Legal Knowledge Management and the Rise of Artificial Intelligence
Robotic Refactoring the Workplace
Why Your Next Insurance Claims Processor Could be a Robot
The Forgotten Element in Your Big Data Strategy
WiFi Networks: Shifting from Providing a Service to Improving the...
Daniel J. Strojny, Interim Associate Director of Network and IT Operations, University of St. Thomas
Operationalize Machine Learning
Zongjie Diao, Director of Product Strategy and Management, Data Center Compute Group, CISCO
Thank you for Subscribing to CIO Applications Weekly Brief
"AI -The Future of Automotive Industry"
By Nitin Sethi, Global IT Director - Business Transformation & Engagement, Visteon Corporation
In spite of much recent excitement about AI, and the unlimited potential it carries, there are still some tactical challenges which need to be addressed, and can potentially become fodder for newer technology innovations. Machine learning / AI requires large amounts of human effort to label the training data necessary for supervised learning. Technology around In-stream supervision, in which data can be marked in the course of natural usage is still at the nascent stage. The challenge of “decision transparency,” or showing which factors led to a decision or prediction, and how, continues to be intriguing. It creates implications on AI applications where trust matters, and in some cases, predictions carry societal implications. Another challenge is that of generalized learning techniques, since AI techniques continue to have difficulties in moving their learnings from one set of data to another. Transfer learning, in which an AI model is trained to accomplish a specific task and then quickly applies that learning to a similar but distinct activity, is potential technology advancement. The progress made till date in AI addresses particular problems/activities which are predictable. The real value of AI will be to tackle the problems, taking into account past success and failure experiences and future predictability based on historical data inputs.
2. Is there a sort of mental checklist or a methodology that you follow to pick out the best possible vendor or partnership to develop products?
To select the possible vendor or developing a partnership to initiate the AI Initiative, following factors needs to be taken into account
1) The business problem which you plan to address through AI, there is a reasonable probability that you may find a vendor who has already has an AI solution which can be tailored to meet the need. Ensure that you do some POC before you select an off the shelf product. There is “Lot of Smoke and Mirrors”
2) It’s essential that vendor selected has the right mix of business domain knowledge and technology know-how.
3) AI is not an overnight phenomenon; it requires a continuous nurturing and training of data, before it evolves into a mature solution to address a business problem, hence it’s essential to calculate ROI for developing in house solution vs. engaging vendor to develop for you.
AI is a journey you embark on, one that demands persistence to obtain the desired results
Also being among the first customers of a startup, or partnering with an established player for a specific business case, may allow you to influence how a product develops.
4) Start Small – Always plan for a pilot, this will help you deploy initial solution faster and enable it to test waters, so that you can see the results, tweak them if necessary to move forward at bigger scale.
3. Is there any particular industry trend you find most exciting or promising to become a game changer of the future? If so, where do you see the industry going?
Till date, significant progress made in AI involves around improving productivity and reducing costs by adding cognitive in or around Intelligent Automation, Natural Language Processing (NLP) and adding value by improving on traditional analytics techniques. Most of this work has been done locally by companies by developing their applications or using products available in the market with a varied level of maturity. The future trend and game-changer of AI applications will depend on deployment diversity and interoperability.
I potentially see a new business model around providing global/regional solutions and compute as service. For example, in the health care industry, there is potential to develop medical databases as an open neural network as a service. This data will comprises of all symptom, case studies of treatments, and responses around the world, which then can be trained by reusing multiple neural network models for pre-diagnostics of many diseases. This database in conjunction with wearable medical / fitness devices and powerful features of pattern recognition which AI has to offer can be applied to oversee early onset of many diseases, helping doctors and other caregivers to better monitor and detect potential life-threatening episodes earlier and at more treatable stages, saving tons of lives and millions of health care dollars. With the shortage of doctors in rural and remote areas, AI has the potential to combine machine learning and systems neurology to develop algorithms that can mimic the human brain. There is already uberization of this model, which I see potentially see a trend in Industry going forward. Beside this, rather than AI serving simulated thinking, the future trend will be to have AI Systems that think and have a conscious mind like humans. General purpose intelligent algorithms that can be applied to any problem.
4. Can you speak to us about a few of your current project initiatives?
A few exciting areas that we are looking into to harvest the power of AI are adding cognitive capabilities in the automation of operational workflows. As a part of the industry 4.0 initiative, we are also looking AI to be used in improving business performance in areas including predictive maintenance, where deep learning’s ability to analyze large amounts of high-dimensional data from audio and images can effectively detect anomalies in factory assembly lines. We are also looking to use AI technologies in infrastructure monitoring.
5. Based on your background and experience in the field of AI, what would be a singular piece of advice that you would like to give toa fellow peer, who is just starting out in this field?
First and foremost for putting AI in your organization is to develop an AI strategy with clearly defined goals, finding talent or right SI partner with the appropriate skill sets, and right domain expertise and most important to have strong commitment and ownership from business top leadership.
According to me, AI is not a shining object or a magic window that can solve everything tonight. It does not automatically solve everything in one go. It is a journey that you embark on, and it demands patience and the right amount of data to be trained. I will highly encourage to have a strong foundation and investment in data, through implementations of Digital Strategies, supported by Big Data and Analytics before venturing into AI.