Over 1,000 people attended the recent VB Transform Conference, which spanned two days in the heart of San Francisco. Executives from Google, Amazon, and Facebook were among those in attendance. With the majority of the guests having ‒ at the least ‒ a base level understanding of AI, I was definitely in the minority as a complete novice. Orchestrated by VentureBeat, the event covered topics such as computer vision, intelligent robotic process automation (RPA) and implementing AI. Sound complicated? Definitely at first, but I left significantly less confused than I had been when I arrived.
Ople was one of the sponsors and had a booth at the ongoing expo, which featured both top-tier AI companies and emerging startups inside Hilton’s Grand Ballroom ‒ a massive space with an adjoining kitchen and high chandeliers. We were able to speak with interested attendees and demo the product, as well as witness the latest in AI technology before hearing from experts during panel sessions. I tested out a virtual reality experience in one booth, putting on a heavy headpiece and seemingly taking an elevator to the top floor of a building before walking off and plummeting down. Other AI companies offered shirts, doughnuts, pins, stickers and a plethora of business cards. Undoubtedly the busiest part of the venue, the bar and self-serving breakfast and lunch were likely major attractions as well.
With the help of the VentureBeat Transform app created specifically for the event, I was able to read up on what topics the sessions would be covering and who would be speaking. Therefore, I wasn’t walking in completely unprepared. The sessions varied from one to three speakers with a moderator, but my favorites were the ones with multiple speakers from different companies. Data science lingo and acronyms were ever-present but it was interesting to see how various companies and industries were utilizing the same type of technology. I made my way up and down the three-story venue and covered a good variety of sessions in two days’ time, realizing little by little that AI was far from the impossible enigma I thought it was only a few days prior.
Women in AI
With AI becoming increasingly trendy, the discussion around diversity was a major topic. A recent New York University study found that over 80 percent of AI professors are men. In addition to specific panels addressing this issue, a Women in AI Breakfast kickstarted the conference and had over 200 women in attendance. Franziska Bell, Director of Data Science at Uber; Jessica Lachs, Head of Analytics at Doordash; and Charina Chou, Global Policy Lead for Emerging Technologies at Google were some of the female speakers.
I was especially impressed by Stephanie Rogers, Senior Product Manager at Pinterest, who spoke at a session about visual search in e-commerce. Since I am familiar with Pinterest, I found the discussion around the company’s implementation of AI to be surprisingly simple to grasp.
Rogers delved into the process of embedding, which allows similar pins to be clustered together ‑ thereby generating related pins that a user would likely be interested in given their past activity on the platform. She also talked about using AI to accurately match pins with their respective retail source. For example, letting a user purchase a dress or a pair of shoes by simply clicking on the pinned item via Pinterest.
At the end of the two-day conference, awards were given specifically to women in AI to acknowledge diversity in the field, across the following categories: responsibility and ethics of AI, AI entrepreneur, AI research, AI mentorship, rising star.
Before breaking off into topic-specific sessions, each day began with a fireside chat with select industry professionals. These appeared to draw the majority of attendees. Vanja Josifovski, CTO of Airbnb, was one of the first speakers and stressed the democratization of AI and the importance of open-source technology. Like many, my first thoughts on “artificial intelligence” were robots or something top secret that would not reach the general public for years to come. However, this initially-niche capability is projected to become the most efficient method of operating ‒ and most likely the only way to stay in business in the future.
AJ Abdallat, CEO of Beyond Limits, discussed the perception of AI and how automation was not a competitor to consumer autonomy. Another popular perception of AI is that it will replace human capability. However, Abdallat acknowledged that this mindset is incorrect: “Make the computer a transparent partner to the human.” AI is intended to function as a counterpart, not a substitute.
Andrew Moore, Head of Google Cloud AI, had a similar view, affirming that machines are not “better” than the human, but humans and machines are “better than anything.”
One of the first presentations was centered on building a best-practice AI platform, and was led by Chris Chapo, SVP of Data & Analytics at Gap Inc. In his company’s case, AI had to be strategically infused in a workplace with limited big data experience. This major retailer was founded in 1969 and grappled with bringing AI capability to the company versus buying AI support from another company. Chapo said the process required experimentation, understanding what drove engagement and analyzing customer purchases. Gap Inc. currently is building AI training courses for its employees, similar to LinkedIn’s Training Academy.
Chapo was also a part of a panel session on defining and enabling AI projects. He was joined by Deborah Leff, Global Leader and Industry CTO for Data Science and AI at IBM. Key points were the need for a user-friendly tool helps making smart data-driven decisions and a focus on improving existing customer experiences.
AI companies seem to be popping up everywhere and with greater frequency. The 2018 AI Index Annual Report showed that active AI startups in the U.S. have increased by 113 percent from 2015 to 2018. While the industry is getting competitive, AI companies’ biggest threat is the customer experiences in places that do not utilize AI. Take the always-updating iPhone; Apple is constantly competing with its previous model and how consumers experience it. In order to be successful, an AI product has to make the existing customer experience even better, causing the non-AI route to appear slow and outdated in comparison.
“Managers that use AI are going to replace managers that don’t,” said Chapo.
How to Make AI Easier to Implement in Business
Ople Founder and CEO, Pedro Alves, moderated one of the final sessions on Transform’s second day. He was joined by Alberto Lopez Toledo, CTO of Freshly; Mike Haley, Head of Machine Intelligence at Autodesk; and Al Brown, SVP of Engineering at Veritone. The panelists compared centralized and decentralized AI expertise in companies and considered which is better. The three seemed to agree that a hybrid approach of the two is most optimal, since it is about finding a balance between scale and specialty. In a decentralized workplace, the entire organization is expected to have some level of data science understanding and all employees must be operating efficiently, which is not possible with a shortage of data scientists around the world. However, there is no head honcho or authoritative data scientist calling the shots or a long line to move the project forward. With a centralized workplace, there are higher level, systemic issues. For example, AI specialists and data scientists are in control, and the rest of the organization is expected to oblige. Moreover, they are not close enough to the actual line of business to fully understand the business objective, often resulting with an AI model that does not get put into production.
With all the panelists sharing similar views, Alves emphasized that both “have a problem with communication,” thereby requiring a fusion of the two practices, and more importantly, a tool that accelerates this fusion, like Ople.
AI Innovation Awards
Alves and VentureBeat Founder Matt Marshall later ended the night and presented their first-ever AI Innovation Awards.
VentureBeat has been leading the movement to bring innovative companies, both small and large, together to advance the tech industry. We have partnered with them in different ways and the AI Innovation Awards was a unique opportunity for us to jump in because of the program’s purpose and its connection with Ople’s mission to make AI ubiquitous.
We want everyone in every business to harness the power of AI and get more from data. In order to achieve this, it is important to acknowledge and learn from others in the field to build a stronger community and better world. Our participation in the AI Innovation Awards is doing just that ‒ congratulating and bringing awareness to fellow AI innovators.
As Transform’s final event, five companies were awarded across different categories for their AI efforts: NLP/NLU innovation, business application innovation, computer vision innovation, AI for good and startup spotlight.
Following their respective rounds of applause, each winner gave a short speech. Guests were then free to indulge in dinner and their next round of drinks ‒ a celebratory conclusion to a fully-packed two days.
“Every time you hear AI, take a sip!” joked Stewart Rogers, VentureBeat analyst-at-large, onstage.