AI, Analytics, Machine Learning, Data Science, Deep Learning Technology Main Developments in 2019 and Key Trends for 2020
AI, Analytics, Machine Learning, Data Science, Deep Learning Technology Main Developments in 2019 and Key Trends for 2020We asked leading experts - what are the most important developments of 2019 and 2020 key trends in AI, Analytics, Machine Learning, Data Science, and Deep Learning? This blog focuses mainly on technology and deployment.
In
2019 (and in previous years), we asked a number of top experts their predictions for 2020.
Some of the trends predicted last year have materialized:
We asked our experts again this year:
We received about 20 replies, and the first part, more focused on research was published earlier.
Here is the second part, more focused on Technology, Industry, and Deployment. Some of the common themes were: AI Hype, AutoML, Cloud, Data, Explainable AI, AI Ethics.
Here are the answers from Meta Brown, Tom Davenport, Carla Gentry, Nikita Johnson, Doug Laney, Bill Schmarzo, Kate Strachnyi, Ronald van Loon, Favio Vazquez, and Jen Underwood.
Meta Brown, @metabrown312, is the author of Data Mining for Dummies and President of A4A Brown
In 2018 we had a dramatic increase in the use of the term "artificial intelligence" to describe everything from truly sophisticated applications and increasingly successful such as self-driving vehicles to mundane use of propensity scores in direct marketing. I predicted that, in 2019, people would catch on that this was all just math. I was half right.
On one hand, more people are starting to see the limitations of what's now labelled as "AI." The public is aware that facial recognition technology can be thwarted by Juggalo makeup, that there's no intelligent life behind those customer service chatbots, and that you can spend millions trying to make software be smarter than doctors, and still fail.
Yet, "artificial intelligence" is still a hot buzzword, and the venture capital money is still rolling in. More than $13 billion went to AI startups in the first 9 months of 2019.
In 2020, look for a growing dichotomy between the two outlooks of artificial intelligence: a public image of growing doubt, suspicion and awareness of AI's limitations, and the business and investment communities that continues to invest hopes, dreams and money in AI promises.
Tom Davenport, @tdav, is the President's Distinguished Professor of Information Technology and Management at Babson College, the co-founder of the International Institute for Analytics, a Fellow of the MIT Initiative on the Digital Economy, and a Senior Advisor to Deloitte Analytics.
Major developments in 2019:
Carla Gentry, @data_nerd, is a consulting Data Scientist and Owner of Analytical-Solution.
Another year of hype and buzz over what can and can't be done with AI, Machine Learning and Data Science, I cringe at the amount of unskilled professionals jumping into these fields and the colleges spitting out so called certifications and degrees with teachers who aren't qualified to teach these courses.
Data Science and machine learning are dependent on large amounts of data but we face another year of misunderstandings of biases, data that needs to be interpreted always takes a risk with bias. Unbiased data stands alone, it's doesn't need to be interpreted, example - Mary increased her sales ROI by 10% verses Mary is a hard worker, which is an opinion and can't be measure.
An article's title caught my eye the other day "Is Data Science dying"? My initial thought before even reading was "No, but all the wanna-bees and hype surely didn't help our field - Data Science is more than writing code". Misunderstanding about tech plus the lack of data and the necessary infrastructure will continue to plague us in 2020 but at least SOME are realizing the sexiest job of the 21st Century isn't so sexy after all, as we spend most of our cleansing and prepping data before we get to glean insight and answer business questions.
In 2020 let's all remember it's about the DATA and make sure we can advance our field with integrity and transparency, the days of the "black box" of AI has to be OVER for us to continue in a positive direction. Remember, the algorithms, models, chat bots, etc, you build may have an effect on someone's life, data points in a database correspond to a life, so remove your bias and let the facts speak for themselves...as always, have fun and play with data responsibly.
