The state of the art in some of the advanced technologies space is evolving at a hyper exponential rate. The growth of Large Language Models (LLMs) in the Natural Language Processing (NLP) space is a case in point – the size of LLMs has grown approximately 3000% in just the last 2 years, leaving even Moore’s law far behind! For the enterprise adopters and users, it might be impossible to stay on top of all the developments, but they must build a basic understanding of foundational technologies, and how their evolution can unlock use cases for life science enterprises.
At the Indegene Digital Summit 2022, Tarun Mathur, CTO of Indegene, broke down some of the advanced high-profile technologies such as blockchain, metaverse, and natural language models, and shared some early use cases in the life science enterprises. The key highlights from his session are summarized below.
At its heart, blockchain is essentially a trusted ledger where valid transactions are appended to the ledger. This ledger is replicated across a network of computer nodes and has ongoing updates to it through the computer nodes. The validity of transactions is determined by algorithm-based consensus (Ref: proof of work and proof of stake)
Blockchains are classified into 3 types that are public, private, and hybrid.
Enterprises across all major industries including life sciences are adopting blockchain for various scalable use cases.
Prominent examples within life sciences include Blockpharma (drug traceability space), Kitchain (clinical supply chain), Pharmaledger (clinical trial supply chain), Patientory (patient data), Embleema (bioinformatics), Mediledger (supply chain platform), and Procredex (verification and credentialing).
Smart contracts are programs stored on a blockchain that are guaranteed to execute when pre-defined conditions are met. Decentralized Apps (Dapps) are distributed user-facing front-end applications that are built on the top of backend blockchain and run on a smart contract system.
Blockchain is also seen as a supporting infrastructure to realize the vision of the future of the web and Internet – Web 3.0. In Web 3.0, data is decentralized, and data ownership and control lie with individuals rather than enterprises. It is touted that the next generation of a globally connected network could be based on the blockchain with smart contracts powering the network at the backend. The front end of Web 3.0 could be a collection of Dapps, following interoperability standards. This will essentially allow the users to jump from one Dapp to another, much in the same way as users jump from one website to another through embedded links.
Despite all the hype and investment surrounding this space, the true Metaverse does not exist today, as Web 3.0 is not a reality yet! Tony Parisi, one of the early thought leaders in this space, coined the 7 rules of Metaverse, essentially describing what Metaverse is.
The 7 rules of Metaverse
Although Metaverse is hardware independent, immersive experiences are a key component of the Metaverse. And by extension, Extended Reality (XR) is intertwined with Metaverse. The XR is an umbrella term used to encompass 3 forms of immersive experiences – virtual reality (VR), augmented reality (AR), and mixed reality (MR).
There are some small-scale to mid-size use cases within life sciences built on the concepts of Metaverse and XR. These use cases include patient networks, telemedicine, training and simulations, brand marketing events, medical congresses, patient simulation, immersive learning, and so on. In the omnichannel engagement space (HCP), immersive experiences are gaining traction and are now seen as a component of the overall omnichannel strategy.
The state of the art in XR space is evolving rapidly too. The user experiences that can be made available to patients, HCPs, payers in the next 12 months could be dramatically different from what it is today, says Tarun.
In the short term, Tarun advocates that it is a good time for enterprises to conduct experiments and pilots on Metaverse but not to aggressively invest in building complex custom experiences and environments until open interoperability standards are established.
Until then, it might be prudent to run pilots on the existing, off-the-shelf platforms by leveraging repurposed "2D" content to understand which types of experiences draw interest from customers. In the mid to long term, enterprises must identify the winning platforms and start partnering with them to build fit-for-purpose experiences for customers on top of those platforms.
NLP is becoming the primary mechanism of interaction between humans and computers. Use cases in life sciences include de-identification of protected health information, extraction of clinical information from unstructured narratives, enterprise document processing, AI-assisted bots for patient experience, and so on.
LLMs are generally built to be domain-agnostic, and they must be contextualized to the life sciences domain by training them. For example, these models are typically trained using Wikipedia or other such domain-agnostic sources. If they are to be used in the clinical information space, they must be trained/fine-tuned with clinical literature sources such as PubMed.
We will see technologies such as blockchain, metaverse, and LLMs becoming more mainstream in the near future. Their evolution will unlock more and more use cases for life sciences enterprises and help drive substantial outcomes.