KNOWLEDGESTREAM AT-A-GLANCE
Developing New Research Methods in the Era of Big Data and AI
ABSTRACT
New Research Methods in the Era of Big Data and AI
PARTICIPANTS
OBJECTIVES
1. : Determine the challenges, drivers, trends and frictions specific to the scope.
2. : Develop a thorough list of alternatives, their strengths and weaknesses, supported by examples.
3. : Define the key characteristics that would need to be true to make the desired impact.
4. : Define the recommended path(s) for a viable solution and identify the critical success factors in deployment.
End Date: Oct 29, 2018
CONTRIBUTIONS
ACTIVITY
62 Days
3 Themes
19 Contributors
813 Posts
253 Comments
68 Followers
OUTPUTS
2 Slide Deck
2 Video
THEME #1
Integrating Big Data and Analytics into AI Research Methods
THEME SUMMARY
The integration of Big Data and AI into basic research faces many hurdles but success will bring global transformation in areas such as medicine, materials science, and climatology.
- Large amounts of data are needed to learn patterns, test algorithms, and advance research but smaller companies and groups face hurdles, including access and price of storage.
- AI can be leveraged for mining large amounts of Data, leading to breakthroughs in our most complex problem sets.
- Big Data is just that – data. Incorporating AI can help to make sense of the data by providing insight and meaning.
Big Data - AI Research Integration is Worth Considering
Integration of Big Data and AI would focus on the development of new knowledge that is an outcome of past learnings, and allow for new developments based on sound data and extrapolation rather than researcher or institutional bias.
AI can help convert Big Data in actionable insights that we can work with. It can help us learn from the data to predict behavior, weather patterns, incidence if disease, etc. AI makes big data valuable.
There are Limitations in AI That, When Overcome, Will Lead to Breakthroughs
Most of the startups to SMBs in AI Research space are facing data quantity problem, as they have exposure to limited data sets, compared to Tech. giants like Google, Microsoft, Facebook.
AI researchers to come to an understanding on what is and what is not okay to train algorithms to perform. Ethics in AI needs to be mutually agreed upon in research and especially for industry
THEME #2
Overcoming Data Needs to Achieve Innovation
THEME SUMMARY
Small businesses and academic groups are finding success in the development of research using Big Data and AI despite challenges in data access and computing requirements.
- Innovation doesn’t require deep pockets – especially in this still-early phase of development. Major costs can be overcome by collaboration (e.g., pooling resources and talent).
- As companies expand in the age of Big Data, more services will become available, including Data As A Service (DAaS), allowing research teams to access information needed to solve critical problems.
Gaining Access to Datasets
Find a collaborator. Startups tend to collaborate with university professors to be able to engage in in-depth research. This gives them access to the university infrastructure for their computing and storage needs
Small businesses should take advantage of the large public datasets to glean customer insights. There are various metrics which would include customer spending habits, frequency of purchase, etc. that can be utilized.
Small Group Success Examples
Figure Eight Inc. ( www.figure-eight.com ) provides a platform to transform unstructured data from the real world – text, images, audio, video – into high-quality large scale structured training datasets at enterprise scale to train, test, and tune machine learning algorithms.
Partnership on AI, a group comprised of IBM, Microsoft, Apple, EBAY, Sony and others. In all, there are more than 70 partners in 9 countries with over 50% or partners being non-profits. “Bringing together diverse, global voices to realize the promise of artificial intelligence.”
THEME #3
Implementing Big Data and AI into Medical Research