I have spent my professional career crawling around large IT systems as a university professor and software entrepreneur. For the last twenty years, I have been an independent industry analyst in Business Intelligence, which is now focused exclusively on Business Analytics. In 2018, I recently completed the 5-course Coursera specialization in Deep Learning, taught by Andrew Ng of Stanford.
The a-ha moment of this DL specialization was the discussions of analytics exceeding human-level performance and realized the profound implications. We humans have created tools smart enough to out-smart ourselves! I wrote my reflection in this article in Towards Data Science, of which the latter half is most relevant…
- Deep Issues Lurking Within Deep Learning (link, pdf)
Taking the social and ethical issues to heart, I started to think how IT managers should approach analytical systems enabled by deep learning. The issues just exploded in my mind. I concluded… Most IT professionals are unprepared to manage properly deep learning for business applications (except in isolated use cases). In the long term, the challenge is managing super-intelligence systems at scale!
These insights prompted an article series about
- How Managers Should Prepare for Deep Learning (link, pdf)
The objective is making Deep Learning relevant and useful to the managers who fund analytical systems, who evaluate their performance, and who are held accountable for their impacts.
The first two articles are published in Towards Data Science, plus the third is in draft (as shown below).
BizSmartAnalytics is presenting the following talks:
- DVEM 2018, Are You Prepared for Deep Learning, Genesee CO, 2018-10-03 (pdf)
- Teradata Analytics Universe, How is Deep Learning Valuable for Your Company, Las Vegas NV, 2018-10-17 (pdf)
BizSmartAnalytics is offering the following services:
- Patreon BizSmartAnalytics which mentors peer-groups of managers involved with DL-enabled systems, along with providing one-on-one advice to executives. Richard organizes and mentors peer groups, which meet online for an hour, twice monthly, plus one hour of preparation. Group size is limited to 5 to 7 persons to facilitate interaction. Group composition is spread across business functions and even industries to maximize a diversity of perspectives. Richard shares the latest concepts, research, and trends occurring in rapidly-evolving DL-enabled analytics. By stressing mentoring-the-mentors, group members are equipped to mentor colleagues within their organizations. The post Guidelines for Peer Groups has more details.
- Custom 1-3 day hands-on workshops for training IT managers about DL-enabled systems. Fees vary depending on duration, class size and travel. Customization to specific use cases relevant to your company is an option. Client supplies the training facility. Ask for a free proposal.
BizSmartAnalytics is conducting interviews with managers who are involved with enterprise analytical systems using neural network technology. The interviews focus the emerging managerial issues and challenges facing managers at all levels. Interviewees do not need in-depth technical knowledge. The interviews are informal and unstructured. Interviews can optionally be conduct as anonymous. These questions are to stimulate discussion:
- Describe your management involvement with analytical systems and deep learning.
- What is the motivation for using neural network technology? Tactical or strategic or mixture?
- What are the issues that you are encountering or expect to encounter? Any surprises?
- Are these issues unique to deep learning? Are the issues technical or managerial?
- Do you have suggestions for best practices or lessons learned for similar professionals?
Past Studies and Articles
A list of these studies and articles are here.