I recently read about the FDA’s investigation into a reported pediatric death in a patient treated with Adzynma. Situations like this are always difficult. At the same time, they remind us why post-marketing safety work is so important for the…
Malignant Pleural Mesothelioma is one of the most difficult cancers to treat. Its biology is driven by inflammation, fibrosis, and a highly suppressive tumor environment. This makes it an important target for new ideas. Using our AI-powered drug repurposing platform…
Over the next few weeks, I will do a deep dive into AI in pharmacovigilance. I will read the key papers in this field and share what I learn with my network. My goals are simple. Become more knowledgeable about…
That was the final question in The Minds of Modern AI, where Geoffrey Hinton, Yoshua Bengio, Yann LeCun, Fei-Fei Li, Bill Dally, and Jensen Huang shared their perspectives. Their answers showed not just different timelines, but different philosophies of intelligence…
IPulse monitors the latest discoveries in biomedical sciences by continuously parsing new literature and extracting structured knowledge from it. We started with ESMO 2025. Our system processed all the abstracts and automatically identified: • Treatments • Targets • Clinical trial…
One surprising outcome from evaluating our drug repurposing system is how far into the future it can generate meaningful hypotheses – years before the biomedical community recognizes them. Here are three examples from our cystic fibrosis (CF) work. 1. Tamoxifen…
They often showcase how well a method can “rediscover” drug–disease links that already appear in the literature. That may look convincing, but it misses the real challenge: finding drug candidates that nobody has talked about before. In a true repurposing…
I’ve decided to go much deeper into drug safety and pharmacovigilance. Over the past year, we discovered that our LLM-based NLP approach works surprisingly well for PV. That led us to develop the IPV tool, and the results convinced me…
One of the toughest challenges in automated drug repurposing and hypothesis generation is validation. Most AI models are tested on data that already include the answers, which often leads to serious information leakage. The model looks good on paper, but…
That question came up in The Minds of Modern AI, a roundtable with Geoffrey Hinton, Yoshua Bengio, Yann LeCun, Fei-Fei Li, Bill Dally, and Jensen Huang — six pioneers who built the foundations of modern AI. Their responses were thoughtful,…
