Kristin Brooks, Contract Pharma04.01.19
Research has found that the costs of bringing on a new drug costs pharmaceutical companies an average of $2 billion, with the main culprits for high costs being high attrition rates and various other costs related to R&D. Vyasa’s Neural Concept Recognition technology, Cortex, aims to help pharma professionals cut through the massive amounts of data needed to begin identifying new chemical compounds during drug R&D. As a result, professionals can utilize data already available to them to identify previously hidden combinations, helping to cut costs and save time.
Dr. Christopher Bouton, chief executive officer of Vyasa Analytics, discusses the company’s deep learning Cortex and how it’s addressing R&D challenges. –KB
Contract Pharma: What are the current obstacles drug developers look to overcome leveraging data?
Christopher Bouton: The drug development process is a fundamentally data-driven activity. As a result, there are numerous elements in the process which involve the generation of large amounts of data of numerous types – including text, image, small compound and quantitative content. The challenge with all of this data, as we found during the “Big Data” phase, is it is not useful just for the sake of having a lot of data. Rather, the challenge is finding more effective ways of deriving insights from the vast troves of data that organizations now have access to.
In particular, having better tools for identifying unexpected patterns in the data which may lead to an “ah-ha” moment for a scientist would be a game-changing value driver. It just so happens that novel deep learning (i.e. AI) technologies are exceptionally good at this type of robust pattern recognition in data sources without the need for a priori rule set definitions.
Once we train a deep learning/AI agent on what to look for in data, it can climb into large data sources and sift through that content on a scale and with a speed that is simply not possible for humans in order to identify interesting trends, patterns and insights in the content.
Contract Pharma: When it comes to data management, what are the greatest hurdles?
CB: Integration across numerous sources and types of content is one of the greatest hurdles in data management today. It’s very difficult to define and construct a common format or structure that will accommodate all of the data that is relevant to a given task and relevant set of data silos. As a result, we need better, smarter approaches to deriving structured information form disparate content types and sources, irrespective of where those sources reside in an organization and what sorts of content they contain.
We’ve developed Vyasa Cortex to be able to connect to a wide range of data sources and to be able to send AI agents into these sources to identify relevant concepts, parameters associated with them and their relationships in a flexible, highly scalable manner.
Contract Pharma: What are the biggest contributing factors that cause delays in R&D?
CB: As is the case in many other industries, one key contributing factor is that delays in R&D can be caused by a wide range of inefficiencies in the constructive use of data to identify trends, patterns, interconnections and insights that would enable better decision making for value generation and path finding during project work.
What is needed are systems capable of doing a better job of recognizing concepts, relationships and parameters in a wide range of content types (images, text, quantitative data) irrespective of formatting and type. We built Cortex to be able to connect to a wide range of data sources and be able to flexibly enable AI agents to crawl through the data in those sources creating a birds eye view of the information and interconnection available in those sources.
CP: What opportunities are available to drive cost reduction and greater efficiencies?
CB: Deep learning/AI systems present a novel capability to identify patterns in data without the need for the definition of rule sets or heuristics. Essentially, we can train an AI agent on the identification of a wide range of what we call “concept types” in disparate data sources.
For example, things like genes, proteins or diseases in text, all the way to certain types of images or the objects in them can be defined as “concepts” in data. The more AI agents that we train to crawl through a given set of data sources, the more insights and efficiencies you can gain from those data sources.
CP: What advances have been made leveraging analytics?
CB: Nearly every organization today needs to reimagine itself as a data analytics company. The amount of digital data being generated on a moment-to-moment basis in the world means that there is an absolute wealth of information available to each business to leverage for efficiency gains, competitive advantage and enhanced decision making potential.
At first, we will see these advantages come from the application of data analytics in a wide range of “low hanging fruit” scenarios where the value of novel deep learning/AI approaches can readily increase the capability of given workflows and processes across a wide range of vertically centered organizations from life sciences to healthcare to legal to fintech.
I feel extremely lucky to be working at a time when we’ve been given this brand new tool (deep learning/AI) when we so desperately need it in order to more effectively utilize the wealth of data available to us across many sectors. Our mission at Vyasa is to enable a wide range of these capabilities for more effective leveraging of digital content through novel analytical approaches.
