The decision by states to re-open their economies amid COVID-19 pandemic is raising concerns and gaining support with President Trump applauding the move.

The State of California announced plans to re-open with more states following suit. The same applies to sports and hotels, which are re-opening in phases.

The big question is: What trends will drive economic recovery as states¹ reopen?

In this article, I will explore how AI will drive economic recovery as reopening begins.

The demand for data scientists is increasing as organizations become data-driven and with the digital transformation² happening, practical regulations are needed to govern these changes.

The role of data scientists is changing and developing legislation will streamline the industry and facilitate the growth of the data science field. A survey by McKinsey in 2019 found data science roles in demand by corporate organizations, as they become smart enterprises.

These and more on our weekly AI update

5 AI Trends That Will Drive Economic Recovery As States Reopen

1. Biology As Technology

Computational biology and bioprinting are bringing forth a golden era of biology where health diagnostics and therapeutics, as well as food, fuel, and materials, will be rapidly advancing.

2. Human-Centered AI For Better Health Outcomes

The advancement of AI algorithms is enabling disease surveillance, telemedicine, virtual diagnostics, fever detection based on facial recognition, and vaccine and drug development.

3. Robotics, Drones, AR/VR

The issues related to supply chain failures will be addressed with the rise of robotic process automation in industrial warehouses and logistics. Robots³ are taking the pulse of patients and assisting in operating rooms in hospitals. Drones are providing security patrols and being tested for the delivery of goods. AR and #virtualreality applications are allowing for touchless experiences.

4. E-commerce For A Sustainable Planet

Poshmark-the AIpowered social discovery platforms have become entertaining, economical, and environmentally-friendly ways to tidy a wardrobe, make money and connect with others.

5. Privacy & Identity For The Surveillance Age

Technologies will be critical to ensure compliance with HIPAA, CCPA, GDPR, and other laws.

Data Scientist Sexiest Job of the 21st Century

AI and augmented intelligence will be an effective way to help humans perform better & be efficient at their tasks. It will assist people to complete tasks in an easier, quicker way, using less energy that can otherwise be spent doing another activity.

The tasks that are not easy to accomplish is being accomplished with AI, as it helps us to be better at completing tasks. Humans may be dependent upon the machines to help make tasks more efficient depending on the situation. The machines will need humans for repairs and ensuring they are operating properly.

#Artificialintelligence needs good clean data⁴ to work properly. The role of data scientists is increasingly needed. Professor @Thomas Davenport wrote an article on the topic with current Chief Data Scientist of the U.S., DJ Patel, called “Data Scientist Sexiest Job of the 21st Century”

The original role of data scientists was to work with data and analytics, and there needs to be a new term for the role of working with AI. The federal government should play a vital role in overall legislation and regulation in the US. But it is really up to the states and jurisdictions to put together practical laws and regulations based on how AI will be actually like most technology-related matters.

AI for Environmental Applications

The primary motivator revolves around either profitability or sometimes sustainability while exploring artificial intelligence (AI) to inform decision-making for retailers.

The concern about the sustainability of AI initiatives is nothing new. The use of AI for environmental applications could contribute up to $5.2 trillion to the global economy by 2030 according to @PwC.

However, the message has yet to hit home entirely when it comes to different industries connecting the notion of machine learning (ML) with their own eco goals.

Organizations are stuck in a conundrum to identify whether AI is profitable or sustainable which is a siloed presumption that’s not true.

The idea of AI driving sustainability ticks two boxes in retail.

– The impacts of the avoidance of waste through the expiration of products across a company’s supply chain and procurement activities.

-The adherence to seasonal trends, where products become redundant, and not bought during a certain period.

Global warming, companies’ carbon footprints, many C-level executives struggle to justify their reluctance to comply with ML when it comes to climate control, eco concerns. The proposition and the way they’re approaching organizations have changed slightly while the actual AI products or services haven’t.

The Future of AI and Machine-Learning Techniques

Turing Award winner and director of the @Montreal Institute for Learning Algorithms @Yoshua Bengio provided a glimpse into the future of AI and machine learning techniques during the @International Conference on Learning Representations (ICLR) 2020 this week which took place virtually.

He also spoke alongside fellow Turing Award recipients @Geoffrey Hinton and @Yann LeCun in February at the @AAAI Conference on Artificial Intelligence 2020 in New York. But Bengio expounded upon some of his earlier themes in a lecture published Monday.

One of those was attention — the mechanism by which a person (or algorithm) focuses on a single element or a few elements at a time in this context.

