发帖
雷达卡
In this wide-ranging interview, we discuss the role of IBM global chief data officer, 4 key ideas of cognitive computing, risks of AI, IBM Data Science Experience, healthcare, basketball, sports analytics, and more.
By Gregory Piatetsky, KDnuggets.I first met Inderpal Bhandari over 20 years ago at a KDD conference, when he was an IBM researcher working on Advanced Scout - probably the first program that applied (very successfully) data mining to basketball and was used by several NBA teams. We recently reconnected after he rejoined IBM as a Global (!) Chief Data Officer, and despite being very busy he was kind enough to answer my questions in this wide ranging interview.
Inderpal Bhandari is Global Chief Data Officer of IBM. In this newly created role, he will leverage his extensive experience to lead the development of IBM's data strategy. Inderpal brings to IBM more than 20 years of experience in leadership roles. He served as Senior VP and Chief Data Officer of Cambia Health Solutions where he led the development of Cambia's data strategy and drove the transformation of the company's consumer experience strategies.
Prior to that, Inderpal served as Vice President of knowledge solutions and Chief Data Officer at Express Scripts/Medco Health Solutions, where he was responsible for maximizing the utility of the company's data and its readiness to respond to emerging market trends. Prior to that, he founded Virtual Gold, growing the company into an international market leader for analytics in call centers and professional sports.
Inderpal is an expert in transforming data into business value and improved customer experiences by delivering strategic, innovative capabilities that use analytic insights to enable growth and productivity. He has been featured as an industry expert by Wall Street Journal, Washington Post, US News & World Report, CNN and FOX.
Inderpal earned MS from the U. of Massachusetts and PhD from CMU.
Gregory Piatetsky: Q1. What do you do as IBM Global Chief Data Officer?
Inderpal Bhandari: This is actually my fourth go-round as Chief Data Officer (CDO). There's been a common theme of creating value from data in the way I approach the CDO role. At Express Scripts/Medco Health Solutions, I was responsible for maximizing the utility of the company's data and its readiness to respond to emerging market trends. As CDO of Cambia Health Solutions, I led the development of Cambia's data strategy and drove the transformation of the company's consumer experience strategies. Here at IBM, we are starting with the company's monetization strategy - how the enterprise derives revenue and where it is placing its strategic bets for the future. For us, this is on becoming a cognitive business.My focus starts with establishing the enterprise wide data strategy. That data strategy then enables our monetization strategy - the internal transformation to become a cognitive business.
Here is my keynote at IBM Chief Data Officer Summit.
I'm focused on five pillars:
- developing the clear data strategy,
- executing enterprise-wide governance and management systems,
- becoming the central and trusted data source for IBM,
- building deep data and analytics partnerships, and
- developing and scaling our talent in this area.
Building a world-class team of data scientists and data engineers is key to our success. We're leveraging internal expertise, and recruiting top talent to join us. We're tackling some of the world's greatest challenges - it's an exciting time to be at IBM.
(Editor: see this job ad if you want to join IBM as a Data Scientist/Data Engineer)
GP: Q2. We discussed IBM Cognitive Computing and you mentioned 4 key ideas that guide its development - "expertise", "white box", "no programming", and "continuous learning". Can you elaborate and describe each idea ?
IB: Yes - cognitive systems are known by four primary attributes. We like to refer to these as the "4 E's."
First, the cognitive system is anexpert in a particular domain. Second, cognitive systems support expression and human interaction.
Third, cognitive systems areeducated - they are trained by experts, not programmed.
And fourth, as they experiencenew information and scenarios, they evolve and learn. They get smarter as they go.
GP: Q3. Some people, including Elon Musk, warn of risks of Artificial Intelligence, while others like Andrew Ng say it is premature to worry about it (like worrying about overpopulation on Mars). Can cognitive computing lead to true Artificial Intelligence and should we worry about the risks?
IB: Our strategy addresses this actually. It is implicit in our approach. If you look back to our 4 E's, cognitive systems are trained to be experts and trained to interact naturally with humans to explain their reasoning, enabling the human to make the final decision.
Ours is a white box approach.
We don't diminish the importance of human analysis, experience, and judgment - but see cognitive systems and people working together to help close the gap between data quantity and data insight.
GP: Q4. IBM has many products of interest to KDnuggets readers, including IBM Watson and its many cognitive APIs, IBM SPSS Modeler, and recently introduced IBM Data Science Experience (datascience.ibm.com). Can you give an overview of IBM Big Data and data science related products and how they fit together?
IB: This is a key advantage IBM has to offer. We offer a complete portfolio of data, analytics, and Watson based solutions to fuel a cognitive business. These solutions enable organizations to engage with data to answer the toughest business questions, uncover patterns and pursue breakthrough ideas. And, we're able to bring these all together in an integrated way.
Within our IBM Watson business, we have many products and offerings that are really changing the way we do business and infusing intelligence into our decision-making. In the healthcare space, we have Watson for Oncology - helping doctors and healthcare professionals make more informed treatment decisions and deliver evidence-based treatment options. We also offer Watson Explorer to analyze both structured and unstructured content to present insights in a single view, Watson Discovery Advisor to look at disparate data sources to identify patterns and recommend solutions, and Watson Engagement Advisor to interact with customers and offer solutions. These are just a few examples.
Yes, the Data Science Experience, released in limited form earlier this month and to be generally available later this year,is the first cloud-based development environment for near real-time, high performance analytics. It's really set up to make the data scientist successful. There's the ability to access and ingest data and deliver insight-driven models to developers. Available on the IBM Cloud Bluemix platform, the Data Science Experience provides 250 curated data sets, open source tools and a collaborative workspace to help data scientists uncover and share meaningful insights with developers, making it easier to rapidly develop applications that are infused with intelligence.
提升卡
置顶卡
沉默卡
变色卡
抢沙发
千斤顶
显身卡
京公网安备 11010802022788号
