Fractal Volume 7

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20 Page 21 Page 22 Page 23 Page 24 Page 25 Page 26 Page 27 Page 28 Page 29 Page 30 Page 31 Page 32

Made with FlippingBook - PDF hosting