AI Summit Silicon Valley 2021: Top Highlights & Insights from AI Experts

To discover how to take your AI project from proof-of-concept to production, the AI Summit Silicon Valley 2021 highlighted industry best practises and guidelines to understand the pitfalls that many AI projects face and how to avoid them. The titans of tech, R&D experts and C-Suite visionaries from the industry suggested game changing initiatives and insights on how early adopters are now scaling AI and ML to solve business challenges and innovate faster.

After a hiatus of almost two years, the AI Summit Silicon Valley 2021 brought forth the chance to listen to AI experts in the epicenter of technological advancement with an extensive two-day aspirational and pragmatic content program.

The event was formally launched by Lisa Gillmor, the mayor of Santa Clara. In her keynote, Gillmor said it had been over two years since she addressed an in-person convention, emphasizing the importance of technology in the community.

On the first day of the conference, Microsoft and IBM Watson specialists shared insights into their work on AI. According to Mitra Azizirad, corporate VP for Microsoft AI and innovation, “business leaders in AI must rely on their teams’ resources rather than technology alone.”

Azizirad said that combining people’s brilliance with technology will “truly change the world.” “Adapting and changing is about so much more of tech – it’s the combination of human and machine that will help organizations both reimagine and transform their businesses,” she said, adding, “human ingenuity with AI is truly a force multiplier.”

During the COVID-19 crisis, 61% of high-performing corporations raised their investment in AI, according to a McKinsey analysis, she said. “This underscores just how integral AI capabilities have become in terms of how work gets done.”

“Even before the pandemic, my team and I were working with many customers around the best ways to inculcate an AI-ready culture in their organizations.”

See More: Top Security Considerations for Transitioning from Private to Public Cloud

Let’s look at the top key highlights of AI Summit Silicon Valley 2021!

Revolutionizing the ‘built environment’ with Tech

Without addressing the problems we see in our built environment, we will not get to our 2030 or 2050 sustainability goals, stated Sridhar Deivasigamani, clean energy technologist and founder & CEO of Intellihot Inc. Speaking during a session on ‘Technology as solutions for the build environment problems’, he said, “About 54% of people live in cities today and by 2050 almost 70% of us will live closer to cities. Cities are filled with buildings, and how energy is consumed in these buildings is not actually well understood.”

Some buildings store hundreds and thousands of gallons of water and keep it hot 24×7. The equipment used is kept running all the time and consumes energy, thus polluting the environment. Deivasigamani, while summarising his concern, said, “What is happening in our built environment is – we have more equipment that runs 24×7 and people have no idea when it is going to fail. When it fails, people replace it.”

He highlighted two points for a sustainable future. The first and foremost is to know how much energy is consumed by us. Secondly, we must know when is the right time to change the equipment. However, he said, none of the businesses would want to change their equipment and spend a lot of money.

Integrating automation in the financial sector

During a session on AI infused automation in the financial sector, Heran Shah, industry CTO of financial services at IBM Automation, said that the pandemic accelerated a default-is-digital requirement. Business processes that were not digitized struggled or halted. “Companies that have not moved beyond a task-driven portfolio of automation initiatives have reported a plateau, decline, or failure of business goals.”

“With AI, things changed. Within the financial sector, investment is increasing in AI infused automation enabling technologies. Covid-19 priorities have process automation as the top focus area for banks, an outcome of the failure of manual legacy systems. Financial institutions understand the need to automate business processes using AI, however they are still searching for use cases with significant uplift of value.”

The most used AI technologies in banking are:

• • RPA:36%
  • Virtual Assistants: 32%
  • Machine learning: 25%

Today, what we are doing is very specific. But what is the next best task, he asked. The hot trends Shah outlined are AI: ML & deep learning, APIs, chatbots, microservice architectures, robotic process automation, and so on. The next trend or technologies on the radar are Cognitive RPA, AR, event-driven architecture, low-code development, and progressive web pages. He said that edge computing, IoT, quantum computing, VR, and sophisticated chatbots are still the hype trends.

See More: Top 5 AWS Misconfigurations That Led to Data Leaks in 2021

AI & ML: Game changers for data center operators

In another session on ‘How AI and machine learning are changing the game for data center operators,’ Tracy Collins, VP of sales at EkkosSense, talked about the current state of optimization. Most data centers, Collins said, see only the cooling unit temperatures. AI & ML helps you understand the data. “AI powered software with machine learning shows operators ‘the why’ and provides instruction sets on how to tune the data center.”

“Companies can spend huge amounts and install sensors in every rack. But it will just tell you what’s happening in the rack. However, an AI based system will measure everything, right from the cooling unit to the power that is being produced. It will eventually correlate all that data real-time and present that to you in an actionable way.”

The question is, why do you need AI and ML to do that? It’s not just the data points but also understanding the relationship between those points, he said. 

“AI and ML are constantly examining the data and learning how it is acting. Through the learning, they are able to determine what organizations can do about it. The operators that bring in AI will learn what is happening in the data center and get a solution for a more balanced state,” added Collins.

IoT, 5G, and business growth

Michael Weller, innovation advisor at Verizon, spoke about tapping into the power of connectivity for IoT and navigating 5G. According to Weller, getting to Industry 4.0 means “overcoming current challenges like coverage, cost, implementation, complexity, scarcity, and security.” 

Weller shared key insights on what 5G will mean to businesses. From a physical perspective, all new technologies that are expected to write on networks viz AI, ML, computer vision, AGVs, AMRs, digital twins, and condition-based maintenance will consume massive amounts of bandwidth. With that, it would be “difficult for existing network infrastructure to keep up.”

The opportunity to use the flexibility of a cellular network to locate and connect to resources will be very attractive to organizations, said Weller. “When it comes to the virtual aspect, network slicing is going to take virtual networks within a private 5G environment and enable end-to-end connections for certain applications and services, essentially creating logical networks within networks.”

“Lastly, the edge computing and importance of the visual tools and methodologies that are coming up, will be a critical part,” he said.

MLOps redefining operations

Enterprises are dramatically accelerating AI investments. About 86% of organizations have increased their ML budgets for 2021, and there is a 76% YoY increase in the average number of data scientists employed by organizations. Putting a thorough perception on MLOps, Atalia Horenshatien, customer-facing data scientist at DataRobot, briefed on the topic ‘Operating at scale with MLOps,’ 

“Successful ML projects must integrate with IT systems. The organization’s competitive advantage comes from how quickly it can deploy and iterate on models in production. But, nothing gets to production without MLOps. When we think about our RoI, we must involve AI and think about MLOps, which will take a vital role as part of the process,” the data scientist believes.

Amid the iteration speed, training, composability, diversity, and lack of reusability challenges, organizations need a single platform to deploy, operate, and govern and secure the models. “To overcome these challenges, the platform should monitor everything everywhere. Secondly, it must inculcate superior ML health. And lastly, the model should involve governance and trust, which is the most important factor. The platform should manage the biases on every deployed model.”

All these components will give organizations trust in their models, move them as fast as possible into production, control and manage them, and get full visibility.

What were your top takeaways from the recently-concluded AI Summit? Let us know on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We’d love to hear from you!