New Cancer Treatment Models
Cancer causes 9.6 million deaths in 2018 and 18 million new cases recorded in the same year. Cancer caused by many causes such as genetic predisposition, environmental pollution, smoking and alcohol use, nutrition patterns. In addition, it can show itself in different forms. And we have so many things for this subject.
Related News: Will Polio be a history?
IBM is looking for cancer treatment methods
In Zurich, a study group of IBM are performing artificial intelligence and machine learning to make cancer easier for us to understand key factors, molecular mechanisms, and tumor composition of this complex disease.
IBM states “ Our goal is to deepen the knowledge about cancer and to equip the industry and academy with information to one day uncover completely new treatments.”.
Here, there are 3 cancer treathment projects which are formed by Artificial Intelligence.
1 – PaccMann Project
The first project of IBM is named “PaccMann”. It is an abbreviation of “prediction of anticancer compound sensitivity with Multi-modal attention-based neural networks”. So, it is not related with Pac-Man computer game.
Fighting with cancer can only take millions of dollars to develop a single drug. In addition, financial constraints can delay and reduce our potential to develop new drugs and therapies.
The algoritm of PaccMann, which is studied by IBM, can automatically analyze chemical compounds and estimate that which ones are more effective in combating cancer resistance and facilitate the process. This machine learning algorithm uses information about the molecular structure of chemical compounds as well as gene definitions. IBM estimates that early detection of anti-cancer compounds will reduce the high costs associated with drug development.
2 – INtERAcT
The second project of the brand is INtERAcT which is abbreviation of “Interaction Network infErence from vectoR representATions of words”. The main aim of this tool is extracting information automatically about cancer treatment methods and projects from online academic publications.
Every year, roughly 17 thousand publications are published about cancer. It is difficult, if not impossible, for researchers to follow the developments in all these publications. INtERAcT provides the ability to automatically extract information from scientific publications by eliminating this difficulty on the academic side.
3 – PIMKL Project
The third and last project of IBM, which are about cancer treatment methods is PIMKL. It is abbreviation of “Pathway-induced multiple kernel learning”. It’s mean is “ path-specific multi-core learning”.
This algorithm allows us to estimate the cancer progression and possible recurrence in patients by using sets of data we know.
PIMKL uses multi-core learning to identify molecular structures. So it categorizes patients for searching an individual cancer treatment method.