DataTalks HFA #2: Found in Translation and NLP for Acute Care 💉

Welcome to the second DataTalks HFA meetup!

DataTalks HFA will bring together local data scientists, researchers, statisticians and data enthusiasts around advanced data science, machine learning and AI topics.

Our second meetup is hosted by GE Healthcare and will demonstrate the potential of machine learning to transform and disrupt healthcare.
○ Rachelly Normand from the Technion will present her groundbreaking research recently published in Nature Methods, utilizing machine learning to promote medical knowledge discovery through mice to human inference.
○ Omer Barkol and Guy Wiener will unmask GE Heathcare's ambitious project to transform patient care in the ICU.

♦ Time: September 23rd, 18:00
♦ Language: Hebrew (both talks)
♦ Location: GE Healthcare (Haifa Life Science Park, 6th floor. Address: Nativ Haor 1, Haifa)
♦ Free parking on Nativ Haor street
♦ Background: Basic knowledge in data science and machine learning is required for understanding - 'seminar level'

Agenda:
• 18:00 - 18:30 - Gathering, snacks & mingling 🍕🍺

• 18:30 - 19:15 - First talk by Rachelly Normand 🐁🧬
Found In Translation: a machine learning model for mouse to human inference

• 19:15 - 20:00 - Second talk by Omer Barkol and Guy Wiener 💉💊
Thinking of Acute Care – not just on #sepsisawarenessmonth

Abstracts for the talks:

• Found In Translation: a machine learning model for mouse to human inference - Rachelly Normand *
  Mice are extremely useful in biomedical research but differ substantially from humans. We developed a machine learning methodology that utilizes public gene expression data to learn the relationship between the species. Applying the model on a new mouse experiment allows the detection of 20-50% more human relevant genes missed by the mouse data alone. This enables better utilization of mouse studies to understand human biology.

• Omer Barkol, PhD: Thinking of Acute Care – not just on #sepsisawarenessmonth *
  GE Healthcare in a collaboration with Roche invest in the development of a new platform to transform how teams of medical professionals deliver patient care. Much of this investment can be credited to the notion that patients don’t suddenly deteriorate - clinicians suddenly notice. While this may be true, detecting that patient deterioration is much more nuanced. One thing is universal: identifying patients at risk of deterioration (prior to requiring heroic efforts to rescue them) is extremely difficult. We use data to allow AI assist both patients and clinicians.

• Guy Wiener, PhD: Understanding Medicalese *
  Medicine is a language, written in English alphabet and grammar, but not always readable to non-clinical English speakers. It consist of drug names, symptoms, body organs, bacteria names, etc., originating from Chemistry, Biology, and Latin, with many acronyms and abbreviations. Modeling it as a language, using publicly-available medical texts, yields surprising results. We will present interesting implications of medical word embeddings, such as overcoming jargon synonyms, generic vs brand medications, relating procedures and conditions, and more.