Periodic Reporting for period 1 - IRIS AI (IRIS.AI: The Artificial Intelligence-powered R&D assistant)
Okres sprawozdawczy: 2018-11-01 do 2019-02-28
IRIS.ai is an AI researcher that discovers and comprehends published scientific literature and explores its interdisciplinary knowledge connections. Using a visual interface, the tool narrows matching selections to a specific reading list and finds relevant information faster. IRIS.ai can peruse openly available scientific content and be modified to access third-party material archived behind a paywall, such as specialist journal subscriptions.
IRIS.ai excels in the area of locating most relevant literature in the least amount of time. This has significant advantages for communities involved in academic and scientific research. The tool does this by reducing 90% of the time previously required to find correct literature sets, with these searches resulting in an 85% accuracy gain versus 70% using manual methods.
Using IRIS.ai will free up a researcher’s annual workload by 432 hours that can be devoted to more creative and value-added activities. Research audiences have an opportunity to discover new knowledge faster, easier and at a lower cost.
IRIS.ai excels in the area of locating most relevant literature in the least amount of time. This has significant advantages for communities involved in academic and scientific research. The tool does this by reducing 90% of the time previously required to find correct literature sets, with these searches resulting in an 85% accuracy gain versus 70% using manual methods.
Using IRIS.ai will free up a researcher’s annual workload by 432 hours that can be devoted to more creative and value-added activities. Research audiences have an opportunity to discover new knowledge faster, easier and at a lower cost.
Technical Feasibility outlined a roadmap to advance our technology from the current TRL7 to a commercially-viable level of TRL9. We translated the project's functional requirements into technical work, where the goal is to have IRIS.ai automatically extract knowledge from a selected document collections and extract intelligence previously unknown to the user. We evaluated potential technical risks and proposed mitigative solutions.
Commercial Feasibility defined our core customer groups and performed an in-depth market analysis to choose the most attractive R&D segment. We explored the areas of Chemical, Materials and Life science to determine the most suitable market for expanding our footprint in selected European markets would immediately benefit from our technology. Our market plan defines how we will reach our target audience. As part of our FTO/IPR analysis, we scrutinised a selection of patents to assess their level of similarity to our innovation.
Financial Feasibility was an exercise in considering several what-if scenarios with different levels of market penetration and speed of our product's acceptance. The key metrics under comparison were market share amongst different customer groups, variation in license fees over time, revenue streams and personnel expenses. We assessed the projects’ cumulative profits and the overall ROI. Based on the discussed assumptions and our research we selected the average-case scenario as the most realistic budget option.
Commercial Feasibility defined our core customer groups and performed an in-depth market analysis to choose the most attractive R&D segment. We explored the areas of Chemical, Materials and Life science to determine the most suitable market for expanding our footprint in selected European markets would immediately benefit from our technology. Our market plan defines how we will reach our target audience. As part of our FTO/IPR analysis, we scrutinised a selection of patents to assess their level of similarity to our innovation.
Financial Feasibility was an exercise in considering several what-if scenarios with different levels of market penetration and speed of our product's acceptance. The key metrics under comparison were market share amongst different customer groups, variation in license fees over time, revenue streams and personnel expenses. We assessed the projects’ cumulative profits and the overall ROI. Based on the discussed assumptions and our research we selected the average-case scenario as the most realistic budget option.
The most commonly used software-powered research assistant tools are nothing more than a search engine bolted on top of scientific publication databases. Instead of turning pages, the researches faces a digital literature discovery process that is tedious, time consuming, of uncertain accuracy and has limited interdisciplinary insight.
IRIS.ai introduces a quantum leap in R&D literature search because users are not restricted to best-guess keyword lists to describe their problem domain but can prime the AI assistant with just a single URL of a paper that is related to the search. The tool will parse the initial document and begin an intelligent search of the available corpus by segmenting it into relevant topics.
IRIS.ai intelligently formulates a hypothesis from a collated literature set. The embedded AI processes perform autonomous scientific documents search, drawing conclusions from the study and reporting back to the requesting researcher.
IRIS.ai introduces a quantum leap in R&D literature search because users are not restricted to best-guess keyword lists to describe their problem domain but can prime the AI assistant with just a single URL of a paper that is related to the search. The tool will parse the initial document and begin an intelligent search of the available corpus by segmenting it into relevant topics.
IRIS.ai intelligently formulates a hypothesis from a collated literature set. The embedded AI processes perform autonomous scientific documents search, drawing conclusions from the study and reporting back to the requesting researcher.