Nikita Johnson, @teamrework, Founder, RE.WORK Deep Learning & AI
In 2019, we have witnessed breakthroughs in a number of areas that have allowed more widespread adoption of AI, on an unprecedented level. Advancing software techniques such as transfer learning and reinforcement learning have also helped to push the progression of AI breakthroughs and adoption forward, helping to separate system improvements with the constraints of our knowledge as humans.
Next year in 2020, we will see a move towards 'Explainable AI' to provide more transparency, accountability, and reproducibility of AI models and techniques. We need to increase our knowledge of the limitations, as well as the advantages and disadvantages of each tool. Enhanced learning will increase our ability to build trust with the products we use, as well as allowing more justifiable decision making by AI!
Doug Laney, @Doug_Laney, Principal Data Strategist, Caserta, best-selling author of "Infonomics," and visiting professor at the University of Illinois Gies College of Business
The resurgence of AI from its halcyon days in the early '90s, along with the mainstreaming of data science, has been fueled by nothing other than data. Today big data is "just data". Its magnitude, even as it continues to swell, no longer overwhelms storage or computing power. At least there's no longer any excuse that any organization being inhibited by data's bigness. (Hint: cloud.) Indeed, incrementally improved technologies and techniques have emerged, but the vast availability of data spewing from social media platforms, exchanged among partners, harvested from websites, and dribbling off connected devices has led to unforeseen insights, automation, and optimization. It has also spawned new data-centric business models.
In 2020, I envision (no pun intended, or was it?) extended information ecosystems to arise, further enabling AI and data-science powered digital coordination among business partners. Some organizations may choose to fabricate their own data exchange solutions to monetize their and others' information assets. Others will fuel their advanced analytics capabilities via blockchain-backed data exchange platforms and/or data aggregators offering an array of alternative data.
Bill Schmarzo, @schmarzo, is CTO, IoT & Analytics Hitachi Vantara .
Main 2019 Developments
Kate Strachnyi, @StorybyData, DATAcated to telling stories with data | Runner | Mom of 2 | Top Voice in Data Science & Analytics.
In 2019, we've seen a consolidation in the data visualization/ business intelligence software space; with Salesforce acquiring Tableau Software, and Google acquiring Looker. This investment in business intelligence tools demonstrates the value companies place on data democratization and enabling users to more easily view and analyze their data.
What we can expect to see in 2020 is the continued shift towards automating data analytics / data science tasks. Data scientists and engineers require tools that allow them to scale and solve more problems. This need will result in the development of automation tools across several stages of the data science process. For example, some data preparation and cleaning task are partially automated; however, they are difficult to fully automate due to the unique needs of companies. Additional candidates for automation include feature engineering, model selection, amongst others.
Ronald van Loon, @Ronald_vanLoon, a Director of Adversitement, Helping Data Driven Companies Generating Success. Top10 Big Data, Data Science, IoT, AI Influencer
In 2019, the industry witnessed an increasing adoption of Explainable AI and augmented analytics that allowed businesses to bridge the gap between the potential of what AI can deliver and the technological complexity of decision making based on unbiased AI outcomes. The full stack approach to AI was another 2019 development that organizations embraced to help speed the path to innovation and support AI growth while improving integration and communication across different teams and individuals.
In 2020 we'll be seeing some customer experience improvement trends emerge due to Conversational AI's ease of use and intuitive interface. This automated solution enables companies to scale and transform the customer experience while opening up a 24/7 pathway to the customer, and offers opportunities for fast problem resolution and reliable self-service. Also, Narrow Intelligence will continue to support how we leverage human and machine strength most effectively as we fit AI into our existing processes and strive to change the questions we ask of AI.
Favio Vazquez, @FavioVaz, CEO at Closter
In the year 2019, we saw amazing developments in the state of the art of artificial intelligence, mainly in deep learning. Data science has the power to use those advances to solve harder problems and shape the world we live in. Data science is the engine that using science it's catalyzing change, and transforming papers into products. Our field is not just "hype" anymore, it's becoming a serious field. We will see an increase in important online and offline education about data science and its friends. Hopefully, we'll become more confident in what we do and how we do it. Semantic technologies, decision intelligence, and knowledge data science will be our companions in the next years, so I recommend people to start exploring graph databases, ontologies, and knowledge representation systems.
