Presentation from Bio IT World, Boston | April 16-18, 2019
Track: Data Visualizations & Exploration Tools: Leveraging Machine Learning for Visualization
Session Title: Simplifying the Adoption of Machine Learning for Real Time Clinical Data Insight and Decision Making
Speaker: Christine Lan, Lucidworks
8. Data Sources Data
Transformation
Data
Visualization
Natural Language Processing
> Synonym Generation
Crowd-Sourced Relevancy
> Personalized Recommendations
Security
> Role and Record Level Access Controls
Scalability
> No cubes and aggregations needed
> 4B+ individual records
User Workflow Oriented
12. Lucidworks Fusion + Pharma
Shortened Product Launch Cycle. Connect patients to appropriate clinical
trials through NLP and personalized ranked best matches using more
sources like social media
More Efficient R&D. Break down data silos across multiple functional areas
by surfacing personalized content, experts, and conversations through
smarter enterprise search and knowledge management
Gathering Competitive Intelligence. Monitor the web (including social media
like Twitter) to see what influential people are saying, keep track of patents
through sources like Cortellis Competitive, determine how biosimilars are
being positioned, and aggregate trends
13. Lucidworks Fusion + Pharma
Shifting Global Regulatory Landscape. Surface relevant regulatory
information, e.g. Cortellis Regulatory Intelligence and FDA sites both in US
and abroad
Compliance. Improve site search so patients can more easily find programs
they’re eligible for to decrease medical non-compliance. Better create and
monitor analytics in near real-time from trip reports and PAEs
Site Search Analysis for Marketing. See what people are searching and
clicking on your website so targeted marketing collateral can be created
14.
15. Signals can include
Industry-Leading
Relevance with
Signals
Fusion captures, stores,
and aggregates signals
from a variety of sources to
drive automated relevancy
tuning to deliver best-in-
class relevance.
Clicks and Queries Geo-location
User Behavior
and Preferences
User Search
History
Device
16. App Maturity
Complexity
Stage 3
Machine Learning
Deliver the "next best action" to
the user. Understand intention
with query classification and
intent, clustering,
recommendations, and
conversational search.
Stage 2
Dynamic
Application of Signals,
Automated Relevancy,
OpenNLP, Gazetteer
Stage 1
Curated
Rules, Signal Capture,
Analytics, Security
Stage 0
Basic
Field Boosting, Synonyms,
Phrase Matching
Relevancy Strategy
Notes de l'éditeur
Roche is for “value-based pricing” vs. voluntary price restraints
R&D lifecycle through …
FDA, compliance ect.
Pricing transparency?
Personalized medication for increased drug efficacy?
Patients are identified to enroll in clinical trials based on more sources—for example, social media—than doctors’ visits. Furthermore, the criteria for including patients in a trial could take significantly more factors (for instance, genetic information) into account to target specific populations, thereby enabling trials that are smaller, shorter, less expensive, and more powerful.
Trials are monitored in real time to rapidly identify safety or operational signals requiring action to avoid significant and potentially costly issues such as adverse events2 and unnecessary delays.
Instead of rigid data silos that are difficult to exploit, data are captured electronically and flow easily between functions, for example, discovery and clinical development, as well as to external partners, for instance, physicians and contract research organizations (CROs). This easy flow is essential for powering the real-time and predictive analytics that generate business value.
Site Search, help reduce medical non-compliance by making it easier for patients to find programs they’re eligible for
Roche is for “value-based pricing” vs. voluntary price restraints
R&D lifecycle through …
FDA, compliance ect.
Pricing transparency?
Personalized medication for increased drug efficacy?
Patients are identified to enroll in clinical trials based on more sources—for example, social media—than doctors’ visits. Furthermore, the criteria for including patients in a trial could take significantly more factors (for instance, genetic information) into account to target specific populations, thereby enabling trials that are smaller, shorter, less expensive, and more powerful.
Trials are monitored in real time to rapidly identify safety or operational signals requiring action to avoid significant and potentially costly issues such as adverse events2 and unnecessary delays.
Instead of rigid data silos that are difficult to exploit, data are captured electronically and flow easily between functions, for example, discovery and clinical development, as well as to external partners, for instance, physicians and contract research organizations (CROs). This easy flow is essential for powering the real-time and predictive analytics that generate business value.
Site Search, help reduce medical non-compliance by making it easier for patients to find programs they’re eligible for
Connectors are tricky to build from scratch…usually it takes somewhere between 1-3 months to build a connector…but luckily for you, we have over 30 OOB
We have a ServiceNow connector OOB which customer like AllState Insurance are leveraging. We also have Slack and Box connectors OOB. I know you mentioned OneDrive during our last meeting and we will have that connector in our next release in July.
Returning the most relevant result for a search query is at the heart of providing a great search experience for your users
Lucidworks View UI is continuously keeping track of the key words users are typing in, the result set being returned and the specific search result users are clicking on or (not).
This information is available to Fusion adminstrators through the Search Analytics dashboard
Fusion is continuously capturing this data and aggregating it to automatically tune relevance and provide the most clicked on information for a particular search term on top of the search results.
For example, if a user types in Holiday Schedule and 2017 Holiday schedule is the most clicked search result by previous users, then the 2017 Holiday Schedule will be boosted to the top of the result set.
The location of the user is similarly used to boost local results over other locations.
This is a fully automated process that is running on Apache Spark and managed by Fusion
User Search History, Preferences and Device used to access search can similarly be used as inputs to automatically provide the most relevant result on top
Apache Solr provides good search relevance out of the box, but needs to be continuously monitored and tuned as the underlying data being searched grows and changes
With the Signals capability Fusion takes this even further and dynamically adjusts relevance that is tuned automatically to the way your users are searching
Machine Learning models can be applied with Fusion to make signals based relevance even more powerful for even more smarter and relevant search
Relevance with Fusion Signals is the best in the industry