Interview with Jay Swartz, Chief Scientist, Blackbox AI - Speaker at Global AI Conf - Sep 2018 Posted on : Sep 14 - 2018

We feature speakers at Global Artificial Intelligence Conference - 2018 Sep 25 - 27 – Boston to catch up and find out what he or she is working on now and what's coming next. This week we're talking to Jay Swartz, Chief Scientist, Blackbox AI Topic - "AI, Blockchain & The Future Of Work"

Interview with Jay Swartz

1. Tell us about yourself and your background.

Currently the Chief Scientist at BlackboxAI and advisor to several startups, Mr. Swartz is a Data Scientist who enables companies to leverage data science (DS) and machine learning (ML) by helping them to establish and extend skilled teams. He has expertise using IBM Watson as well as other AI & blockchain technologies, enriched by over 40 years of experience in IT, software engineering, research, marketing and consulting with Apple Computer, University of Colorado, IBM, Seagate Technology, Sun Microsystems (Oracle), Welltok and start-ups. He helps identify revenue and cost reduction opportunities for ML implementations. Mr. Swartz leverages a rich, multi-disciplinary perspective based on his work with ML, virtual reality, information systems and web/mobile solutions. He is an active advisor to startups on best practice for deploying ML based solutions and is a Quora Top Writer (Jacque Swartz). He shares his experience with the larger community, regularly presenting to DSI students at Galvanize as well at MeetUps and industry conferences.

Mr. Swartz has extensive experience building tools that generate revenue, increase margin and improve client satisfaction. He combines technical expertise and customer requirements identification skills to define and implement world-class products and tools that delight clients.

2. What have you been working on recently?
I have been building specifications and SOWs for several AI solutions; facial recognition for detecting loyal clients entering an establishment and two variations of a resource optimization system matching resources to projects.

3. Tell me about the right tool you used recently to solve customer problem?
I identified an OpenSource alternative for a reverse auction solution that a customer asked to have built. It eliminated a revenue project, but was the right thing to do for our customer.

4. Where are we now today in terms of the state of artificial intelligence, and where do you think we’ll go over the next five years?
AI is at a stage equivalent to hot air balloons for air travel. We don’t yet have anything that is truly intelligent. The current technology is very powerful for detecting patterns at the surface level of a problem, but has limited semantic ability. DNN variations are improving rapidly. In particular, GANs and transfer learning techniques are lowering the bar for large training sets. Many aspects of ML models are improving, such as the recent comprehension of ELUs for the logistic function in NNs. Just four years ago, discerning cats from dogs was considered a challenge. Today, cats, dogs and anything else you define can be concurrently detected in live video streams. Five more years of improvement, defined here not as core discovery but simply as implementation at scale of capabilities already identified in research publications, will see the advent of more powerful intelligent agents that assist users in sophisticated ways.

Today’s appointment schedulers will mature to manage a wider scope of life events, suggesting when and where to go to improve your chances of achieving increasingly abstract goals. For example, you will give your assistant the objective to make you into a web developer, and it will plan out your education schedule, suggest projects to complete, find competitions that will advance your skills and send you to restaurants frequented by hiring managers in the industry where you have the best matched skills.

5. There is a negative perception around AI and even some leading technology folks have come out against it or saying that it’s actually potentially harmful to society. Where are you coming down on those discussions? How do you explain this in a way that maybe has a more positive beneficial impact for society?

This is a matter of time horizons. For the next decade or two, AI will be a boon to humanity that eliminates the onerous, boring, and dangerous tasks from many jobs, making work more rewarding. Over time, there will be considerable pressure on less skilled individuals where most of the tasks that they perform are automated, and this will require action on the part of society to resolve. It will almost certainly take two or more decades for politicians to find a reasonable solution. Those who believe AI to be beneficial are considering a shorter time horizon than those raising alarms.

6. When you’re hiring, what types of people are you hiring? The job market for traditional programmers, engineers is very difficult to get into AI space. Are you hiring from that talent pool or is that a different talent pool? In terms of talent, how do you go about ensuring you get the best AI people at your company?

Blackbox has developed an ecosystem model where individuals join our community and express interest in participating in both internal and client projects. Participants for projects are selected based on best fit with the project requirements. We are constantly seeking out new talent who have proven skills in engineering, data science and experience in technical settings. Our ecosystem model is designed to optimize individual’s effectiveness and desires for work.

7. Will progress in AI and robotics take away the majority of jobs currently done by humans? Which jobs are most at risk?
Bear in mind, that AI shines given a task, but can only perform the most menial of jobs. AI and robotics are still at a stage where only relatively narrow tasks can be automated. Job descriptions will be modified at an accelerating rate, removing tasks that have been automated with AI, robotics, and other technologies. Tasks that lend themselves to automation have one or more of the following features; high volume of data, follow a pattern, have low risk for wrong responses, require limited judgement to perform, or are highly repetitive.

8. What can AI systems do now?
AI systems can detect patterns in complex data sets far better than humans. For example, when trained with desired patterns, it can detect anomalies, such as spotting a shoplifter taking products from a shelf and putting them into a pocket when the trained pattern is to take the product to the checkout area. Image recognition in general is very powerful.
NLP technologies have transformed the NLP community. Once state-of-the art language tagging methodologies have been made largely obsolete with the maturation of RNNs and attention networks. The rapid growth of corpora is enabling translation into more languages.The important thought here is again the idea of tasks being powered by specialized AIs. Radiologists are no longer better than AI at identifying breast cancer, but these AIs were not trained to detect other conditions. Over time we will see AIs tailored to the specific patterns of other conditions.

9. When will AI systems become more intelligent than people?

The short answer is that no one knows. We don’t yet know how humans manifest intelligence, so building an equivalent AI is impossible to understand. AI and CCN (computational cognitive neuroscience) research are both uncovering the mechanics of intelligence. AI today is a misnomer as today’s implementations are machine learning models with no true intelligence, just exceptional pattern detection. Research is making inroads to enabling AI to include comprehension of semantics, emergent features and many other areas. DNNs still have many aspects that have not been fully explored. New architectures and metadata elements remain to be discovered and optimized.

10. You’ve already hired Y number of people approximately. What would be your pitch to folks out there to join your Organization? Why does your organization matter in the world?

Blackbox AI is based on a collaborative model where the resources in our ecosystem are not tied to directly job or role descriptions. Our environment enables collaborators to follow their passion and explore new techniques. This expands their capabilities while adding value to the organization. Collaborators work on discrete projects and can participate on teams irrespective of their formal education or work history. Blackbox AI is about client deliverable results and teamwork.
Our long term vision is no less than enabling the future of work through intelligent smart contract infrastructure. We are building the components required to empower individuals to focus on the tasks they are good at and interested in, and have others support them in the tasks they are not interested in performing. Where you live and when you work will be at your discretion, not HQ’s.

11. What are some of the best takeaways that the attendees can have from your talk?

Understanding where AI and blockchain can add value, as well as where it is too early to be applied. The talk identifies the arc of these technologies and how they contribute to the long term vision for the future of work.

12. What are the top 5 AI Use cases in enterprises?

Image recognition, anomaly detection, NLP, process automation and translation.

13. Which company do you think is winning the global AI race?

Google is in the lead and there are many others. Due to AI still being confined to narrow use cases, many companies are pursuing solutions tied to their specific markets.

14 Any closing remarks

Looking forward to the event!