Jen Underwood, @idigdata, a force of nature taking organizations forward faster
In 2019, we reached the tipping point for organizations to get serious about competing in an algorithm economy. Rather sponsor than one-off projects, market-leading companies elevated data science prominence by planning enterprise-wide AI strategies. Meanwhile, mature data science organizations launched ethics, governance and ML Ops initiatives. Unfortunately, while machine learning adoption rates improved success eluded most.
From a technology perspective, we witnessed the rise of hybrid distributed computing and serverless architectures. Concurrently, algorithms, frameworks, and AutoML solutions rapidly advanced from innovation to commoditization.
In 2020, I anticipate personal data security, regulation, algorithm bias, and deep fake topics to dominate headline news. On a brighter note, advances in explainable AI along with natural language generation and optimization techniques to enhance human understanding will help bridge the gap between data science and the business. With the further emergence of data literacy and citizen data science programs, machine learning practitioners should continue to thrive.
Here is the word cloud based on their predictions
Related:
Some of the trends predicted last year have materialized:
- more attention to ethics in AI
- democratization of Data Science
- Reinforcement learning advances
- China growing success in AI
We asked our experts again this year:
What were the main developments in AI, Data Science, Deep Learning, and Machine Learning in 2019, and what key trends do you expect in 2020?
We received about 20 replies, and the first part, more focused on research was published earlier.
Here is the second part, more focused on Technology, Industry, and Deployment. Some of the common themes were: AI Hype, AutoML, Cloud, Data, Explainable AI, AI Ethics.
Here are the answers from Meta Brown, Tom Davenport, Carla Gentry, Nikita Johnson, Doug Laney, Bill Schmarzo, Kate Strachnyi, Ronald van Loon, Favio Vazquez, and Jen Underwood.
Meta Brown, @metabrown312, is the author of Data Mining for Dummies and President of A4A Brown
In 2018 we had a dramatic increase in the use of the term "artificial intelligence" to describe everything from truly sophisticated applications and increasingly successful such as self-driving vehicles to mundane use of propensity scores in direct marketing. I predicted that, in 2019, people would catch on that this was all just math. I was half right.
On one hand, more people are starting to see the limitations of what's now labelled as "AI." The public is aware that facial recognition technology can be thwarted by Juggalo makeup, that there's no intelligent life behind those customer service chatbots, and that you can spend millions trying to make software be smarter than doctors, and still fail.
Yet, "artificial intelligence" is still a hot buzzword, and the venture capital money is still rolling in. More than $13 billion went to AI startups in the first 9 months of 2019.
In 2020, look for a growing dichotomy between the two outlooks of artificial intelligence: a public image of growing doubt, suspicion and awareness of AI's limitations, and the business and investment communities that continues to invest hopes, dreams and money in AI promises.
Tom Davenport, @tdav, is the President's Distinguished Professor of Information Technology and Management at Babson College, the co-founder of the International Institute for Analytics, a Fellow of the MIT Initiative on the Digital Economy, and a Senior Advisor to Deloitte Analytics.
Major developments in 2019:
- Widespread deployment of automated machine learning tools for the more structured aspects of data science.
- Broad realization that analytics and AI have an ethical dimension that needs to be consciously addressed
- Increasing recognition that most analytical and AI models aren't deployed and don't have value to the organizations creating them as a result
- Availability of tools to create, manage and monitor an organization's suite of machine learning models, with continuous retraining of drifting models and a focus on model inventory management.
- Improved status and recognition for analytical and AI translators, who work with business users and leaders to translate business requirements into high-level specifications for models
- Recognition that the fit of a model to data is only one consideration in whether it is useful or not.