Like @Google’s Transformer, it is central both to machine learning model architectures and to the bottleneck neuro scientific theory of consciousness, so information is distilled down in the brain to only its salient bits.

As like natural language processing, models⁵ with attention have already achieved state-of-the-art results and they could form the foundation of enterprise AI that assists employees in a range of cognitively demanding tasks.

Bengio is confident that the interplay between biological and AI research will unlock the key to machines that can reason like humans.

Applicant Screening by AI and Machine-Learning

To generate screening questions for active job postings, @LinkedIn started using AI and #machinelearning. Co-authors describe Job 2 Questions, a model that helps recruiters quickly find applicants by reducing the need for manual screening in a paper published this week on the preprint server @Arxiv.org.

A LinkedIn study found that roughly 70% of manual phone screenings uncover missing basic applicant qualifications. Screening is a necessary evil, the timing of Job2Questions’ deployment is fortuitous.

Companies are adopting alternatives, with some showing a willingness to pilot AI and machine learning tools⁶ as the pandemic increasingly impacts traditional hiring processes.

Job2Questions generates a number of screening question candidates, given the content of a job posting, the researchers explained. It is designed to reduce the time recruiters spend asking questions they should already have answers.

The LinkedIn researchers had annotators label sentence-question pairs, which enabled the prediction of the templates from sentences to collect data to train the machine learning models underpinning Job2Questions. The researchers claim that only 18.67% of applicants who didn’t answer screening questions correctly were rated as a “good fit” by recruiters

Failure of AI and ML Systems

Tech giants including @Google, @Amazon, @Microsoft, @Uber, and @Tesla, have had their artificial intelligence (AI) and machine learning (ML) systems tricked, evaded, or unintentionally misled.

Most organizations’ leaders are largely unaware of their own risk when creating and using AI and ML technologies despite these high profile failures. AI/ML-specific insurance could be an emerging solution to cover such situations.

AI and ML systems are brittle and their failures can lead to real-world disasters as recent events have shown. AI failures⁷ revealed that ML systems can fail in two ways intentionally and unintentionally according to research.

Gartner issued a dire warning to executives in its report on attacking machine learning models: “Application leaders must anticipate and prepare to mitigate potential risks of data corruption, model theft, and adversarial samples.” But organizations are woefully underprepared.

As the head of security of one of the largest banks in the United States told us, “We want to protect client information used in ML models but we don’t know how to get there.”

AI and ML systems can help create large amounts of value for many organizations but the risks must be understood — and mitigated — before the technology is fully integrated.

AI is revolutionizing Customer Experience

The way companies interact with customers transforming rapidly because of AI. “The global AI agenda,” found that customer service is the most active department for AI deployment today according to a survey of 1,004 business leaders conducted by @MIT Technology Review Insights.

It will remain the leading area of AI use in companies by 2022 (say 73% of respondents), followed by sales and marketing (59%), a part of the business that just a third of surveyed executives had tapped into as of 2019.

Companies have invested in customer service AI⁸ primarily to improve efficiency, by decreasing call processing and complaint resolution times in recent years. The leading organizations in the customer experience field have also looked toward AI to increase intimacy — to bring a deeper level of customer understanding, drive customization, and create personalized journeys.

@Genesys¹⁰, a software company works with thousands of organizations all over the world with solutions for contact centers, voice, chat and messaging. @Tony Bates, CEO Genesys, says, “The goal across each one of these 70 billion annual interactions. delight someone at the moment and create an end-to-end experience that makes all of us as individuals feel unique.”

OpenAI unveils Jukebox

Artificial intelligence is taking aim at the talents of songwriters. @OpenAI unveiled Jukebox this week, a #neuralnetwork that generates songs complete with lyrics and singing. 1.2 million songs along with lyrics and metadata used to feed Jukebox trained by #OpenAI.

OpenAI said in a blog post “The metadata includes artist, album genre, and year of the songs, along with common moods or playlist keywords associated with each song,”

Jukebox creates songs in the style of specific artists wandering into an audio version of the uncanny valley. A selection of song samples shared by OpenAI worth a listen. Jukebox’s Alan Jackson song has a distinctive uptempo country flavor, and its Katy Perry pop song is a torchy ballad with reverb-soaked drums.

You can almost hear Elvis’s songs sound hauntingly real. The researchers of OpenAI stepped in to co-write some lyrics. The Presley tune starts with “From dust we came with humble start/from dirt to lipid to the cell to heart.” That’s pretty trippy for an Elvis song.