Dr. Christopher Bouton, Vyasa’s founder and CEO, received his BA in Neuroscience from Amherst College in 1996 and his Ph.D. in Molecular Neurobiology from Johns Hopkins University in 2001. Previously Dr. Bouton was the CEO of Entagen a software company founded in 2008 that provided innovative Big Data products including Extera and TripleMap. Entagen's technologies won numerous awards including the "Innovative Technology of the Year Award for Big Data" from the Massachusetts Technology Leadership Council in 2012 and Entagen was recognized as a Gartner "Cool Vendor" in the Life Sciences in 2013. Entagen was acquired by Thomson Reuters in 2013. Dr. Bouton is an author on over a dozen scientific papers and book chapters and his work has been covered in a number of industry news articles.
Dr. Christopher Bouton, chief executive officer of Vyasa Analytics, discusses the company’s deep learning Cortex and how it’s addressing R&D challenges. –KB
Contract Pharma: What are the current obstacles drug developers look to overcome leveraging data?
Christopher Bouton: The drug development process is a fundamentally data-driven activity. As a result, there are numerous elements in the process which involve the generation of large amounts of data of numerous types – including text, image, small compound and quantitative content. The challenge with all of this data, as we found during the “Big Data” phase, is it is not useful just for the sake of having a lot of data. Rather, the challenge is finding more effective ways of deriving insights from the vast troves of data that organizations now have access to.
In particular, having better tools for identifying unexpected patterns in the data which may lead to an “ah-ha” moment for a scientist would be a game-changing value driver. It just so happens that novel deep learning (i.e. AI) technologies are exceptionally good at this type of robust pattern recognition in data sources without the need for a priori rule set definitions.
Once we train a deep learning/AI agent on what to look for in data, it can climb into large data sources and sift through that content on a scale and with a speed that is simply not possible for humans in order to identify interesting trends, patterns and insights in the content.
Contract Pharma: When it comes to data management, what are the greatest hurdles?
CB: Integration across numerous sources and types of content is one of the greatest hurdles in data management today. It’s very difficult to define and construct a common format or structure that will accommodate all of the data that is relevant to a given task and relevant set of data silos. As a result, we need better, smarter approaches to deriving structured information form disparate content types and sources, irrespective of where those sources reside in an organization and what sorts of content they contain.
We’ve developed Vyasa Cortex to be able to connect to a wide range of data sources and to be able to send AI agents into these sources to identify relevant concepts, parameters associated with them and their relationships in a flexible, highly scalable manner.
Contract Pharma: What are the biggest contributing factors that cause delays in R&D?
CB: As is the case in many other industries, one key contributing factor is that delays in R&D can be caused by a wide range of inefficiencies in the constructive use of data to identify trends, patterns, interconnections and insights that would enable better decision making for value generation and path finding during project work.
What is needed are systems capable of doing a better job of recognizing concepts, relationships and parameters in a wide range of content types (images, text, quantitative data) irrespective of formatting and type. We built Cortex to be able to connect to a wide range of data sources and be able to flexibly enable AI agents to crawl through the data in those sources creating a birds eye view of the information and interconnection available in those sources.
CP: What opportunities are available to drive cost reduction and greater efficiencies?
CB: Deep learning/AI systems present a novel capability to identify patterns in data without the need for the definition of rule sets or heuristics. Essentially, we can train an AI agent on the identification of a wide range of what we call “concept types” in disparate data sources.
For example, things like genes, proteins or diseases in text, all the way to certain types of images or the objects in them can be defined as “concepts” in data. The more AI agents that we train to crawl through a given set of data sources, the more insights and efficiencies you can gain from those data sources.
CP: What advances have been made leveraging analytics?
CB: Nearly every organization today needs to reimagine itself as a data analytics company. The amount of digital data being generated on a moment-to-moment basis in the world means that there is an absolute wealth of information available to each business to leverage for efficiency gains, competitive advantage and enhanced decision making potential.
At first, we will see these advantages come from the application of data analytics in a wide range of “low hanging fruit” scenarios where the value of novel deep learning/AI approaches can readily increase the capability of given workflows and processes across a wide range of vertically centered organizations from life sciences to healthcare to legal to fintech.
I feel extremely lucky to be working at a time when we’ve been given this brand new tool (deep learning/AI) when we so desperately need it in order to more effectively utilize the wealth of data available to us across many sectors. Our mission at Vyasa is to enable a wide range of these capabilities for more effective leveraging of digital content through novel analytical approaches.