Carla Gentry, @data_nerd, is a consulting Data Scientist and Owner of Analytical-Solution.
Another year of hype and buzz over what can and can't be done with AI, Machine Learning and Data Science, I cringe at the amount of unskilled professionals jumping into these fields and the colleges spitting out so called certifications and degrees with teachers who aren't qualified to teach these courses.
Data Science and machine learning are dependent on large amounts of data but we face another year of misunderstandings of biases, data that needs to be interpreted always takes a risk with bias. Unbiased data stands alone, it's doesn't need to be interpreted, example - Mary increased her sales ROI by 10% verses Mary is a hard worker, which is an opinion and can't be measure.
An article's title caught my eye the other day "Is Data Science dying"? My initial thought before even reading was "No, but all the wanna-bees and hype surely didn't help our field - Data Science is more than writing code". Misunderstanding about tech plus the lack of data and the necessary infrastructure will continue to plague us in 2020 but at least SOME are realizing the sexiest job of the 21st Century isn't so sexy after all, as we spend most of our cleansing and prepping data before we get to glean insight and answer business questions.
In 2020 let's all remember it's about the DATA and make sure we can advance our field with integrity and transparency, the days of the "black box" of AI has to be OVER for us to continue in a positive direction. Remember, the algorithms, models, chat bots, etc, you build may have an effect on someone's life, data points in a database correspond to a life, so remove your bias and let the facts speak for themselves...as always, have fun and play with data responsibly.
Nikita Johnson, @teamrework, Founder, RE.WORK Deep Learning & AI
In 2019, we have witnessed breakthroughs in a number of areas that have allowed more widespread adoption of AI, on an unprecedented level. Advancing software techniques such as transfer learning and reinforcement learning have also helped to push the progression of AI breakthroughs and adoption forward, helping to separate system improvements with the constraints of our knowledge as humans.
Next year in 2020, we will see a move towards 'Explainable AI' to provide more transparency, accountability, and reproducibility of AI models and techniques. We need to increase our knowledge of the limitations, as well as the advantages and disadvantages of each tool. Enhanced learning will increase our ability to build trust with the products we use, as well as allowing more justifiable decision making by AI!
Doug Laney, @Doug_Laney, Principal Data Strategist, Caserta, best-selling author of "Infonomics," and visiting professor at the University of Illinois Gies College of Business
The resurgence of AI from its halcyon days in the early '90s, along with the mainstreaming of data science, has been fueled by nothing other than data. Today big data is "just data". Its magnitude, even as it continues to swell, no longer overwhelms storage or computing power. At least there's no longer any excuse that any organization being inhibited by data's bigness. (Hint: cloud.) Indeed, incrementally improved technologies and techniques have emerged, but the vast availability of data spewing from social media platforms, exchanged among partners, harvested from websites, and dribbling off connected devices has led to unforeseen insights, automation, and optimization. It has also spawned new data-centric business models.
In 2020, I envision (no pun intended, or was it?) extended information ecosystems to arise, further enabling AI and data-science powered digital coordination among business partners. Some organizations may choose to fabricate their own data exchange solutions to monetize their and others' information assets. Others will fuel their advanced analytics capabilities via blockchain-backed data exchange platforms and/or data aggregators offering an array of alternative data.
Bill Schmarzo, @schmarzo, is CTO, IoT & Analytics Hitachi Vantara .
Main 2019 Developments
- Growing "consumer proof points" with respect to AI integration into our everyday lives via smart phones, web sites, home devices and vehicles.
- Formalization of DataOps category as an acknowledgment of the growing importance of the Data Engineering role
- Growing respect for the business potential of data science within the executive suite.
- CIOs continue to struggle to deliver on the data monetization promise; Data Lake disillusionment leading to Data Lake "second surgeries"
- More real-world examples of industrial companies leveraging sensors, edge analytics and AI to create products that get more intelligent through usage; they appreciate, not depreciate, in value with usage
- Grandiose smart spaces projects continue to struggle to grow beyond initial pilots due to inability to deliver reasonable financial or operational impact.