AI still has some big limitations⁹ which are good news for human songwriters. OpenAI is riding on a big 2019 investment from @Microsoft and previously developed a system that beats humans at a video game, and trained a robot to solve the Rubik’s Cube.

PPP Lending AI

Google developed an AI solution called PPP Lending AI that integrates with existing document ingestion tools to help lenders expedite the processing of applications for the US @Small Business Administration’s (SBA) Paycheck Protection Program, which aims to keep workers employed during the coronavirus pandemic. It’s available to eligible lending institutions through June 30.

AI can automate the handling of volumes of loan applications by identifying patterns as Google explained in a whitepaper. PPP Lending AI can specifically classify and extract data in critical paperwork before reading documents for submission to the SBA.

Google says PPP Lending AI includes three parts solutions

-Loan Processing Portal, a web-based app that serves as a user interface and self-servicing center is the first part. End-users and loan applicants are allowed to create, submit, and view the status of their PPP loan.

-Document AI PPP Parser which allows lenders to use AI to extract structured information from loan documents submitted by the loan applicants is the second part.

-Loan Analytics is the third, let’s lenders onboard structured historical loan data, perform de-identification anonymization on sensitive information, store it securely with fine-grained data access control.

AI and ML assisting in adapting to post-pandemic changes

We are slowly moving into a post-pandemic world. The Covid19 pandemic affected everyone in the United States and is starting to open up. Many people have been working from home, adapting to the country’s social distancing protocols.

This pandemic has changed drastically the way that we work and view work. Companies will have a new set of organizational challenges in the new norm of post-pandemic uncertainty. Will the company’s culture change and adapt?

Many large companies (>10,000 employees) that needed to create an innovative culture across their organization have used @BetterUp for their cultural transformation initiatives in the last decade. BetterUp¹¹ uses AI and machine learning algorithms to match employees with the right coaches.

Recommendations of articles, podcasts, interactive exercises, and games are generated by another set of #algorithms to allow employees to improve their skills after the sessions with personalized coaches.

Gaurav, VP of Product, BetterUp says, “The algorithm is really tailoring the right recommendation to each individual based on what their area of interest is and based on their learning preferences. The algorithm is taking that into account and making those recommendations.”

Generative Adversarial Networks (GANs)

The concept of Generative Adversarial Networks (GANs) is relatively new. GAN sounds like a competition between two deep neural networks. One is a generator that creates novel content with the desired set of criteria and another, called the discriminator, tests whether the generator’s output is true or false.

This technology-powered some interesting results almost immediately. A few teams that used GANs created photo realistic images from natural language in 2016.

The Future of AI in the Pharmaceutical industry looks prominent because of the combination and integration of methods like generative reinforcement learning — and the intriguing prospects of quantum computing. But biology, chemistry, and clinical trials are very complex to be perfectly transparent.

The key to success in pharmaAI is the massive integration of the systems used to identify biological targets, systems that help design novel molecules and systems that personalize the treatments and predict the clinical trial outcomes.

Big pharma brain is required for the discovery and development cycles that can take many years to integrate clinical data back into target discovery. AI powered drug discovery scientists will need to be mixed-martial artists to accelerate small molecule drug discovery.

Determined AI Open Sourcing AI Infrastructure Product

Machine learning is a crucial component of innumerable software stacks. Enterprise-grade and often enterprise-only tools are required to create & manage it. By open-sourcing its entire AI infrastructure product, @Determined AI aims to make them more accessible.

Determined Training Platform created by the company for developing AI in an organized way. They raised an $11 million Series A last year.

Evan Sparks, CEO, Determined AI¹² said “Machinelearning is going to be a big part of how software is developed going forward. But in order for companies like @Google and @Amazon to be productive, they had to build all this software infrastructure,”

ML is being experimented with small teams using tools intended for academic work and individual research at smaller companies. There aren’t a lot of options to scale that up to dozens of engineers developing a real product.

Chief Scientist at Determined AI, @Ameet Talwalkar said, they’re using things like @TensorFlow and @PyTorch. The founders of Determined AI, whose started out at @UC Berkeley’s AmpLab (home of ApacheSpark), has been developing its platform for a few years, with feedback and validation from some paying customers.

Determined AI may grow as a new defacto standard for AI development in businesses.

Works Cited

¹States Re-opening, ²Digital Transformation, ³Robots, ⁴Clean Data, ⁵NLP Models, ⁶AI and ML Tools, ⁷AI Failures, ⁸Customer Service AI, ⁹Limitations of AI

Companies Cited

¹⁰Genesys, ¹¹BetterUP, ¹²Determined AI