At Gramener, our Subject Matter Experts (SMEs) share their perspectives on the latest innovations in the data & AI world. In this article, Santosh Shevade, our Principal Data Consultant from the Pharma Analytics team, shares information on GPTs, the EU Clinical Trials Information System (CTIS), Med-PaLM2, ClinicalGPT, PMC-LLAMA, and MedAlpaca. Santosh has also been recognized as LinkedIn's top Artificial Intelligence (AI) voice.
The All New GPTs from OpenAI
OpenAI has introduced GPTs, which is a special version of ChatGPT designed for specific purposes.
People can easily make these versions without needing to know how to code. This lets users customize ChatGPT to be more helpful in their everyday lives, at work, or at home.
Recently, Bill Gates wrote a blog about how the progress in Generative AI will change how we use computers. We command our computers which apps to use for different tasks, but Bill Gates thinks that will be different in the next five years. He imagines a future where AI agents, which understand natural language, will know about our personal and professional lives. This means we won't have to use different apps for different things.
These AI agents will be much better than what we have now. They'll give personalized help and can do lots of different things with the user's permission. According to Gates, these agents are the most significant change in computing since we switched from typing commands to using graphical interfaces.
What do GPTs mean?
When we look at OpenAI's introduction of GPTs, which are customized versions of ChatGPT for specific jobs, we notice a shift towards AI tools that are more personalized, smart, and versatile. GPTs let people create special AI helpers without needing to know how to code, and these can be used in different parts of our daily lives and work.
OpenAI is also putting a lot of effort into making sure that GPTs are private and safe, which matches Bill Gates' idea of having AI helpers deeply connected to our personal and professional lives. Putting all this together, it looks like the future will have AI helpers – GPTs and others changing how we use technology. This change will make technology easier to use, more effective, and more tailored to each person.
Transparency Rules & The EU Clinical Trials Information System (CTIS)
The European Medicines Agency (EMA) is asking the public for input to improve the rules about sharing information on clinical trials in the EU. They want to make sure that the rules are clear and that people's privacy is protected.
The goal is to make it easier for people to use the system and to keep information safe.
If the suggested changes are accepted after getting feedback from the public, they will follow the rules set by the Clinical Trials Regulation (CTR).
The idea is to simplify things and reduce paperwork while protecting personal and private information. The Clinical Trials Information System (CTIS) helps make it easy to apply for and oversee clinical trials in Europe, making it a good place for this kind of research.
The CTIS is like a main hub for clinical trial applications and assessments, and there's a public database that doctors and the public can check. The EMA is in charge of the CTIS, and the European Commission is watching over how it's used.
What does this mean?
Generative AI can really help organizations follow the new rules of the EU Clinical Trials Information System (CTIS). It automatically checks if everything follows the rules, keeps private information safe, makes reports easier, and provides effective training. This technology improves how data is analyzed while making sure that clinical trials are transparent and keep things confidential.
But, it's important to use advanced analytics with a reliable partner to make sure the data is correct and follows the changing rules. This way, we can keep the right balance between being transparent about clinical trials and keeping things private and confidential.
Large Language Models (LLMs) in Pharma – A New Dawn
This research paper thoroughly explores the shift from basic Pretrained Language Models (PLMs) to more advanced LLMs. The authors looked at nearly 18 models, such as Med-PaLM2, ClinicalGPT, PMC-LLAMA, and MedAlpaca, in their survey.
The survey talks about how LLMs make healthcare applications better and more efficient compared to older PLMs. It shows the improvements in LLMs, such as how they can deal with problems PLMs had before (not being expressive enough, hard to explain, and needing a lot of data).
LLMs, with their Chain-of-Thought (CoT) ability and impressive few-shot or zero-shot capabilities, are now better integrated into healthcare.
What does this mean?
This is a new trend for the future growth of generative AI in healthcare. It talks about trying to improve medical knowledge using methods like Retrieval-based LLMs. It also discusses the difficulties in connecting these AI applications with other systems, making them work well with patients and doctors, and a crucial aspect such as handling LLM hallucinations, misunderstandings, and prompt brittleness.
That's it for this edition, and we'll be back soon with another edition of the AI Chronicles from Gramener.
The post Latest in Pharma AI: GPTs, EU Clinical Trials Information System (CTIS), Med-PaLM2, and ClinicalGPT appeared first on Datafloq.