- Recession will drive chasm between "Have's" and "Have Not's" with respect to organizations that leverage data and analytics to drive meaningful business results
Kate Strachnyi, @StorybyData, DATAcated to telling stories with data | Runner | Mom of 2 | Top Voice in Data Science & Analytics.
In 2019, we've seen a consolidation in the data visualization/ business intelligence software space; with Salesforce acquiring Tableau Software, and Google acquiring Looker. This investment in business intelligence tools demonstrates the value companies place on data democratization and enabling users to more easily view and analyze their data.
What we can expect to see in 2020 is the continued shift towards automating data analytics / data science tasks. Data scientists and engineers require tools that allow them to scale and solve more problems. This need will result in the development of automation tools across several stages of the data science process. For example, some data preparation and cleaning task are partially automated; however, they are difficult to fully automate due to the unique needs of companies. Additional candidates for automation include feature engineering, model selection, amongst others.
Ronald van Loon, @Ronald_vanLoon, a Director of Adversitement, Helping Data Driven Companies Generating Success. Top10 Big Data, Data Science, IoT, AI Influencer
In 2019, the industry witnessed an increasing adoption of Explainable AI and augmented analytics that allowed businesses to bridge the gap between the potential of what AI can deliver and the technological complexity of decision making based on unbiased AI outcomes. The full stack approach to AI was another 2019 development that organizations embraced to help speed the path to innovation and support AI growth while improving integration and communication across different teams and individuals.
In 2020 we'll be seeing some customer experience improvement trends emerge due to Conversational AI's ease of use and intuitive interface. This automated solution enables companies to scale and transform the customer experience while opening up a 24/7 pathway to the customer, and offers opportunities for fast problem resolution and reliable self-service. Also, Narrow Intelligence will continue to support how we leverage human and machine strength most effectively as we fit AI into our existing processes and strive to change the questions we ask of AI.
Favio Vazquez, @FavioVaz, CEO at Closter
In the year 2019, we saw amazing developments in the state of the art of artificial intelligence, mainly in deep learning. Data science has the power to use those advances to solve harder problems and shape the world we live in. Data science is the engine that using science it's catalyzing change, and transforming papers into products. Our field is not just "hype" anymore, it's becoming a serious field. We will see an increase in important online and offline education about data science and its friends. Hopefully, we'll become more confident in what we do and how we do it. Semantic technologies, decision intelligence, and knowledge data science will be our companions in the next years, so I recommend people to start exploring graph databases, ontologies, and knowledge representation systems.
Jen Underwood, @idigdata, a force of nature taking organizations forward faster
In 2019, we reached the tipping point for organizations to get serious about competing in an algorithm economy. Rather sponsor than one-off projects, market-leading companies elevated data science prominence by planning enterprise-wide AI strategies. Meanwhile, mature data science organizations launched ethics, governance and ML Ops initiatives. Unfortunately, while machine learning adoption rates improved success eluded most.
From a technology perspective, we witnessed the rise of hybrid distributed computing and serverless architectures. Concurrently, algorithms, frameworks, and AutoML solutions rapidly advanced from innovation to commoditization.
In 2020, I anticipate personal data security, regulation, algorithm bias, and deep fake topics to dominate headline news. On a brighter note, advances in explainable AI along with natural language generation and optimization techniques to enhance human understanding will help bridge the gap between data science and the business. With the further emergence of data literacy and citizen data science programs, machine learning practitioners should continue to thrive.
Here is the word cloud based on their predictions
Related:
- AI, Analytics, Machine Learning, Data Science, Deep Learning Research Main Developments in 2019 and Key Trends for 2020
- Machine Learning & AI Main Developments in 2018 and Key Trends for 2019
- AI, Data Science, Analytics Main Developments in 2018 and Key Trends for 2019