5 Key Lessons Learned from 2018 AI Summit, NYC
2018 AI Summit, NYC, https://theaisummit.com/newyork/

5 Key Lessons Learned from 2018 AI Summit, NYC

Earlier this week, I had the great pleasure of attending 2018 AI Summit (@Business_AI #AISummit) at Jacob Javits Center in the New York City. It was an excellent opportunity to meet leading AI researchers, thought leaders, and fellow practitioners.


The AI Summit:

Now in its 3rd year, AI Summit is one of the most prestigious conferences focusing on enterprise AI applications. With 3,000+ attendees, 300+ speakers and 150+ sessions, the summit had something for everyone. All the usual suspects were in attendance, including Google, Amazon, Microsoft, and of course, Sophia the Robot (@RealSophiaRobot), Hanson Robotic's AI initiative. Even Financial Institutions were well represented by Goldman Sachs, J.P.Morgan, Citi, Capital One, Bank of America, and Barclays. The sessions were a good mix of case studies, business insights, and technical solutions from every industry sector. Here are 5 key takeaways from my experience at the summit.


#1: We are beyond the hype

A couple of years back, AI was offered as a solution to every problem in the world. It was supposed to make the world a 'better' place so that we all can live happily ever after! This year, we heard about stories of AI failing miserably including the sombre story of a self-driving car accident in Tempe, AZ and hilarity of Amazon Alexa starting her own party in Munich, Germany.

Another reality check is the realization that AI is only as good as the data that it can learn from. Speaker after speaker pointed out that 'AI Strategy' starts with 'Data Strategy' and we need to climb the 'AI pyramid' in the right order. This back-to-basics pivot showed that we are at the inflection point and ready to move beyond the early AI hype.


#2: 'AI-as-a-Service' is an exciting proposition

Democratization of AI was one of the most prominent themes of the conference. Gone are the days when you need a PhD to build a Machine Learning algorithm. With emergence of cloud service providers, powerful algorithms are just a simple API call away. These solutions significantly lower the entry barrier for a much larger business community.

"AI is all about automating the ineffable" - Cassie Kozyrkov, Google

Cassie Kozyrkov (@quaesita) delivered an excellent keynote on how Google is democratizing AI through cloud computing and developing Decision Intelligence discipline. Other cloud service providers such as AWS and Microsoft Azure showcased powerful solutions for taking advantage of off-the-shelf AI capabilities. Furthermore, Ben Goertzel (@bengoertzel) from SingularityNET championed 'Decentralized AI' for infusing AI and blockchain together to create an open market. It is an interesting proposal for ensuring democratization of future AI development.


#3: Explainable AI is a real thing:

Deep Learning has reputation of being a 'black box' methodology but model interpretability is an active research field. Several success stories for building 'glass box' models were shared for improving transparency and interpretability. Here are few interesting projects in the Explainable AI domain:

  • Reversed Time Attention Model (RETAIN): The RETAIN model was developed to help doctors understand why a model predicts patients at risk of heart failure.
  • Local Interpretable Model-Agnostic Explanations (LIME): A post-hoc model that provides an explanation of a decision after it has been made.
  • Layer-wise Relevance Propagation (LRP): This is particularly useful for computer vision for identifying important pixels by running a backward pass in the neural network.


#4: Financial Institutions are ripe for disruption

The summit has a separate stream for 'AI in Finance' where speakers from global banks and fintechs discussed the latest trends in Finance. The key AI use cases include anomaly detection for fraud/compliance/cyber-defense, expert systems for robo-advisors and chat-bots/virtual agents for customer experience.

"When we think about AI/ML, its about removing the friction to give customers seamless experience" - Gavin Michael, Citi

Especially both Amazon Connect and Google Contact Center AI recently announced contact center AI solutions that promise better efficiency and lower operating costs. Experts agreed that manual back-office bank operations would likely be the target of AI investment in the near future.


#5: Responsible AI is imperative

Last July, the two biggest personalities in Technology got into a very public dispute over the future of AI. Elon Musk believes that AI will be the downfall of human race while Mark Zuckerburg thinks that AI will actually help innovate even further. The two fought over social media, trading petty remarks, but essentially both represent the opposite sides of the AI: Apocalypse vs. Nirvana. Regardless the business community has realized that general or 'human-like' intelligence is much farther away and focused its efforts on domain-specific impactful AI applications with tremendous job growth potential.

"We're seeing that AI is helping people at their jobs... as well as opening up new jobs" - Beth Smith, IBM Watson

Especially, the 'Responsible AI' theme transcended through every talk. Several speakers talked about integrating 'trust' as one of the engineering principles and embedding ethics and values as a part of AI deliverables.

"We need to create responsible AI so we can bring trust to the technology" - David Carmona, Microsoft

IBM has created AIFairness360, an open-source toolkit to mitigate unwanted bias in the Machine Learning models and Adversarial Robustness Toolbox for analysis of attacks and defense methods for machine learning models. New York City recently passed the first bill to examine 'algorithmic bias' in government agencies. Responsible AI will likely be the focus of technological, sociological and political discussions for the foreseeable future.


Overall this was an amazing learning experience with great talks, demos, and panel discussion. I truly enjoyed interacting with fellow AI professionals and was able to forge valuable connections within the industry. Thank you AI Summit (@Business_AI #AISummit) for putting together a phenomenal conference!

要查看或添加评论,请登录

Vishal Morde的更多文章

社区洞察

其他会员也浏览了